From: Jerin Jacob <jerin.jacob@caviumnetworks.com>
To: Sergio Gonzalez Monroy <sergio.gonzalez.monroy@intel.com>
Cc: dev@dpdk.org
Subject: Re: [dpdk-dev] [PATCH] example/ipsec-secgw: ipsec security gateway
Date: Mon, 1 Feb 2016 16:56:49 +0530 [thread overview]
Message-ID: <20160201112646.GA10490@localhost.localdomain> (raw)
In-Reply-To: <56AF3CDC.5090105@intel.com>
On Mon, Feb 01, 2016 at 11:09:16AM +0000, Sergio Gonzalez Monroy wrote:
> On 31/01/2016 14:39, Jerin Jacob wrote:
> >On Fri, Jan 29, 2016 at 08:29:12PM +0000, Sergio Gonzalez Monroy wrote:
> >>Sample app implementing an IPsec Security Geteway.
> >>The main goal of this app is to show the use of cryptodev framework
> >>in a real world application.
> >>
> >>Currently only supported static IPv4 IPsec tunnels using AES-CBC
> >>and HMAC-SHA1.
> >>
> >>Also, currently not supported:
> >>- SA auto negotiation (No IKE support)
> >>- chained mbufs
> >>
> >>Signed-off-by: Sergio Gonzalez Monroy <sergio.gonzalez.monroy@intel.com>
> >Hi Sergio,
> >
> >Great effort, one of the best highly integrated sample DPDK application
> >(ethernet, acl, crypto and lpm)
> >
> >>---
> >> MAINTAINERS | 4 +
> >> doc/guides/sample_app_ug/index.rst | 1 +
> >> doc/guides/sample_app_ug/ipsec_secgw.rst | 440 +++++++++++
> >> examples/Makefile | 2 +
> >> examples/ipsec-secgw/Makefile | 58 ++
> >> examples/ipsec-secgw/esp.c | 256 +++++++
> >> examples/ipsec-secgw/esp.h | 66 ++
> >> examples/ipsec-secgw/ipip.h | 100 +++
> >> examples/ipsec-secgw/ipsec-secgw.c | 1218 ++++++++++++++++++++++++++++++
> >> examples/ipsec-secgw/ipsec.c | 138 ++++
> >> examples/ipsec-secgw/ipsec.h | 184 +++++
> >> examples/ipsec-secgw/rt.c | 131 ++++
> >> examples/ipsec-secgw/sa.c | 391 ++++++++++
> >> examples/ipsec-secgw/sp.c | 324 ++++++++
> >> 14 files changed, 3313 insertions(+)
> >> create mode 100644 doc/guides/sample_app_ug/ipsec_secgw.rst
> >> create mode 100644 examples/ipsec-secgw/Makefile
> >> create mode 100644 examples/ipsec-secgw/esp.c
> >> create mode 100644 examples/ipsec-secgw/esp.h
> >> create mode 100644 examples/ipsec-secgw/ipip.h
> >> create mode 100644 examples/ipsec-secgw/ipsec-secgw.c
> >> create mode 100644 examples/ipsec-secgw/ipsec.c
> >> create mode 100644 examples/ipsec-secgw/ipsec.h
> >> create mode 100644 examples/ipsec-secgw/rt.c
> >> create mode 100644 examples/ipsec-secgw/sa.c
> >> create mode 100644 examples/ipsec-secgw/sp.c
> >>
> >>diff --git a/MAINTAINERS b/MAINTAINERS
> >>index b90aeea..996eda6 100644
> >>--- a/MAINTAINERS
> >>+++ b/MAINTAINERS
> >>@@ -596,3 +596,7 @@ F: doc/guides/sample_app_ug/vmdq_dcb_forwarding.rst
> >> M: Pablo de Lara <pablo.de.lara.guarch@intel.com>
> >> M: Daniel Mrzyglod <danielx.t.mrzyglod@intel.com>
> >> F: examples/ptpclient/
> >>+
> >>+M: Sergio Gonzalez Monroy <sergio.gonzalez.monroy@intel.com>
> >>+F: doc/guides/sample_app_ug/ipsec-secgw.rst
> >>+F: examples/ipsec-secgw/
> >>diff --git a/doc/guides/sample_app_ug/index.rst b/doc/guides/sample_app_ug/index.rst
> >>index 8a646dd..88375d2 100644
> >>--- a/doc/guides/sample_app_ug/index.rst
> >>+++ b/doc/guides/sample_app_ug/index.rst
> >>@@ -73,6 +73,7 @@ Sample Applications User Guide
> >> proc_info
> >> ptpclient
> >> performance_thread
> >>+ ipsec_secgw
> >> **Figures**
> >>diff --git a/doc/guides/sample_app_ug/ipsec_secgw.rst b/doc/guides/sample_app_ug/ipsec_secgw.rst
> >>new file mode 100644
> >>index 0000000..7be3f7e
> >>--- /dev/null
> >>+++ b/doc/guides/sample_app_ug/ipsec_secgw.rst
> >>@@ -0,0 +1,440 @@
> >>+.. BSD LICENSE
> >>+ Copyright(c) 2010-2016 Intel Corporation. All rights reserved.
