test suite reviews and discussions
 help / color / mirror / Atom feed
From: "Tu, Lijuan" <lijuan.tu@intel.com>
To: "Han, YingyaX" <yingyax.han@intel.com>, "dts@dpdk.org" <dts@dpdk.org>
Cc: "Han, YingyaX" <yingyax.han@intel.com>
Subject: Re: [dts] [PATCH V2]tests/vmdq: vmdq script refactoring
Date: Tue, 31 Mar 2020 03:05:47 +0000	[thread overview]
Message-ID: <8CE3E05A3F976642AAB0F4675D0AD20E0BC034FF@SHSMSX101.ccr.corp.intel.com> (raw)
In-Reply-To: <1585112902-137050-1-git-send-email-yingyax.han@intel.com>

Applied, thanks

> -----Original Message-----
> From: dts [mailto:dts-bounces@dpdk.org] On Behalf Of hanyingya
> Sent: Wednesday, March 25, 2020 1:08 PM
> To: dts@dpdk.org
> Cc: Han, YingyaX <yingyax.han@intel.com>
> Subject: [dts] [PATCH V2]tests/vmdq: vmdq script refactoring
> 
> script refactoring and enable rss.
> 
> Signed-off-by: hanyingya <yingyax.han@intel.com>
> ---
>  tests/TestSuite_vmdq.py | 328 +++++++++++++++++-----------------------
>  1 file changed, 142 insertions(+), 186 deletions(-)
> 
> diff --git a/tests/TestSuite_vmdq.py b/tests/TestSuite_vmdq.py index
> 623d728..cb6bdf0 100644
> --- a/tests/TestSuite_vmdq.py
> +++ b/tests/TestSuite_vmdq.py
> @@ -11,44 +11,52 @@ import os
>  import re
>  from test_case import TestCase
>  from time import sleep
> +from settings import HEADER_SIZE
>  from pktgen import PacketGeneratorHelper
> 
> +
>  class TestVmdq(TestCase):
> -    dut_ports = []
> -    ip_dot1q_header_size = 22
> -    default_framesize = 64
> -    default_payload = default_framesize - ip_dot1q_header_size
> -    current_frame_size = 0
> -    destmac_port0 = "52:54:00:12:00:00"
> -    destmac_port1 = "52:54:00:12:01:00"
> -    da_repeat = 1
> -    vlan_repeat = 1
> -    queues = 8
> 
>      def set_up_all(self):
>          """
>          Run at the start of each test suite.
>          """
>          self.tester.extend_external_packet_generator(TestVmdq, self)
> -
> +
> +        self.dut_ports = self.dut.get_ports(self.nic)
> +        self.verify(len(self.dut_ports) >= 2, "Insufficient ports")
> +        self.ports_socket = self.dut.get_numa_id(self.dut_ports[0])
> +
>          self.dut.send_expect("sed -i
> 's/CONFIG_RTE_MAX_QUEUES_PER_PORT=256/CONFIG_RTE_MAX_QUEUES_
> PER_PORT=1024/' ./config/common_base", "# ", 5)
> -
>          self.dut.build_install_dpdk(self.target)
>          # Update the max queue per port for Fortville.
>          self.dut.send_expect("sed -i 's/define MAX_QUEUES 128/define
> MAX_QUEUES 1024/' ./examples/vmdq/main.c", "#", 5)
> -
> -        self.dut_ports = self.dut.get_ports(self.nic)
> -        self.verify(len(self.dut_ports) >= 2, "Insufficient ports")
> +        out = self.dut.send_expect("make -C examples/vmdq", "#", 10)
> +        self.verify("Error" not in out, "Compilation error")
> 
> +        self.frame_size = 64
> +        self.header_size = HEADER_SIZE['ip'] + HEADER_SIZE['eth']
> +        self.destmac_port = ["52:54:00:12:0%d:00" % i for i in
> + self.dut_ports]
>          self.core_configs = []
>          self.core_configs.append({'cores': '1S/1C/1T', 'mpps': {}})
>          self.core_configs.append({'cores': '1S/2C/1T', 'mpps': {}})
>          self.core_configs.append({'cores': '1S/2C/2T', 'mpps': {}})
>          self.core_configs.append({'cores': '1S/4C/1T', 'mpps': {}})
> 
> -        self.ports_socket = self.dut.get_numa_id(self.dut_ports[0])
> -        out = self.dut.send_expect("make -C examples/vmdq", "#", 10)
> -        self.verify("Error" not in out, "Compilation error")
> +        # Put different number of pools: in the case of 10G 82599 Nic is 64, in
> the case
> +        # of FVL spirit is 63,in case of FVL eagle is 34.
