[dts] [PATCH V1 1/2] Packet capture: upload automation test plan

[Mo, YufengX]

hi, Marvin



see below.



> -----Original Message-----

> From: Liu, Yong

> Sent: January 16, 2017 11:08 AM

> To: Mo, YufengX; dts at dpdk.org

> Cc: Mo, YufengX

> Subject: RE: [dts] [PATCH V1 1/2] Packet capture: upload automation test plan

>

> Yufen,

> Please make sure all lines in test plan and commit are limited to 79 characters.

> And some comments and questions below.

>

> Thanks,

> Marvin

>

> > -----Original Message-----

> > From: dts [mailto:dts-bounces at dpdk.org] On Behalf Of yufengx.mo at intel.com

> > Sent: Monday, January 16, 2017 10:49 AM

> > To: dts at dpdk.org

> > Cc: Mo, YufengX <yufengx.mo at intel.com>

> > Subject: [dts] [PATCH V1 1/2] Packet capture: upload automation test plan

> >

> > From: yufengmx <yufengx.mo at intel.com>

> >

> >

> > Packet capture framework feature support packet capturing on dpdk ethernet

> > devices.

> > DPDK provides dpdk-pdump tool under app/pdump directory for packet

> > capturing on dpdk.

> >

> > The  tool is a Data Plane Development Kit (DPDK) tool that runs as a DPDK

> > secondary process and is capable of enabling packet capture on dpdk ports.

> >   * The  tool depends on libpcap based PMD which is disabled

> >     by default in the build configuration files,

> >     owing to an external dependency on the libpcap development files

> >     which must be installed on the board.

> >     Once the libpcap development files are installed, the libpcap based PMD

> >     can be enabled by setting CONFIG_RTE_LIBRTE_PMD_PCAP=y and

> > recompiling the DPDK.

> >

> > Signed-off-by: yufengmx <yufengx.mo at intel.com>

> > ---

> >  test_plans/packet_capture_test_plan.rst | 376

> > ++++++++++++++++++++++++++++++++

> >  1 file changed, 376 insertions(+)

> >  create mode 100644 test_plans/packet_capture_test_plan.rst

> >

> > diff --git a/test_plans/packet_capture_test_plan.rst

> > b/test_plans/packet_capture_test_plan.rst

> > new file mode 100644

> > index 0000000..6a51cf9

> > --- /dev/null

> > +++ b/test_plans/packet_capture_test_plan.rst

> > @@ -0,0 +1,376 @@

> > +.. Copyright (c) <2016> 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.

> > +

> > +==================

> > +  packet capture framework feature

> > +==================

> > +

> > +Packet capture framework feature support packet capturing on dpdk ethernet

> > devices.

> > +DPDK provides dpdk-pdump tool under app/pdump directory for packet

> > capturing on dpdk.

> > +

> > +The dpdk-pdump application will act as the secondary process. The EAL

> > +thread polls for packet capture fd. If fd polled matches packet capture

> > +fd, it will initiate packet capture processing.

> > +

> > +The testpmd application will act as the primary process. The primary

> > +process create socket for packet capture connection with the secondary

> > +process and registers socket with packet capture epoll event. Packet capture

> > event will be polled as part of interrupt thread.

> > +

> > +The primary process creates mempool and two rte_rings for packets

> > +duplication and sharing packet info with the secondary process respectively.

> > +

> > +Upon receiving packet capture event, the primary process receive either

> > +register RX/TX callbacks or remove RX/TX callbacks message from the

> > +secondary process on socket.If packet matches, reference count of

> > +packet will be incremented and enqueued to second rte_ring for the secondary

> > process to use.

> > +

> > +Prerequisites

> > +=============

> > +* 2x NICs (2 full duplex ports per NIC) plugged into the available

> > +slots on a platform, another two nic ports are linked with cables.

> > +

> > +Test cases

> > +=============

> > +The testpmd application act as server process with port-topology

> > +chained mode, the dpdk-pdump act as client process to dump capture

> > +packet with different options setting. Select one port of tester as tx

> > +port, another port of tester as rx port, send different type packets

> > +from two ports, check pcap files' content dumped by scapy and tcpdump to

> > confirm testpmd working correctly, check pcap files' content dumped by

> > tcpdump and dpdk-pdump to confirm dpdk-pdump working correctly.