> >>+ 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.
> >>+
> >>+IPsec Security Gateway Sample Application
> >>+=========================================
> >>+
> >>+The IPsec Security Gateway application is a minimal example of real work
> >>+application using DPDK cryptodev framework.
> >>+
> >>+Overview
> >>+--------
> >>+
> >>+The application demonstrates the implementation of a minimal Security Gateway
> >>+(see Constraints bellow) using DPDK based on RFC4301, RFC4303, RFC3602 and
> >>+RFC2404.
> >>+
> >>+Note that IKE is not supported (therefore not IPsec compliant), so only manual
> >>+setting of Security Policies and Associations is allowed.
> >>+
> >>+The Security Policies (SP) are implemented as ACL rules, the Security
> >>+Associations (SA) are stored in a table and the Routing is implemented
> >>+using LPM.
> >>+
> >>+The application splits the ports between Protected and Unprotected.
> >>+Based on the port the trafic is coming from, the traffic would be consider
> >>+IPsec Inbound or Outbound.
> >>+Traffic from Unprotected ports is consider IPsec Inbound, and traffic from
> >>+Protected ports is consider IPsec Outbound.
> >>+
> >>+Path for IPsec Inbound traffic:
> >>+* Read packets from the port
> >>+* Classify packets between IPv4 and ESP.
> >>+* Inbound SA lookup for ESP packets based on their SPI
> >>+* Verification/Decryption
> >>+* Removal of ESP and outter IP header
> >>+* Inbound SP check using ACL of decrypted packets and any other IPv4 packet
> >>+we read.
> >>+* Routing
> >>+* Write packet to port
> >>+
> >>+Path for IPsec Outbound traffic:
> >>+* Read packets from the port
> >>+* Outbound SP check using ACL of all IPv4 traffic
> >>+* Outbound SA lookup for packets that need IPsec protection
> >>+* Add ESP and outter IP header
> >>+* Encryption/Digest
> >>+* Routing
> >>+* Write packet to port
> >>+
> >>+Constraints
> >>+-----------
> >>+* IPv4 traffic
> >>+* ESP tunnel mode
> >>+* EAS-CBC and HMAC-SHA1
> >>+* Each SA must be handle by a unique lcore (1 RX queue per port)
> >>+* No chained mbufs
> >>+
> >>+Compiling the Application
> >>+-------------------------
> >>+
> >>+To compile the application:
> >>+
> >>+#. Go to the sample application directory:
> >>+
> >>+ .. code-block:: console
> >>+
> >>+ export RTE_SDK=/path/to/rte_sdk
> >>+ cd ${RTE_SDK}/examples/ipsec-secgw
> >>+
> >>+#. Set the target (a default target is used if not specified). For example:
> >>+
> >>+ .. code-block:: console
> >>+
> >>+ export RTE_TARGET=x86_64-native-linuxapp-gcc
> >>+
> >>+ See the *DPDK Getting Started Guide* for possible RTE_TARGET values.