> +        if self.nic in ("niantic", "springfountain"):
> +            self.pools = 64
> +        elif self.nic in ("fortville_spirit", "fortville_spirit_single"):
> +            self.pools = 63
> +        elif self.nic in ("fortville_eagle"):
> +            self.pools = 34
> +        else:
> +            self.pools = 8
> +
> +        # get dts output path
> +        if self.logger.log_path.startswith(os.sep):
> +            self.output_path = self.logger.log_path
>          # get dts output path
>          if self.logger.log_path.startswith(os.sep):
>              self.output_path = self.logger.log_path @@ -58,208 +66,137 @@
> class TestVmdq(TestCase):
>              self.output_path = os.sep.join([cur_path, self.logger.log_path])
>          # create an instance to set stream field setting
>          self.pktgen_helper = PacketGeneratorHelper()
> +        self.prios = range(8)
> 
> -    def validateApproxEqual(self, lines):
> +    def set_up(self):
>          """
> -        Check that all the rx queue stats are within a 30% range.
> +        Run before each test case.
>          """
> -        minimum = 1000000
> -        maximun = 0
> +        pass
> 
> -        # Need to use Python re package because utils.regexp only handles 1
> group,
> -        # we need 4.
> -        scanner = re.compile(
> -            "^Pool [0-9]+: ([0-9]+) ([0-9]+) ([0-9]+) ([0-9]+)$")
> -        for l in lines:
> -            m = scanner.search(l)
> -            if m is None:
> -                # Line at the end, "Finished handling signal", ignore
> -                pass
> -            else:
> -                for stat in m.groups():
> -                    if stat < minimum:
> -                        minimum = stat
> -                    if stat > maximun:
> -                        maximun = stat
> -        self.verify(maximun - minimum <= minimum *
> -                    0.3, "Too wide variation in queue stats")
> -
> -    def Npools_128queues(self, npools):
> +    def start_application(self, npools, core_config):
>          """
> -        MAX queues is 128
> -        queues/pools = 128/npools
> +        Prepare the commandline and start vmdq app
>          """
> -        self.current_frame_size = self.default_framesize
> -
> -        self.dut_ports = self.dut.get_ports(self.nic)
> -
> -        core_list = self.dut.get_core_list("1S/4C/1T", socket=self.ports_socket)
> +        core_list = self.dut.get_core_list(core_config, socket=self.ports_socket)
> +        self.verify(core_list is not None, "Requested cores failed")
>          core_mask = utils.create_mask(core_list)
> -
> -        port_mask = utils.create_mask([self.dut_ports[0], self.dut_ports[1]])
> +        port_mask = utils.create_mask(self.dut_ports)
>          # Run the application
> -        out = self.dut.send_expect("./examples/vmdq/build/vmdq_app -n 4 -
> c %s -- -p %s --nb-pools %s&" %
> -                                   (core_mask, port_mask, str(npools)), "reading queues",
> 120)
> +        self.dut.send_expect("./examples/vmdq/build/vmdq_app -c %s -n 4 -- -
> p %s --nb-pools %s --enable-rss" %
> +                             (core_mask, port_mask, str(npools)),
> + "reading queues", 120)
> 
> -        # Transmit traffic
> +    def get_tgen_input(self, prios):
> +        """
> +        create streams for ports.
