From patchwork Thu Sep 22 14:29:47 2022 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 8bit X-Patchwork-Submitter: "Li, HongboX" X-Patchwork-Id: 116621 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 EC492A0540; Thu, 22 Sep 2022 08:18:26 +0200 (CEST) Received: from [217.70.189.124] (localhost [127.0.0.1]) by mails.dpdk.org (Postfix) with ESMTP id E664D4067C; Thu, 22 Sep 2022 08:18:26 +0200 (CEST) Received: from mga03.intel.com (mga03.intel.com [134.134.136.65]) by mails.dpdk.org (Postfix) with ESMTP id 2320C40156 for ; Thu, 22 Sep 2022 08:18:24 +0200 (CEST) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=intel.com; i=@intel.com; q=dns/txt; s=Intel; t=1663827505; x=1695363505; h=from:to:cc:subject:date:message-id:in-reply-to: references:mime-version:content-transfer-encoding; bh=FFu1zG70gDTuqoCXXC/h09KCMyxMgr+Xaf2enIHejGE=; b=BMQpm74AJqS9VDKyp4j/ZuCxOQ3s5dK6qf4PqCVAuLw7lgyGMlOfQ+mf ilQLkFRCzeCg356QwhvlMD73Q6BDhv6JoJiHZ3aozSjMNhITWHAohXjbx DzAotO7PK2o2QXvCtq8Cpe9jD4QaAuAIfqqn+vJBoFIgCKyxEu+ptznuT bHzeg3H+1e/PJmqj7LipEK42FpJAdwiD7ps0MY4xdVxdEP4N1HcsOnb1x XUwscH8Tf0mWZ6BAeto7pVSZQ6FPBnAFwsFWOlNCwqGu4hhaz1oaDtV4G Y/2z34CfaKKzSiBqJ5mxpu7osWChP8qJ5dpFxKZkkVXb4l+RP6BYhOakY A==; X-IronPort-AV: E=McAfee;i="6500,9779,10477"; a="301617773" X-IronPort-AV: E=Sophos;i="5.93,335,1654585200"; d="scan'208";a="301617773" Received: from fmsmga005.fm.intel.com ([10.253.24.32]) by orsmga103.jf.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 21 Sep 2022 23:18:24 -0700 X-IronPort-AV: E=Sophos;i="5.93,335,1654585200"; d="scan'208";a="948466152" Received: from unknown (HELO localhost.localdomain) ([10.239.252.92]) by fmsmga005-auth.fm.intel.com with ESMTP/TLS/ECDHE-RSA-AES256-GCM-SHA384; 21 Sep 2022 23:17:47 -0700 From: Hongbo Li To: dts@dpdk.org Cc: Hongbo Li Subject: [dts][PATCH V1 4/7] tests/tso:Separated performance cases Date: Thu, 22 Sep 2022 14:29:47 +0000 Message-Id: <20220922142950.398902-4-hongbox.li@intel.com> X-Mailer: git-send-email 2.25.1 In-Reply-To: <20220922142950.398902-1-hongbox.li@intel.com> References: <20220922142950.398902-1-hongbox.li@intel.com> 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 Separated performance cases Signed-off-by: Hongbo Li --- test_plans/perf_tso_test_plan.rst | 82 +++++++++++ test_plans/tso_test_plan.rst | 35 ----- tests/TestSuite_perf_tso.py | 221 ++++++++++++++++++++++++++++++ tests/TestSuite_tso.py | 128 ----------------- 4 files changed, 303 insertions(+), 163 deletions(-) create mode 100644 test_plans/perf_tso_test_plan.rst create mode 100644 tests/TestSuite_perf_tso.py diff --git a/test_plans/perf_tso_test_plan.rst b/test_plans/perf_tso_test_plan.rst new file mode 100644 index 00000000..9dcdd4c3 --- /dev/null +++ b/test_plans/perf_tso_test_plan.rst @@ -0,0 +1,82 @@ +.. SPDX-License-Identifier: BSD-3-Clause + Copyright(c) 2015-2017 Intel Corporation + +========================================= +Transmit Segmentation Offload (TSO) Tests +========================================= + +Description +=========== + +This document provides the plan for testing the TSO (Transmit Segmentation +Offload, also called Large Send offload - LSO) feature of +Intel Ethernet Controller, including Intel 82599 10GbE Ethernet Controller and +Intel® Ethernet Converged Network Adapter XL710-QDA2. TSO enables the TCP/IP stack to +pass the network device a larger ULP datagram than the Maximum Transmit +Unit Size (MTU). NIC divides the large ULP datagram to multiple segments +according to the MTU size. + + +Prerequisites +============= + +The DUT must take one of the Ethernet controller ports connected to a port on another +device that is controlled by the Scapy packet generator. + +The Ethernet interface identifier of the port that Scapy will use must be