> >>+
> >>+#. Build the application:
> >>+
> >>+ .. code-block:: console
> >>+
> >>+ make
> >>+
> >>+Running the Application
> >>+-----------------------
> >>+
> >>+The application has a number of command line options:
> >>+
> >>+.. code-block:: console
> >>+
> >>+ ./build/ipsec-secgw [EAL options] -- -p PORTMASK -P -u PORTMASK --config
> >>+ (port,queue,lcore)[,(port,queue,lcore] --EP0|--EP1 --cdev AESNI|QAT
> >>+
> >>+where,
> >>+
> >>+* -p PORTMASK: Hexadecimal bitmask of ports to configure
> >>+
> >>+* -P: optional, sets all ports to promiscuous mode so that packets are
> >>+ accepted regardless of the packet's Ethernet MAC destination address.
> >>+ Without this option, only packets with the Ethernet MAC destination address
> >>+ set to the Ethernet address of the port are accepted (default is enabled).
> >>+
> >>+* -u PORTMASK: hexadecimal bitmask of unprotected ports
> >>+
> >>+* --config (port,queue,lcore)[,(port,queue,lcore)]: determines which queues
> >>+ from which ports are mapped to which cores
> >>+
> >>+* --cdev AESNI/QAT: choose whether to use QAT HW crypto or EASNI SW crypto
> >>+
> >>+* --EP0: configure the app as Endpoint 0.
> >>+
> >>+* --EP1: configure the app as Endpoint 1.
> >>+
> >>+Either --EP0 or --EP1 must be specified.
> >>+
> >>+The mapping of lcores to port/queues is similar to other l3fwd applications.
> >>+
> >>+For example, given the following command line:
> >>+.. code-block:: console
> >>+
> >>+ ./build/ipsec-secgw -l 20,21 -n 4 --socket-mem 0,2048 -- -p 0xf -P -u 0x3
> >>+ --config="(0,0,20),(1,0,20),(2,0,21),(3,0,21)" --cdev AESNI --EP0
> >>+
> >>+where each options means:
> >>+
> >>+* The -l option enables cores 20 and 21
> >>+
> >>+* The -n option sets memory 4 channels
> >>+
> >>+* The --socket-mem to use 2GB on socket 1
> >>+
> >>+* The -p option enables ports (detected) 0, 1, 2 and 3
> >>+
> >>+* The -P option enables promiscous mode
> >>+
> >>+* The -u option sets ports 1 and 2 as unprotected, leaving 2 and 3 as protected
> >>+
> >>+* The --config option enables one queue per port with the following mapping:
> >>+
> >>++----------+-----------+-----------+---------------------------------------+
> >>+| **Port** | **Queue** | **lcore** | **Description** |
> >>+| | | | |
> >>++----------+-----------+-----------+---------------------------------------+
> >>+| 0 | 0 | 20 | Map queue 0 from port 0 to lcore 20. |
> >>+| | | | |
> >>++----------+-----------+-----------+---------------------------------------+
> >>+| 1 | 0 | 20 | Map queue 0 from port 1 to lcore 20. |
> >>+| | | | |
> >>++----------+-----------+-----------+---------------------------------------+
> >>+| 2 | 0 | 21 | Map queue 0 from port 2 to lcore 21. |
> >>+| | | | |
> >>++----------+-----------+-----------+---------------------------------------+
> >>+| 3 | 0 | 21 | Map queue 0 from port 3 to lcore 21. |
> >>+| | | | |
> >>++----------+-----------+-----------+---------------------------------------+
> >>+
> >>+Refer to the *DPDK Getting Started Guide* for general information on running applications and the Environment Abstraction Layer (EAL) options.
> >>+
> >>+Configurations
> >>+--------------
> >>+
> >>+The following sections provide some details on the default values used to
> >>+initialize the SP, SA and Routing tables.
> >>+Currently all the configuration is hardcoded into the application.
> >>+
> >>+Security Policy Initialization
> >>+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> >>+
> >>+As mention in the overview, the Security Policies are ACL rules.
> >>+There are two ACLs, Inbound and Outboud.
> >>+The application would also replicate the set of two ACLs per socket in use.