> +        """
>          tx_port = self.tester.get_local_port(self.dut_ports[0])
>          rx_port = self.tester.get_local_port(self.dut_ports[1])
> -        tx_mac = self.tester.get_mac(tx_port)
> -
> -        self.vlan_repeat = npools
> -        self.da_repeat = npools
>          tgen_input = []
> -        for p in range(8):
> -            self.tester.scapy_append('dmac="%s"' % self.destmac_port0)
> -            self.tester.scapy_append('smac="%s"' % tx_mac)
> -            self.tester.scapy_append(
> -                'flows = [Ether(src=smac, dst=dmac)/Dot1Q(vlan=0,prio=%d)]'%p)
> +        for prio in prios:
> +            pcap = os.sep.join([self.output_path, "%s%d.pcap" %
> (self.suite_name, prio)])
> +            tgen_input.append((tx_port, rx_port, "%s" % pcap))
> +        return tgen_input
> 
> -            pcap = os.sep.join([self.output_path, "test%d.pcap" % p])
> -            self.tester.scapy_append('wrpcap("%s", flows)' %pcap)
> -
> -            self.tester.scapy_execute()
> -            tgen_input.append((tx_port, rx_port, "%s" %pcap))
> -
> -        self.tester.pktgen.clear_streams()
> -        vm_config = self.set_fields()
> -        # run packet generator
> -        streams = self.pktgen_helper.prepare_stream_from_tginput(tgen_input,
> 10,
> -                                                        vm_config, self.tester.pktgen)
> -        loss = self.tester.pktgen.measure_loss(stream_ids=streams)
> -        self.logger.info("loss is {}!".format(loss))
> +    def create_pcaps(self, prios):
> +        """
> +        create traffic flows to pcap files
> +        """
> +        payload = self.frame_size - self.header_size
> +        for prio in prios:
> +            self.tester.scapy_append(
> +                'flows = [Ether(dst="%s")/Dot1Q(vlan=0,prio=%d)/IP(src="1.2.3.4",
> dst="1.1.1.1")/("X"*%d)]'
> +                % (self.destmac_port[0], prio, payload))
> +            pcap = os.sep.join([self.output_path, "%s%d.pcap" %
> (self.suite_name, prio)])
> +            self.tester.scapy_append('wrpcap("%s", flows)' % pcap)
> +        self.tester.scapy_execute()
> 
> -        # Verify the accurate
> -        self.verify(loss[0]/100 < 0.001, "Excessive packet loss")
> +    def verify_all_vmdq_stats(self):
> +        """
> +        Every RX queue should have received approximately (+/-15%) the same
> number of incoming packets.
> +        """
>          out = self.get_vmdq_stats()
> -        self.validateApproxEqual(out.split("\r\n"))
> +        lines_list = out.split("\r\n")
> +        nb_packets = []
> +        for pool_info in lines_list:
> +            if pool_info.startswith('Pool'):
> +                nb_packets += pool_info.split()[2:]
> +        nb_packets = list(map(int, nb_packets))
> +        self.verify(min(nb_packets) > 0, "Some queues don't get any packet!")
> +        self.verify(float((max(nb_packets) - min(nb_packets))/max(nb_packets))
> <= 0.15,
> +                    "Too wide variation in queue stats")
> 
>      def get_vmdq_stats(self):
> -        vmdq_dcb_session = self.dut.new_session()
> -        vmdq_dcb_session.send_expect("kill -s SIGHUP  `pgrep -fl vmdq_app |
> awk '{print $1}'`", "#", 20)
> +        vmdq_session = self.dut.new_session()
> +        vmdq_session.send_expect("kill -s SIGHUP  `pgrep -fl vmdq_app |
> + awk '{print $1}'`", "#", 20)
>          out = self.dut.get_session_output()
>          self.logger.info(out)
> -        vmdq_dcb_session.close()
> +        vmdq_session.close()
>          return out
> 
> -    def set_up(self):
> +    def test_perf_vmdq_max_queues(self):
>          """
> -        Run before each test case.
> +        Every RX queue should have received approximately (+/-15%) the same
> number of
> +        incoming packets.
>          """
> -        self.dut.kill_all()
> -
> -    def test_perf_vmdq_64pools_queues(self):
> -        """
> -        This function call "Npools_128queues" with different number
> -        of pools. Details see below. If not sure, set it as 8 pools.
> -        """
> -        if self.nic in ("niantic", "springfountain"):
> -            self.Npools_128queues(64)
> -        elif self.nic in ("fortville_spirit", "fortville_spirit_single"):
> -            self.Npools_128queues(63)
> -        elif self.nic in ("fortville_eagle"):
> -            self.Npools_128queues(34)
> -        else:
> -            self.Npools_128queues(8)
> +        # Run the application
> +        self.start_application(self.pools, "1S/4C/1T")
> +        # Transmit traffic
> +        self.create_pcaps(self.prios)
> +        tgen_input = self.get_tgen_input(self.prios)
> +        vm_config = self.set_fields(self.pools, self.pools)
> +        # Start traffic transmission using approx 10% of line rate.