> > +

> > +dpdk-pdump command format

> > +--------------------------

> > +* packet capture framework tool dpdk-pdump command format, parameters

> > +inside the parenthesis represents the mandatory parametersm parameters

>

> Typo issue, "parametersm" should be ?

>

> > +inside the square brackets represents optional

> > +parameters::

> > +./x86_64-native-linuxapp-gcc/app/dpdk-pdump -- --pdump= '(port =

> > +<port_id> |device_id = <pci address>), (queue=<queue number>),

> > +(rx-dev=<iface/path to pcap file> | tx-dev=<iface/path to pcap file>),

> > +[ring-size=<size>], [mbuf-size=<size>], [total-num-mbufs=<size>]'

> > +

> > +transmission packets' format

> > +--------------------------

> > +* IP_RAW

> > +  [Ether()/IP()/Raw('\0'*60)]

> > +* TCP

> > +  [Ether()/IP()/TCP()/Raw('\0'*60)]

> > +* UDP

> > +  [Ether()/IP()/UDP()/Raw('\0'*60)]

> > +* SCTP

> > +  [Ether()/IP()/SCTP()/Raw('\0'*40)]

> > +* IPv6_TCP

> > +  [Ether()/IPv6()/TCP()/Raw('\0'*60)]

> > +* IPv6_UDP

> > +  [Ether()/IPv6()/UDP()/Raw('\0'*60)]

> > +* IPv6_SCTP

> > +  [Ether()/IP()/IPv6()/SCTP()/Raw('\0'*40)]

> > +* VLAN_UDP

> > +  [Ether()/Dot1Q()/IP()/UDP()/Raw('\0'*40)]

> > +* TIMESYNC

> > +  [Ether(dst='FF:FF:FF:FF:FF:FF',type=0x88f7)/"\\x00\\x02"]

> > +* ARP

> > +  [Ether(dst='FF:FF:FF:FF:FF:FF')/ARP()]

> > +* LLDP

> > +  [Ether()/LLDP()/LLDPManagementAddress()]

> > +  notice: LLDP()/LLDPManagementAddress() method are in dts/dep/lldp.py

> > +

> > +port configuration

> > +--------------------------

> > +* confirm two NICs physical link on a platform:

> > +  dut port 0 <---> tester port 0

> > +  dut port 1 <---> tester port 1

> > +

> > +* Bind two port on DUT::

> > + ./tools/dpdk_nic_bind.py --bind=igb_uio <dut port 0 pci address> <dut

> > +port 1 pci address>

> > +

> Tools folder has been changed to usertools.

>

> > +* On dut, use port 0 as rx/tx port.

> > +  If dut port 0 rx dump is set, scapy send packet from tester port 0

> > +and tcpdump dumps tester

> > +  port 1's packet.

> > +

> > +  If dut port 0 tx dump is set, scapy send packet from tester port 1

> > + and tcpdump dumps tester  port 0's packet.

> > +

> > +* If using interfaces as dpdk-pdump vdev, prepare two ports on DUT,

> > +which haven't been binded

> > +  to dpdk and have been in linked status

> > +

> > +Test Case 1: test pdump port

> > +=================================

> > +test different port options::

> > +* port=<dut port id>

> > +* device_id=<dut pci address>

> > +

> > +Boot up dpdk's testpmd with chained option::

> > +      ./x86_64-native-linuxapp-gcc/app/testpmd -c 0x6 -n 4 -- -i --port-

> > topology=chained

> > +      testpmd> set fwd io

> > +      testpmd> start

> > +

> Recommend to remove targets which will be change between executions.

>

> > +When test VLAN_UDP type packet transmission, set vlan:

> > +    testpmd> vlan set filter off 1

> > +    testpmd> start

> > +

> > +Boot up dpdk-pdump::

> > +    ./x86_64-native-linuxapp-gcc/app/dpdk-pdump -- --pdump  '<port

> > option>,queue=*,\

> > +    tx-dev=/tmp/pdump-tx.pcap,rx-dev=/tmp/pdump-rx.pcap'

> > +

> > +Set up linux's tcpdump to receiver packet on tester::

> > +      tcpdump -i <rx port name> -w /tmp/sniff-<rx port name>.pcap

> > +      tcpdump -i <tx port name> -w /tmp/sniff-<tx port name>.pcap

> > +

> > +Send packet on tester by port 0::

> > +    sendp(<packet format>, iface=<port 0 name>)

> > +

> > +Compare pcap file of scapy with the pcap file dumped by tcpdump.