> >>+
> >>+Following are the default rules:
> >>+
> >>+Endpoint 0 Outbound Security Policies:
> >>+
> >>++---------+------------------+-----------+------------+
> >>+| **Src** | **Dst** | **proto** | **SA idx** |
> >>+| | | | |
> >>++---------+------------------+-----------+------------+
> >>+| Any | 192.168.105.0/24 | Any | 5 |
> >>+| | | | |
> >>++---------+------------------+-----------+------------+
> >>+| Any | 192.168.106.0/24 | Any | 6 |
> >>+| | | | |
> >>++---------+------------------+-----------+------------+
> >>+| Any | 192.168.107.0/24 | Any | 7 |
> >>+| | | | |
> >>++---------+------------------+-----------+------------+
> >>+| Any | 192.168.108.0/24 | Any | 8 |
> >>+| | | | |
> >>++---------+------------------+-----------+------------+
> >>+
> >>+Endpoint 0 Inbound Security Policies:
> >>+
> >>++---------+------------------+-----------+------------+
> >>+| **Src** | **Dst** | **proto** | **SA idx** |
> >>+| | | | |
> >>++---------+------------------+-----------+------------+
> >>+| Any | 192.168.115.0/24 | Any | 5 |
> >>+| | | | |
> >>++---------+------------------+-----------+------------+
> >>+| Any | 192.168.116.0/24 | Any | 6 |
> >>+| | | | |
> >>++---------+------------------+-----------+------------+
> >>+| Any | 192.168.117.0/24 | Any | 7 |
> >>+| | | | |
> >>++---------+------------------+-----------+------------+
> >>+| Any | 192.168.118.0/24 | Any | 8 |
> >>+| | | | |
> >>++---------+------------------+-----------+------------+
> >>+
> >>+Endpoint 1 Outbound Security Policies:
> >>+
> >>++---------+------------------+-----------+------------+
> >>+| **Src** | **Dst** | **proto** | **SA idx** |
> >>+| | | | |
> >>++---------+------------------+-----------+------------+
> >>+| Any | 192.168.115.0/24 | Any | 5 |
> >>+| | | | |
> >>++---------+------------------+-----------+------------+
> >>+| Any | 192.168.116.0/24 | Any | 6 |
> >>+| | | | |
> >>++---------+------------------+-----------+------------+
> >>+| Any | 192.168.117.0/24 | Any | 7 |
> >>+| | | | |
> >>++---------+------------------+-----------+------------+
> >>+| Any | 192.168.118.0/24 | Any | 8 |
> >>+| | | | |
> >>++---------+------------------+-----------+------------+
> >>+
> >>+Endpoint 1 Inbound Security Policies:
> >>+
> >>++---------+------------------+-----------+------------+
> >>+| **Src** | **Dst** | **proto** | **SA idx** |
> >>+| | | | |
> >>++---------+------------------+-----------+------------+
> >>+| Any | 192.168.105.0/24 | Any | 5 |
> >>+| | | | |
> >>++---------+------------------+-----------+------------+
> >>+| Any | 192.168.106.0/24 | Any | 6 |
> >>+| | | | |
> >>++---------+------------------+-----------+------------+
> >>+| Any | 192.168.107.0/24 | Any | 7 |
> >>+| | | | |
> >>++---------+------------------+-----------+------------+
> >>+| Any | 192.168.108.0/24 | Any | 8 |
> >>+| | | | |
> >>++---------+------------------+-----------+------------+
> >>+
> >>+
> >>+Security Association Initialization
> >>+~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
> >>+
> >>+The SAs are kept in a array table.
> >>+
> >>+For Inbound, the SPI is used as index module the table size.
> >>+This means that on a table for 100 SA, SPI 5 and 105 would use the same index
> >>+and that is not currently supported.
> >>+
> >>+Notice that it is not an issue for Outbound traffic as we store the index and
> >>+not the SPI in the Security Policy.
> >>+
> >>+All SAs are configured with the same cipher algo, block size and key (AES-CBC
> >>+16B block), and same authentication algo, digest size and key (HMAC-SHA1, 12B
> >>+digest).