known. +On tester, all offload feature should be disabled on tx port, and start rx port capture:: + + ethtool -K rx off tx off tso off gso off gro off lro off + ip l set up + tcpdump -n -e -i -s 0 -w /tmp/cap + + +On DUT, run pmd with parameter "--enable-rx-cksum". Then enable TSO on tx port +and checksum on rx port. The test commands is below:: + + #enable hw checksum on rx port + tx_checksum set ip hw 0 + tx_checksum set udp hw 0 + tx_checksum set tcp hw 0 + tx_checksum set sctp hw 0 + set fwd csum + + # enable TSO on tx port + *tso set 800 1 + + + +Test case: TSO performance +========================== + +Set the packet stream as below before testing + ++-------+---------+---------+---------+----------+----------+ +| Frame | 1S/1C/1T| 1S/1C/1T| 1S/2C/1T| 1S/2C/2T | 1S/2C/2T | +| Size | | | | | | ++-------+---------+---------+---------+----------+----------+ +| 64 | | | | | | ++-------+---------+---------+---------+----------+----------+ +| 65 | | | | | | ++-------+---------+---------+---------+----------+----------+ +| 128 | | | | | | ++-------+---------+---------+---------+----------+----------+ +| 256 | | | | | | ++-------+---------+---------+---------+----------+----------+ +| 512 | | | | | | ++-------+---------+---------+---------+----------+----------+ +| 1024 | | | | | | ++-------+---------+---------+---------+----------+----------+ +| 1280 | | | | | | ++-------+---------+---------+---------+----------+----------+ +| 1518 | | | | | | ++-------+---------+---------+---------+----------+----------+ + +Then run the test application as below:: + + ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -c 0xffffffff -n 2 -- -i --rxd=512 --txd=512 + --burst=32 --rxfreet=64 --mbcache=128 --portmask=0x3 --txpt=36 --txht=0 --txwt=0 + --txfreet=32 --txrst=32 --enable-rx-cksum + +The -n command is used to select the number of memory channels. It should match the +number of memory channels on that setup. diff --git a/test_plans/tso_test_plan.rst b/test_plans/tso_test_plan.rst index d0f96b2b..7b7077b7 100644 --- a/test_plans/tso_test_plan.rst +++ b/test_plans/tso_test_plan.rst @@ -131,38 +131,3 @@ Test nvgre() in scapy:: sendp([Ether(dst="%s",src="52:00:00:00:00:00")/IP(src="192.168.1.1",dst="192.168.1.2",proto=47)/NVGRE()/Ether(dst=%s,src="52:00:00:00:00:00")/IP(src="192.168.1.1",dst="192.168.1.2")/TCP(sport="1021",dport="1021")/("X"*%s)], iface="%s") -Test case: TSO performance -========================== - -Set the packet stream to be sent out from packet generator before testing as -below. - -+-------+---------+---------+---------+----------+----------+ -| Frame | 1S/1C/1T| 1S/1C/1T| 1S/2C/1T| 1S/2C/2T | 1S/2C/2T | -| Size | | | | | | -+-------+---------+---------+---------+----------+----------+ -| 64 | | | | | | -+-------+---------+---------+---------+----------+----------+ -| 65 | | | | | | -+-------+---------+---------+---------+----------+----------+ -| 128 | | | | | | -+-------+---------+---------+---------+----------+----------+ -| 256 | | | | | | -+-------+---------+---------+---------+----------+----------+ -| 512 | | | | | | -+-------+---------+---------+---------+----------+----------+ -| 1024 | | | | | | -+-------+---------+---------+---------+----------+----------+ -| 1280 | | | | | | -+-------+---------+---------+---------+----------+----------+ -| 1518 | | | | | | -+-------+---------+---------+---------+----------+----------+ - -Then run the test application as below:: - - ./x86_64-native-linuxapp-gcc/app/dpdk-testpmd -c 0xffffffff -n 2 -- -i --rxd=512 --txd=512 - --burst=32 --rxfreet=64 --mbcache=128 --portmask=0x3 --txpt=36 --txht=0 --txwt=0 - --txfreet=32 --txrst=32 --enable-rx-cksum - -The -n command is used to select the number of memory channels. It should match the -number