> > +Compare pcap file dumped by dpdk-pdump with pcap files dumped by tcpdump.

> > +

> > +Send packet on tester by port 1::

> > +    sendp(<packet format>, iface=<port 1 name>)

> > +

> > +Compare pcap file of scapy with the pcap file dumped by tcpdump.

> > +Compare pcap file dumped by dpdk-pdump with pcap files dumped by tcpdump.

> > +

> > +Test Case 2: test pdump queue

> > +=================================

> > +test different queue options::

> > +* first queue:

> > +  queue=0

> > +* all:

> > +  queue=*

>

> Please add some basic description of this parameter. Like "Capture first queue traffic".

> > +

> > +Boot up dpdk's testpmd with chained option::

> > +      ./x86_64-native-linuxapp-gcc/app/testpmd -c 0x6 -n 4 -- -i --port-

> > topology=chained

> > +      testpmd> set fwd io

> > +      testpmd> start

> > +

> > +When test VLAN_UDP type packet transmission, set vlan:

> > +    testpmd> vlan set filter off 1

> > +    testpmd> start

> > +

> > +Boot up dpdk-pdump::

> > +    ./x86_64-native-linuxapp-gcc/app/dpdk-pdump -- --pdump  'port=0,<queue

> > option>,\

> > +    tx-dev=/tmp/pdump-tx.pcap,rx-dev=/tmp/pdump-rx.pcap'

> > +

> > +Set up linux's tcpdump to receiver packet on tester::

> > +      tcpdump -i <rx port name> -w /tmp/sniff-<rx port name>.pcap

> > +      tcpdump -i <tx port name> -w /tmp/sniff-<tx port name>.pcap

> > +

> > +Send packet on tester by port 0::

> > +    sendp(<packet format>, iface=<port 0 name>)

> > +

> > +Compare pcap file of scapy with the pcap file dumped by tcpdump.

> > +Compare pcap file dumped by dpdk-pdump with pcap files dumped by tcpdump.

> > +

> > +Send packet on tester by port 1::

> > +    sendp(<packet format>, iface=<port 1 name>)

> > +

> > +Compare pcap file of scapy with the pcap file dumped by tcpdump.

> > +Compare pcap file dumped by dpdk-pdump with pcap files dumped by tcpdump.

> > +

> > +Test Case 3: test pdump dev pcap

> > +=================================

>

> Underscore please align with test case name.

>

> > +test different dump options::

> > +* tx-dev=/tmp/pdump-tx.pcap,rx-dev=/tmp/pdump-rx.pcap

> > +* rx-dev=/tmp/pdump-rx.pcap

> > +* tx-dev=/tmp/pdump-tx.pcap

>

> Same as previous comment, add basic descriptions for these parameters.

>

> > +

> > +Boot up dpdk's testpmd with chained option::

> > +      ./x86_64-native-linuxapp-gcc/app/testpmd -c 0x6 -n 4 -- -i --port-

> > topology=chained

> > +      testpmd> set fwd io

> > +      testpmd> start

> > +

> > +When test VLAN_UDP type packet transmission, set vlan:

> > +    testpmd> vlan set filter off 1

> > +    testpmd> start

> > +

> > +Boot up dpdk-pdump with pdump options::

> > +    ./x86_64-native-linuxapp-gcc/app/dpdk-pdump -- --pdump

> > 'port=0,queue=*,<dump object>'

> > +

> > +Set up linux's tcpdump to receiver packet on tester::

> > +      tcpdump -i <rx port name> -w /tmp/sniff-<rx port name>.pcap

> > +      tcpdump -i <tx port name> -w /tmp/sniff-<tx port name>.pcap

> > +

> > +Send packet on tester by port 0::

> > +    sendp(<packet format>, iface=<port 0 name>)

> > +

> > +Compare pcap file of scapy with the pcap file dumped by tcpdump.

> > +Compare pcap file dumped by dpdk-pdump with pcap files dumped by

> > tcpdump(ignore when only set tx-dev).

> > +

> > +Send packet on tester by port 1::

> > +    sendp(<packet format>, iface=<port 1 name>)

> > +

> > +Compare pcap file of scapy with the pcap file dumped by tcpdump.

