From: "Liu, Yong" <yong.liu@intel.com>
To: "Xu, Qian Q" <qian.q.xu@intel.com>, "dts@dpdk.org" <dts@dpdk.org>
Subject: Re: [dts] [PATCH 1/2][dts v2] Update vmdq test scripts for supporting FVL/Niantic.
Date: Wed, 16 Sep 2015 02:45:38 +0000 [thread overview]
Message-ID: <86228AFD5BCD8E4EBFD2B90117B5E81E10EF46BF@SHSMSX103.ccr.corp.intel.com> (raw)
In-Reply-To: <1442371032-25142-1-git-send-email-qian.q.xu@intel.com>
Thanks Qian. Applied.
> -----Original Message-----
> From: dts [mailto:dts-bounces@dpdk.org] On Behalf Of Qian Xu
> Sent: Wednesday, September 16, 2015 10:37 AM
> To: dts@dpdk.org
> Subject: [dts] [PATCH 1/2][dts v2] Update vmdq test scripts for supporting
> FVL/Niantic.
>
> [v2]
> Minor changes on the code format and value assignment.
> [v1]
> Updated vmdq test suite, and it works for both FVL and Niantic. The major
> changes are as below:
> 1. Support FVL 4x10g, 2x10g, 2x40g.
> 2. Update the traffic to make the MAC address and VLAN id changes together.
> Then the packet loss is 0 at low traffic.
> 3. Set the max pools number for major NICs: FVL 40G, 10G, Niantic, 1G.
> 4. Update the core config.
>
> Signed-off-by: Qian Xu <qian.q.xu@intel.com>
>
> diff --git a/tests/TestSuite_vmdq.py b/tests/TestSuite_vmdq.py
> new file mode 100644
> index 0000000..de5a86f
> --- /dev/null
> +++ b/tests/TestSuite_vmdq.py
> @@ -0,0 +1,262 @@
> +# <COPYRIGHT_TAG>
> +
> +"""
> +DPDK Test suite.
> +
> +Tests for vmdq.
> +
> +"""
> +
> +import dts
> +import re
> +from etgen import IxiaPacketGenerator
> +from test_case import TestCase
> +from time import sleep
> +
> +
> +class TestVmdq(TestCase, IxiaPacketGenerator):
> + 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.send_expect("sed -i
> 's/CONFIG_RTE_MAX_QUEUES_PER_PORT=256/CONFIG_RTE_MAX_QUEUES_PER_PORT=1024/
> ' ./config/common_linuxapp", "# ", 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)
> + print self.dut_ports
> + self.verify(len(self.dut_ports) >= 2, "Insufficient 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")
> +
> +
> + def validateApproxEqual(self, lines):
> + """
> + Check that all the rx queue stats are within a 30% range.
> + """
> +
> + minimum = 1000000
> + maximun = 0
> +
> + # Need to use Python re package because dts.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):
> + """
> + MAX queues is 128
> + queues/pools = 128/npools
> + """
> +
> + 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_mask = dts.create_mask(core_list)
> +
> + port_mask = dts.create_mask([self.dut_ports[0],
> self.dut_ports[1]])
> + # 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)
> +
> + # Transmit traffic
> + 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)
> + self.tester.scapy_append('wrpcap("test%d.pcap", flows)' %p)
> + self.tester.scapy_execute()
> + tgen_input.append((tx_port, rx_port, "test%d.pcap" %p))
> +
> + loss = self.tester.traffic_generator_loss(tgen_input, 10)
> + print "loss is %s !" % loss
> +
> + # Verify the accurate
> + self.verify(loss < 0.001, "Excessive packet loss")
> + self.validateApproxEqual(out.split("\r\n"))
> +
> + def set_up(self):
> + """
> + Run before each test case.
> + """
> + self.dut.kill_all()
> +
> + def test_perf_vmdq_64pools_queues(self):
> + """
> + This function call " Npools_128queues" with differen 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)
> +
> + def test_perf_vmdq_performance(self):
> + """
> + Try different configuration and different packe 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:
> +
> + print dts.BLUE(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 = dts.create_mask(core_list)
> + portmask = dts.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])
> +
> + print dts.GREEN("Waiting for application to initialize")
> + sleep(5)
> +
> + for frame_size in frame_sizes:
> +
> + TestVmdq.current_frame_size = frame_size
> +
> + print dts.BLUE(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))
> + self.tester.scapy_append('wrpcap("test1.pcap", flows)')
> + 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))
> + self.tester.scapy_append('wrpcap("test2.pcap", flows)')
> + self.tester.scapy_execute()
> +
> + self.vlan_repeat = self.queues
> + self.da_repeat = self.queues
> +
> + tgen_input = []
> + tgen_input.append((tx_port, rx_port, "test1.pcap"))
> + tgen_input.append((rx_port, tx_port, "test2.pcap"))
> + _, pps =
> self.tester.traffic_generator_throughput(tgen_input)
> + 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")
> +
> + # Print results
> + dts.results_table_add_header(
> + ['Frame size'] + [n['cores'] for n in self.core_configs])
> +
> + for size in frame_sizes:
> + dts.results_table_add_row(
> + [size] + [n['mpps'][size] for n in self.core_configs])
> +
> + dts.results_table_print()
> +
> + # Override etgen.dot1q function
> + def dot1q(self, port, prio, id, vlan, type):
> + """
> + Change Ixia configuration
> + """
> +
> + self.add_tcl_cmd("vlan config -mode vIncrement")
> + self.add_tcl_cmd("vlan config -step 1")
> + self.add_tcl_cmd("vlan config -repeat %d" % self.vlan_repeat)
> + self.add_tcl_cmd("stream config -framesize %d" %
> + TestVmdq.current_frame_size)
> + super(TestVmdq, self).dot1q(port, prio, id, vlan, type)
> +
> + def ether(self, port, src, dst, type):
> + """
> + Configure Ether protocal.
> + """
> + self.add_tcl_cmd("protocol config -ethernetType ethernetII")
> + self.add_tcl_cmd('stream config -sa "%s"' %
> self.macToTclFormat(src))
> + self.add_tcl_cmd('stream config -da "%s"' %
> self.macToTclFormat(dst))
> + self.add_tcl_cmd('stream config -daRepeatCounter increment')
> + self.add_tcl_cmd('stream config -daStep 1')
> + self.add_tcl_cmd('stream config -numDA %d' % self.da_repeat)
> +
> +
> --
> 2.1.0
prev parent reply other threads:[~2015-09-16 2:45 UTC|newest]
Thread overview: 3+ messages / expand[flat|nested] mbox.gz Atom feed top
2015-09-16 2:37 Qian Xu
2015-09-16 2:37 ` [dts] [PATCH 2/2][dts v2] Update vmdq test plan to align with scripts Qian Xu
2015-09-16 2:45 ` Liu, Yong [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=86228AFD5BCD8E4EBFD2B90117B5E81E10EF46BF@SHSMSX103.ccr.corp.intel.com \
--to=yong.liu@intel.com \
--cc=dts@dpdk.org \
--cc=qian.q.xu@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).