> >>+
> >>+Following are the default values:
> >>+
> >>+Endpoint 0 Outbound Security Associations:
> >>+
> >>++---------+----------------+------------------+
> >>+| **SPI** | **Tunnel src** | **Tunnel dst** |
> >>+| | | |
> >>++---------+----------------+------------------+
> >>+| 5 | 172.16.1.5 | 172.16.2.5 |
> >>+| | | |
> >>++---------+----------------+------------------+
> >>+| 6 | 172.16.1.6 | 172.16.2.6 |
> >>+| | | |
> >>++---------+----------------+------------------+
> >>+| 7 | 172.16.1.7 | 172.16.2.7 |
> >>+| | | |
> >>++---------+----------------+------------------+
> >>+| 8 | 172.16.1.8 | 172.16.2.8 |
> >>+| | | |
> >>++---------+----------------+------------------+
> >>+
> >>+Endpoint 0 Inbound Security Associations:
> >>+
> >>++---------+----------------+------------------+
> >>+| **SPI** | **Tunnel src** | **Tunnel dst** |
> >>+| | | |
> >>++---------+----------------+------------------+
> >>+| 5 | 172.16.2.5 | 172.16.1.5 |
> >>+| | | |
> >>++---------+----------------+------------------+
> >>+| 6 | 172.16.2.6 | 172.16.1.6 |
> >>+| | | |
> >>++---------+----------------+------------------+
> >>+| 7 | 172.16.2.7 | 172.16.1.7 |
> >>+| | | |
> >>++---------+----------------+------------------+
> >>+| 8 | 172.16.2.8 | 172.16.1.8 |
> >>+| | | |
> >>++---------+----------------+------------------+
> >>+
> >>+Endpoint 1 Outbound Security Associations:
> >>+
> >>++---------+----------------+------------------+
> >>+| **SPI** | **Tunnel src** | **Tunnel dst** |
> >>+| | | |
> >>++---------+----------------+------------------+
> >>+| 5 | 172.16.2.5 | 172.16.1.5 |
> >>+| | | |
> >>++---------+----------------+------------------+
> >>+| 6 | 172.16.2.6 | 172.16.1.6 |
> >>+| | | |
> >>++---------+----------------+------------------+
> >>+| 7 | 172.16.2.7 | 172.16.1.7 |
> >>+| | | |
> >>++---------+----------------+------------------+
> >>+| 8 | 172.16.2.8 | 172.16.1.8 |
> >>+| | | |
> >>++---------+----------------+------------------+
> >>+
> >>+Endpoint 1 Inbound Security Associations:
> >>+
> >>++---------+----------------+------------------+
> >>+| **SPI** | **Tunnel src** | **Tunnel dst** |
> >>+| | | |
> >>++---------+----------------+------------------+
> >>+| 5 | 172.16.1.5 | 172.16.2.5 |
> >>+| | | |
> >>++---------+----------------+------------------+
> >>+| 6 | 172.16.1.6 | 172.16.2.6 |
> >>+| | | |
> >>++---------+----------------+------------------+
> >>+| 7 | 172.16.1.7 | 172.16.2.7 |
> >>+| | | |
> >>++---------+----------------+------------------+
> >>+| 8 | 172.16.1.8 | 172.16.2.8 |
> >>+| | | |
> >>++---------+----------------+------------------+
> >>+
> >>+Routing Initialization
> >>+~~~~~~~~~~~~~~~~~~~~~~
> >>+
> >>+The Routing is implemented using LPM table.
> >>+
> >>+Following default values:
> >>+
> >>+Endpoint 0 Routing Table:
> >>+
> >>++------------------+----------+
> >>+| **Dst addr** | **Port** |
> >>+| | |
> >>++------------------+----------+
> >>+| 172.16.2.5/32 | 0 |
> >>+| | |
> >>++------------------+----------+
> >>+| 172.16.2.6/32 | 0 |
> >>+| | |
> >>++------------------+----------+
> >>+| 172.16.2.7/32 | 1 |
> >>+| | |
> >>++------------------+----------+
> >>+| 172.16.2.8/32 | 1 |
> >>+| | |
> >>++------------------+----------+
> >>+| 192.168.115.0/24 | 2 |
> >>+| | |
> >>++------------------+----------+
> >>+| 192.168.116.0/24 | 2 |
> >>+| | |
> >>++------------------+----------+
> >>+| 192.168.117.0/24 | 3 |
> >>+| | |
> >>++------------------+----------+
> >>+| 192.168.118.0/24 | 3 |
> >>+| | |
> >>++------------------+----------+
> >>+
> >>+Endpoint 1 Routing Table:
> >>+
> >>++------------------+----------+
> >>+| **Dst addr** | **Port** |
> >>+| | |
> >>++------------------+----------+
> >>+| 172.16.1.5/32 | 2 |
> >>+| | |
> >>++------------------+----------+
> >>+| 172.16.1.6/32 | 2 |
> >>+| | |
> >>++------------------+----------+
> >>+| 172.16.1.7/32 | 3 |
> >>+| | |
> >>++------------------+----------+
> >>+| 172.16.1.8/32 | 3 |
> >>+| | |
> >>++------------------+----------+
> >>+| 192.168.105.0/24 | 0 |
> >>+| | |
> >>++------------------+----------+
> >>+| 192.168.106.0/24 | 0 |
> >>+| | |
> >>++------------------+----------+
> >>+| 192.168.107.0/24 | 1 |
> >>+| | |
> >>++------------------+----------+
> >>+| 192.168.108.0/24 | 1 |
> >>+| | |
> >>++------------------+----------+
> >>diff --git a/examples/Makefile b/examples/Makefile
> >>index 1cb4785..65ce6ce 100644
> >>--- a/examples/Makefile
> >>+++ b/examples/Makefile
> >>@@ -1,6 +1,7 @@
> >> # BSD LICENSE
> >> #
> >> # Copyright(c) 2014 6WIND S.A.