> > +Compare pcap file dumped by dpdk-pdump with pcap files dumped by

> > tcpdump(ignore when only set rx-dev).

> > +

> > +Test Case 4: test pdump dev iface

> > +=================================

> > +test different dump options::

> > +* tx-dev=<dut tx port name>,rx-dev=<dut rx port name>

> > +* rx-dev=<dut rx port name>

> > +* tx-dev=<dut tx port name>

> > +

> > +Boot up dpdk's testpmd with chained option::

> > +      ./x86_64-native-linuxapp-gcc/app/testpmd -c 0x6 -n 4 -- -i --port-

> > topology=chained

> > +      testpmd> set fwd io

> > +      testpmd> start

> > +

> > +When test VLAN_UDP type packet transmission, set vlan:

> > +    testpmd> vlan set filter off 1

> > +    testpmd> start

> > +

> > +Boot up dpdk-pdump with pdump options::

> > +    ./x86_64-native-linuxapp-gcc/app/dpdk-pdump -- --pdump

> > 'port=0,queue=*,<dump object>'

> > +

> > +Set up linux's tcpdump to receiver packet on tester::

> > +      tcpdump -i <rx port name> -w /tmp/sniff-<rx port name>.pcap

> > +      tcpdump -i <tx port name> -w /tmp/sniff-<tx port name>.pcap

> > +

> > +Set up linux's tcpdump to receiver packet of dpdk-pdump on Dut::

> > +      when rx-dev is set

> > +      tcpdump -i <dut rx port name> -w /tmp/pdump-rx.pcap

> > +

> > +      when tx-dev is set

> > +      tcpdump -i <dut tx port name> -w /tmp/pdump-tx.pcap

> > +

> > +Send packet on tester by port 0::

> > +    sendp(<packet format>, iface=<port 0 name>)

> > +

> > +Compare pcap file of scapy with the pcap file dumped by tcpdump.

> > +Compare pcap file dumped by dpdk-pdump with pcap files dumped by

> > tcpdump(ignore when only set tx-dev).

> > +

> > +Send packet on tester by port 1::

> > +    sendp(<packet format>, iface=<port 1 name>)

> > +

> > +Compare pcap file of scapy with the pcap file dumped by tcpdump.

> > +Compare pcap file dumped by dpdk-pdump with pcap files dumped by

> > tcpdump(ignore when only set rx-dev).

> > +

> > +Test Case 5: test pdump ring size

> > +=================================

> > +test ring size option, set value within 2^[1~27]

> > +

> > +Boot up dpdk's testpmd with chained option::

> > +      ./x86_64-native-linuxapp-gcc/app/testpmd -c 0x6 -n 4 -- -i --port-

> > topology=chained

> > +      testpmd> set fwd io

> > +      testpmd> start

> > +

> > +When test VLAN_UDP type packet transmission, set vlan:

> > +    testpmd> vlan set filter off 1

> > +    testpmd> start

> > +

> > +Boot up dpdk-pdump with pdump options::

> > +    ./x86_64-native-linuxapp-gcc/app/dpdk-pdump -- --pdump

> > 'port=0,queue=*,\

> > +    tx-dev=/tmp/pdump-tx.pcap,rx-dev=/tmp/pdump-rx.pcap,ring-size=1024'

> > +

> > +Set up linux's tcpdump to receiver packet on tester::

> > +      tcpdump -i <rx port name> -w /tmp/sniff-<rx port name>.pcap

> > +      tcpdump -i <tx port name> -w /tmp/sniff-<tx port name>.pcap

> > +

> > +Send packet on tester by port 0::

> > +    sendp(<packet format>, iface=<port 0 name>)

> > +

> > +Compare pcap file of scapy with the pcap file dumped by tcpdump.

> > +Compare pcap file dumped by dpdk-pdump with pcap files dumped by tcpdump.

> > +

> > +Send packet on tester by port 1::

> > +    sendp(<packet format>, iface=<port 1 name>)

> > +

> > +Compare pcap file of scapy with the pcap file dumped by tcpdump.

> > +Compare pcap file dumped by dpdk-pdump with pcap files dumped by tcpdump.

> > +

> > +Test Case 6: test pdump mbuf size

> > +=================================

> > +test mbuf size option, set value within [252~59520]. min value is

> > +decided by single packet size, max value is decided by test platform memery

> > size.

>

> Typo issue, should be "memory".