> +        ratePercent = 10
> +        # run packet generator
> +        streams =
> self.pktgen_helper.prepare_stream_from_tginput(tgen_input, ratePercent,
> vm_config, self.tester.pktgen)
> +        # set traffic option
> +        options = {'duration': 15}
> +        loss = self.tester.pktgen.measure_loss(stream_ids=streams,
> options=options)
> +        self.logger.info("loss is [loss rate, SendNumbers,
> ReceNumbers]{}!".format(loss))
> +        # Verify there is no packet loss
> +        self.verify(loss[1] == loss[2], "Packet Loss! Send: %d, but only
> Receive: %d!".format(loss[1], loss[2]))
> +        self.verify_all_vmdq_stats()
> +
> +    def create_throughput_traffic(self, frame_size):
> +        payload = frame_size - self.header_size
> +        tgen_Input = []
> +        for _port in self.dut_ports:
> +            if _port % len(self.dut_ports) == 0 or len(self.dut_ports) % _port == 2:
> +                txIntf = self.tester.get_local_port(self.dut_ports[_port + 1])
> +            else:
> +                txIntf = self.tester.get_local_port(self.dut_ports[_port - 1])
> +            rxIntf = self.tester.get_local_port(self.dut_ports[_port])
> +            self.tester.scapy_append('flows =
> [Ether(dst="%s")/Dot1Q(vlan=0)/("X"*%d)]'
> +                                     % (self.destmac_port[txIntf], payload))
> +            pcap = os.sep.join([self.output_path, "%s-%d.pcap" %
> (self.suite_name, _port)])
> +            self.tester.scapy_append('wrpcap("%s", flows)' % pcap)
> +            self.tester.scapy_execute()
> +            tgen_Input.append((txIntf, rxIntf, pcap))
> +        return tgen_Input
> 
>      def test_perf_vmdq_performance(self):
>          """
>          Try different configuration and different packet size
>          """
> -        self.tester.get_interface(
> -            self.tester.get_local_port(self.dut_ports[0]))
> -
>          frame_sizes = [64, 128, 256, 512, 1024, 1280, 1518]
>          for config in self.core_configs:
> -
>              self.logger.info(config["cores"])
>              self.dut.kill_all()
> -
>              core_config = config['cores']
> -            core_list = self.dut.get_core_list(core_config,socket=self.ports_socket)
> -            core_mask = utils.create_mask(core_list)
> -            portmask = utils.create_mask(self.dut.get_ports())
> -            if self.nic in ("niantic", "springfountain"):
> -                self.queues = 64
> -                self.dut.send_expect(
> -                    "examples/vmdq/build/vmdq_app -n %d -c %s -- -p %s --nb-
> pools 64&" %
> -                    (self.dut.get_memory_channels(), core_mask, portmask),
> "reading queues", 30)
> -            elif self.nic in ("fortville_spirit", "fortville_spirit_single"):
> -                self.queues = 63
> -                self.dut.send_expect(
> -                    "examples/vmdq/build/vmdq_app -n %d -c %s -- -p %s --nb-
> pools 63&" %
> -                    (self.dut.get_memory_channels(), core_mask, portmask),
> "reading queues", 30)
> -            elif self.nic in ("fortville_eagle"):
> -                self.queues = 34
> -                self.dut.send_expect(
> -                    "examples/vmdq/build/vmdq_app -n %d -c %s -- -p %s --nb-
> pools 34&" %
> -                    (self.dut.get_memory_channels(), core_mask, portmask),
> "reading queues", 30)
> -            else:
> -                self.queues = 8
> -                self.dut.send_expect(
> -                    "examples/vmdq/build/vmdq_app -n %d -c %s -- -p %s --nb-
> pools 8&" %
> -                    (self.dut.get_memory_channels(), core_mask, portmask),
> "reading queues", 30)
> -
> -            tx_port = self.tester.get_local_port(self.dut_ports[0])
> -            rx_port = self.tester.get_local_port(self.dut_ports[1])
> -
> +            self.start_application(self.pools, core_config)
>              self.logger.info("Waiting for application to initialize")
>              sleep(5)
> -
>              for frame_size in frame_sizes:
> -                TestVmdq.current_frame_size = frame_size
>                  self.logger.info(str(frame_size))
> -
> -                self.tester.scapy_append('dstmac="%s"' % self.destmac_port0)
> -                tx_mac = self.tester.get_mac(tx_port)
> -                self.tester.scapy_append('srcmac="%s"' % tx_mac)
> -                self.tester.scapy_append(
> -                        'flows =
> [Ether(src=srcmac,dst=dstmac)/Dot1Q(vlan=0)/("X"*%d)]' %