of memory channels on that setup. diff --git a/tests/TestSuite_perf_tso.py b/tests/TestSuite_perf_tso.py new file mode 100644 index 00000000..3821ecd1 --- /dev/null +++ b/tests/TestSuite_perf_tso.py @@ -0,0 +1,221 @@ +import os +import time + +import framework.utils as utils +from framework.pktgen import PacketGeneratorHelper +from framework.settings import HEADER_SIZE +from framework.test_case import TestCase + +DEFAULT_MUT = 1500 +TSO_MTU = 9000 + + +class TestTSO(TestCase): + dut_ports = [] + + def set_up_all(self): + """ + Run at the start of each test suite. + """ + # Based on h/w type, choose how many ports to use + self.dut_ports = self.dut.get_ports(self.nic) + + # Verify that enough ports are available + self.verify(len(self.dut_ports) >= 2, "Insufficient ports for testing") + + # Verify that enough threads are available + self.all_cores_mask = utils.create_mask(self.dut.get_core_list("all")) + self.portMask = utils.create_mask([self.dut_ports[0], self.dut_ports[1]]) + self.ports_socket = self.dut.get_numa_id(self.dut_ports[0]) + + self.loading_sizes = [128, 800, 801, 1700, 2500] + self.rxfreet_values = [0, 8, 16, 32, 64, 128] + + self.test_cycles = [{"cores": "", "Mpps": {}, "pct": {}}] + + self.table_header = ["Frame Size"] + for test_cycle in self.test_cycles: + self.table_header.append("%s Mpps" % test_cycle["cores"]) + self.table_header.append("% linerate") + + self.eal_param = self.dut.create_eal_parameters( + cores="1S/2C/1T", socket=self.ports_socket, ports=self.dut_ports + ) + + self.headers_size = HEADER_SIZE["eth"] + HEADER_SIZE["ip"] + HEADER_SIZE["tcp"] + self.tester.send_expect( + "ifconfig %s mtu %s" + % ( + self.tester.get_interface( + self.tester.get_local_port(self.dut_ports[0]) + ), + TSO_MTU, + ), + "# ", + ) + # get dts output path + if self.logger.log_path.startswith(os.sep): + self.output_path = self.logger.log_path + else: + cur_path = os.path.dirname(os.path.dirname(os.path.realpath(__file__))) + 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.path = self.dut.apps_name["test-pmd"] + + def set_up(self): + """ + Run before each test case. + """ + pass + + def test_perf_TSO_2ports(self): + """ + TSO Performance Benchmarking with 2 ports. + """ + + # prepare traffic generator input + tgen_input = [] + + # run testpmd for each core config + for test_cycle in self.test_cycles: + core_config = test_cycle["cores"] + cores = self.dut.get_core_list(core_config, socket=self.ports_socket) + self.coreMask = utils.create_mask(cores) + if len(cores) > 2: + queues = len(cores) // 2 + else: + queues = 1 + + command_line = ( + "%s %s -- -i --rxd=512 --txd=512 --burst=32 --rxfreet=64 --mbcache=128 --portmask=%s --max-pkt-len=%s --txpt=36 --txht=0 --txwt=0 --txfreet=32 --txrst=32 " + % (self.path, self.eal_param, self.portMask, TSO_MTU) + ) + info = "Executing PMD using %s\n" % test_cycle["cores"] + self.logger.info(info) + self.rst_report(info, annex=True) + self.rst_report(command_line + "\n\n", frame=True, annex=True) + + self.dut.send_expect(command_line, "testpmd> ", 120) + self.dut.send_expect("port stop all", "testpmd> ", 120) + self.dut.send_expect( + "csum set ip hw %d" % self.dut_ports[0], "testpmd> ", 120 + ) + self.dut.send_expect( + "csum set udp hw %d" % self.dut_ports[0], "testpmd> ", 120 + ) + self.dut.send_expect( + "csum set tcp hw %d" % self.dut_ports[0], "testpmd> ", 120 + ) + self.dut.send_expect( + "csum set sctp hw %d" % self.dut_ports[0], "testpmd> ", 120 + ) + self.dut.send_expect( + "csum set outer-ip hw %d" % self.dut_ports[0], "testpmd> ", 120 + ) + self.dut.send_expect( + "csum parse-tunnel on %d" % self.dut_ports[0], "testpmd> ", 120 + ) + self.dut.send_expect( + "csum set ip hw %d" % self.dut_ports[1], "testpmd> ", 120 + ) + self.dut.send_expect( + "csum set udp hw %d" % self.dut_ports[1], "testpmd> ", 120 + ) + self.dut.send_expect( + "csum set tcp hw %d" % self.dut_ports[1], "testpmd> ", 120 + ) + self.dut.send_expect( + "csum set sctp hw %d" % self.dut_ports[1], "testpmd> ", 120 + ) + self.dut.send_expect( + "csum set outer-ip hw %d" % self.dut_ports[1], "testpmd> ", 120 + ) + self.dut.send_expect( + "csum parse-tunnel on %d" % self.dut_ports[1], "testpmd> ", 120 + ) + self.dut.send_expect("tso set 800 %d" % self.dut_ports[1], "testpmd> ", 120) + self.dut.send_expect("set fwd csum", "testpmd> ", 120) + self.dut.send_expect("port start all", "testpmd> ", 120) + self.dut.send_expect("set promisc all off", "testpmd> ", 120) + self.dut.send_expect("start", "testpmd> ") + for loading_size in self.loading_sizes: + frame_size = loading_size + self.headers_size + wirespeed = self.wirespeed(self.nic, frame_size, 2) + + # create pcap file + self.logger.info("Running with frame size %d " % frame_size) + payload_size = frame_size - self.headers_size + for _port in range(2): + mac = self.dut.get_mac_address(self.dut_ports[_port]) + + pcap = os.sep.join([self.output_path, "dts{0}.pcap".format(_port)]) + self.tester.scapy_append( + 'wrpcap("%s", [Ether(dst="%s",src="52:00:00:00:00:01")/IP(src="192.168.1.1",dst="192.168.1.2")/TCP(sport=1021,dport=1021)/("X"*%d)])' + % (pcap, mac, payload_size) + ) + tgen_input.append( + ( + self.tester.get_local_port(self.dut_ports[_port]), + self.tester.get_local_port(self.dut_ports[1 - _port]), + "%s" % pcap, + ) + ) + self.tester.scapy_execute() + + # clear streams before add new streams + self.tester.pktgen.clear_streams() + # run packet generator + streams = self.pktgen_helper.prepare_stream_from_tginput( + tgen_input, 100, None, self.tester.pktgen + ) + _, pps = self.tester.pktgen.measure_throughput(stream_ids=streams) + + pps /= 1000000.0 + test_cycle["Mpps"][loading_size] = pps + test_cycle["pct"][loading_size] = pps * 100 // wirespeed + + self.dut.send_expect("stop", "testpmd> ") + self.dut.send_expect("quit", "# ", 30) + time.sleep(5) + + for n in range(len(self.test_cycles)): + for loading_size in self.loading_sizes: + self.verify( + self.test_cycles[n]["Mpps"][loading_size] > 0, "No traffic detected" + ) + + # Print results + self.result_table_create(self.table_header) + for loading_size in self.loading_sizes: + table_row = [loading_size] + for test_cycle in self.test_cycles: + table_row.append(test_cycle["Mpps"][loading_size]) + table_row.append(test_cycle["pct"][loading_size]) + + self.result_table_add(table_row) + + self.result_table_print() + + def tear_down(self): + """ + Run after each test case. + """ + self.dut.send_expect("quit", "# ") + self.dut.kill_all() + time.sleep(2) + + def tear_down_all(self): + """ + Run after each test suite. + """ + self.tester.send_expect( + "ifconfig %s mtu %s" + % ( + self.tester.get_interface( + self.tester.get_local_port(self.dut_ports[0]) + ), + DEFAULT_MUT, + ), + "# ", + ) diff --git a/tests/TestSuite_tso.py b/tests/TestSuite_tso.py index 49db011d..c4f25afc 100644 --- a/tests/TestSuite_tso.py +++ b/tests/TestSuite_tso.py @@ -487,134 +487,6 @@ class TestTSO(TestCase): ) self.get_chksum_value_and_verify(dump_pcap, save_file, Nic_list) - def test_perf_TSO_2ports(self): - """ - TSO Performance Benchmarking with 2 ports. - """ - - # prepare traffic generator input - tgen_input = [] - - # run testpmd for each core config - for test_cycle in self.test_cycles: - core_config = test_cycle["cores"] - cores = self.dut.get_core_list(core_config, socket=self.ports_socket) - self.coreMask = utils.create_mask(cores) - if len(cores) > 2: - queues = len(cores) // 2 - else: - queues = 1 - - command_line = ( - "%s %s -- -i --rxd=512 --txd=512 --burst=32 --rxfreet=64 --mbcache=128 --portmask=%s --max-pkt-len=%s --txpt=36 --txht=0 --txwt=0 --txfreet=32 --txrst=32 " - % (self.path, self.eal_param, self.portMask, TSO_MTU) - ) - info = "Executing PMD using %s\n" % test_cycle["cores"] - self.logger.info(info) - self.rst_report(info, annex=True) - self.rst_report(command_line + "\n\n", frame=True, annex=True) - - self.dut.send_expect(command_line, "testpmd> ", 120) - self.dut.send_expect("port stop all", "testpmd> ", 120) - self.dut.send_expect( - "csum set ip hw %d" % self.dut_ports[0], "testpmd> ", 120 - ) - self.dut.send_expect( - "csum set udp hw %d" % self.dut_ports[0], "testpmd> ", 120 - ) - self.dut.send_expect( - "csum set tcp hw %d" % self.dut_ports[0], "testpmd> ", 120 - ) - self.dut.send_expect( - "csum set sctp hw %d" % self.dut_ports[0], "testpmd> ", 120 - ) - self.dut.send_expect( - "csum set outer-ip hw %d" % self.dut_ports[0], "testpmd> ", 120 - ) - self.dut.send_expect( - "csum parse-tunnel on %d" % self.dut_ports[0], "testpmd> ", 120 - ) - self.dut.send_expect( - "csum set ip hw %d" % self.dut_ports[1], "testpmd> ", 120 - ) - self.dut.send_expect( - "csum set udp hw %d" % self.dut_ports[1], "testpmd> ", 120 - ) - self.dut.send_expect( - "csum set tcp hw %d" % self.dut_ports[1], "testpmd> ", 120 - ) - self.dut.send_expect( - "csum set sctp hw %d" % self.dut_ports[1], "testpmd> ", 120 - ) - self.dut.send_expect( - "csum set outer-ip hw %d" % self.dut_ports[1], "testpmd> ", 120 - ) - self.dut.send_expect( - "csum parse-tunnel on %d" % self.dut_ports[1], "testpmd> ", 120 - ) - self.dut.send_expect("tso set 800 %d" % self.dut_ports[1], "testpmd> ", 120) - self.dut.send_expect("set fwd csum", "testpmd> ", 120) - self.dut.send_expect("port start all", "testpmd> ", 120) - self.dut.send_expect("set promisc all off", "testpmd> ", 120) - self.dut.send_expect("start", "testpmd> ") - for loading_size in self.loading_sizes: - frame_size = loading_size + self.headers_size - wirespeed = self.wirespeed(self.nic, frame_size, 2) - - # create pcap file - self.logger.info("Running with frame size %d " % frame_size) - payload_size = frame_size - self.headers_size - for _port in range(2): - mac = self.dut.get_mac_address(self.dut_ports[_port]) - - pcap = os.sep.join([self.output_path, "dts{0}.pcap".format(_port)]) - self.tester.scapy_append( - 'wrpcap("%s", [Ether(dst="%s",src="52:00:00:00:00:01")/IP(src="192.168.1.1",dst="192.168.1.2")/TCP(sport=1021,dport=1021)/("X"*%d)])' - % (pcap, mac, payload_size) - ) - tgen_input.append( - ( - self.tester.get_local_port(self.dut_ports[_port]), - self.tester.get_local_port(self.dut_ports[1 - _port]), - "%s" % pcap, - ) - ) - self.tester.scapy_execute() - - # clear streams before add new streams - self.tester.pktgen.clear_streams() - # run packet generator - streams = self.pktgen_helper.prepare_stream_from_tginput( - tgen_input, 100, None, self.tester.pktgen - ) - _, pps = self.tester.pktgen.measure_throughput(stream_ids=streams) - - pps /= 1000000.0 - test_cycle["Mpps"][loading_size] = pps - test_cycle["pct"][loading_size] = pps * 100 // wirespeed - - self.dut.send_expect("stop", "testpmd> ") - self.dut.send_expect("quit", "# ", 30) - time.sleep(5) - - for n in range(len(self.test_cycles)): - for loading_size in self.loading_sizes: - self.verify( - self.test_cycles[n]["Mpps"][loading_size] > 0, "No traffic detected" - ) - - # Print results - self.result_table_create(self.table_header) - for loading_size in self.loading_sizes: - table_row = [loading_size] - for test_cycle in self.test_cycles: - table_row.append(test_cycle["Mpps"][loading_size]) - table_row.append(test_cycle["pct"][loading_size]) - - self.result_table_add(table_row) - - self.result_table_print() - def tear_down(self): """ Run after each test case.