> Should here to verify what will be happened when packet size is over mbuf size?

[Mo, YufengX] when size is beyon the range, dpdk-pdump won't boot up. Do you want to add a verification.

will it do the same verification for other range checking?

> Suggest add one case for jumbo frame packet capture, not sure whether jumbo frame should be validated here.

[Mo, YufengX]  Do you mean to write a new case for jumbo frame?

>

> > +

> > +

> > +Boot up dpdk's testpmd with chained option::



> > +      ./x86_64-native-linuxapp-gcc/app/testpmd -c 0x6 -n 4 -- -i --port-

> > topology=chained

> > +      testpmd> set fwd io

> > +      testpmd> start

> > +

> > +When test VLAN_UDP type packet transmission, set vlan:

> > +    testpmd> vlan set filter off 1

> > +    testpmd> start

> > +

> > +Boot up dpdk-pdump with pdump options::

> > +    ./x86_64-native-linuxapp-gcc/app/dpdk-pdump -- --pdump

> > 'port=0,queue=*,\

> > +    tx-dev=/tmp/pdump-tx.pcap,rx-dev=/tmp/pdump-rx.pcap,mbuf-size=2048'

> > +

> > +Set up linux's tcpdump to receiver packet on tester::

> > +      tcpdump -i <rx port name> -w /tmp/sniff-<rx port name>.pcap

> > +      tcpdump -i <tx port name> -w /tmp/sniff-<tx port name>.pcap

> > +

> > +Send packet on tester by port 0::

> > +    sendp(<packet format>, iface=<port 0 name>)

> > +

> > +Compare pcap file of scapy with the pcap file dumped by tcpdump.

> > +Compare pcap file dumped by dpdk-pdump with pcap files dumped by tcpdump.

> > +

> > +Send packet on tester by port 1::

> > +    sendp(<packet format>, iface=<port 1 name>)

> > +

> > +Compare pcap file of scapy with the pcap file dumped by tcpdump.

> > +Compare pcap file dumped by dpdk-pdump with pcap files dumped by tcpdump.

> > +

> > +Test Case 7: test pdump total num mbufs

> > +=================================

> > +test total-num-mbufs option, set value within [1025~65535]

>

> What will be affected when change the number?  I think mbuf will be freed after packet saved into pcap file.

> Will it affect the performance?

[Mo, YufengX] no things happen when total-num-mbufs changes. Currently, test case will reboot dpdk-pdump/testpmd

after each different option finish testing.

Within those email discussed with Reshma, she has mentioned that dpdk-pdump don't work for performance, you are are in

maillist too.

>

> > +

> > +Boot up dpdk's testpmd with chained option::

> > +      ./x86_64-native-linuxapp-gcc/app/testpmd -c 0x6 -n 4 -- -i --port-

> > topology=chained

> > +      testpmd> set fwd io

> > +      testpmd> start

> > +

> > +When test VLAN_UDP type packet transmission, set vlan:

> > +    testpmd> vlan set filter off 1

> > +    testpmd> start

> > +

> > +Boot up dpdk-pdump with pdump options::

> > +    ./x86_64-native-linuxapp-gcc/app/dpdk-pdump -- --pdump

> > 'port=0,queue=*,\

> > +    tx-dev=/tmp/pdump-tx.pcap,rx-dev=/tmp/pdump-rx.pcap,total-num-

> > mbufs=8191'

> > +

> > +Set up linux's tcpdump to receiver packet on tester::

> > +      tcpdump -i <rx port name> -w /tmp/sniff-<rx port name>.pcap

> > +      tcpdump -i <tx port name> -w /tmp/sniff-<tx port name>.pcap

> > +

> > +Send packet on tester by port 0::

> > +    sendp(<packet format>, iface=<port 0 name>)

> > +

> > +Compare pcap file of scapy with the pcap file dumped by tcpdump.

> > +Compare pcap file dumped by dpdk-pdump with pcap files dumped by tcpdump.

> > +

> > +Send packet on tester by port 1::

> > +    sendp(<packet format>, iface=<port 1 name>)

> > +

> > +Compare pcap file of scapy with the pcap file dumped by tcpdump.

> > +Compare pcap file dumped by dpdk-pdump with pcap files dumped by tcpdump.

> > +

> > --

> > 1.9.3


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://dpdk.org/ml/archives/dts/attachments/20170116/4f589d57/attachment.html>