> -                        (frame_size - TestVmdq.ip_dot1q_header_size))
> -
> -                pcap = os.sep.join([self.output_path, "test1.pcap"])
> -                self.tester.scapy_append('wrpcap("%s", flows)' %pcap)
> -
> -                self.tester.scapy_execute()
> -
> -                self.tester.scapy_append('dstmac="%s"' % self.destmac_port1)
> -                tx_mac = self.tester.get_mac(rx_port)
> -                self.tester.scapy_append('srcmac="%s"' % tx_mac)
> -                self.tester.scapy_append(
> -                        'flows =
> [Ether(src=srcmac,dst=dstmac)/Dot1Q(vlan=0)/("X"*%d)]' %
> -                        (frame_size - TestVmdq.ip_dot1q_header_size))
> -
> -                pcap = os.sep.join([self.output_path, "test2.pcap"])
> -                self.tester.scapy_append('wrpcap("%s", flows)' % pcap)
> -
> -                self.tester.scapy_execute()
> -
> -                self.vlan_repeat = self.queues
> -                self.da_repeat = self.queues
> -
> -                tgen_input = []
> -
> -                pcap1 = os.sep.join([self.output_path, "test1.pcap"])
> -                pcap2 = os.sep.join([self.output_path, "test2.pcap"])
> -
> -                tgen_input.append((tx_port, rx_port, pcap1))
> -                tgen_input.append((rx_port, tx_port, pcap2))
> -
> +                tgen_input = self.create_throughput_traffic(frame_size)
>                  # clear streams before add new streams
>                  self.tester.pktgen.clear_streams()
> -                vm_config = self.set_fields()
> +                vm_config = self.set_fields(self.pools, self.pools)
>                  # run packet generator
> -                streams =
> self.pktgen_helper.prepare_stream_from_tginput(tgen_input, 100,
> -                                                    vm_config, self.tester.pktgen)
> +                streams =
> + self.pktgen_helper.prepare_stream_from_tginput(tgen_input, 100,
> + vm_config, self.tester.pktgen)
>                  _, pps =
> self.tester.pktgen.measure_throughput(stream_ids=streams)
> -
> -                config['mpps'][frame_size] = pps/1000000.0
> -
> -        for n in range(len(self.core_configs)):
> -            for size in frame_sizes:
> -                self.verify(
> -                    self.core_configs[n]['mpps'][size] is not 0, "No traffic detected")
> -
> +                self.verify(pps > 0, "No traffic detected")
> +                config['mpps'][frame_size] = pps / 1000000.0
>          # Print results
>          self.result_table_create(
>              ['Frame size'] + [n['cores'] for n in self.core_configs]) @@ -271,12
> +208,31 @@ class TestVmdq(TestCase):
>          self.result_table_print()
> 
>      # Override etgen.dot1q function
> -    def set_fields(self):
> -        ''' set ip protocol field behavior '''
> +    def set_fields(self, vid_range, dmac_range):
> +        """
> +        set ip protocol field behavior
> +        """
>          fields_config = {
> -        'vlan':  {
> -            0: {'range': self.vlan_repeat, 'action': 'inc'}},
> -        'mac':  {
> -            'dst': {'range': self.da_repeat, 'action': 'inc'}},}
> -
> +            'vlan': {
> +                0: {'range': vid_range, 'action': 'inc'}},
> +            'mac': {
> +                'dst': {'range': dmac_range, 'action': 'inc'}},
> +            'ip': {'src': {'action': 'random'}},
> +        }
>          return fields_config
> +
> +    def tear_down(self):
> +        """
> +        Run after each test case.
> +        """
> +        self.dut.kill_all()
> +
> +    def tear_down_all(self):
> +        """
> +        Run after each test suite.
> +        """
> +        # resume setting
> +        self.dut.send_expect(
> +            "sed -i
> 's/CONFIG_RTE_MAX_QUEUES_PER_PORT=1024/CONFIG_RTE_MAX_QUEUES
> _PER_PORT=256/' ./config/common_base",
> +            "# ", 5)
> +        self.dut.build_install_dpdk(self.target)
> --
> 2.17.1


      reply	other threads:[~2020-03-31  3:05 UTC|newest]

Thread overview: 2+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2020-03-25  5:08 hanyingya
2020-03-31  3:05 ` Tu, Lijuan [this message]

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=8CE3E05A3F976642AAB0F4675D0AD20E0BC034FF@SHSMSX101.ccr.corp.intel.com \
    --to=lijuan.tu@intel.com \
    --cc=dts@dpdk.org \
    --cc=yingyax.han@intel.com \
    /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).