> >>+# Copyright(c) 2010-2016 Intel Corporation. All rights reserved.
> >> #
> >> # Redistribution and use in source and binary forms, with or without
> >> # modification, are permitted provided that the following conditions
> >>@@ -78,5 +79,6 @@ DIRS-y += vmdq
> >> DIRS-y += vmdq_dcb
> >> DIRS-$(CONFIG_RTE_LIBRTE_POWER) += vm_power_manager
> >> DIRS-$(CONFIG_RTE_LIBRTE_CRYPTODEV) += l2fwd-crypto
> >>+DIRS-$(CONFIG_RTE_LIBRTE_CRYPTODEV) += ipsec-secgw
> >> include $(RTE_SDK)/mk/rte.extsubdir.mk
> >>diff --git a/examples/ipsec-secgw/Makefile b/examples/ipsec-secgw/Makefile
> >>new file mode 100644
> >>index 0000000..5f893b8
> >>--- /dev/null
> >>+++ b/examples/ipsec-secgw/Makefile
> >>@@ -0,0 +1,58 @@
> >>+# BSD LICENSE
> >>+#
> >>+# Copyright(c) 2010-2016 Intel Corporation. All rights reserved.
> >>+# All rights reserved.
> >>+#
> >>+# Redistribution and use in source and binary forms, with or without
> >>+# modification, are permitted provided that the following conditions
> >>+# are met:
> >>+#
> >>+# * Redistributions of source code must retain the above copyright
> >>+# notice, this list of conditions and the following disclaimer.
> >>+# * Redistributions in binary form must reproduce the above copyright
> >>+# notice, this list of conditions and the following disclaimer in
> >>+# the documentation and/or other materials provided with the
> >>+# distribution.
> >>+# * Neither the name of Intel Corporation nor the names of its
> >>+# contributors may be used to endorse or promote products derived
> >>+# from this software without specific prior written permission.
> >>+#
> >>+# THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
> >>+# "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
> >>+# LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
> >>+# A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT
> >>+# OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL,
> >>+# SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT
> >>+# LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
> >>+# DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
> >>+# THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
> >>+# (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE
> >>+# OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
> >>+
> >>+ifeq ($(RTE_SDK),)
> >>+ $(error "Please define RTE_SDK environment variable")
> >>+endif
> >>+
> >>+# Default target, can be overridden by command line or environment
> >>+RTE_TARGET ?= x86_64-native-linuxapp-gcc
> >>+
> >>+include $(RTE_SDK)/mk/rte.vars.mk
> >>+
> >>+APP = ipsec-secgw
> >>+
> >>+CFLAGS += -O3 -gdwarf-2
> >>+CFLAGS += $(WERROR_FLAGS)
> >>+
> >>+VPATH += $(SRCDIR)/librte_ipsec
> >>+
> >>+#
> >>+# all source are stored in SRCS-y
> >>+#
> >>+SRCS-y += ipsec.c
> >>+SRCS-y += esp.c
> >>+SRCS-y += sp.c
> >>+SRCS-y += sa.c
> >>+SRCS-y += rt.c
> >>+SRCS-y += ipsec-secgw.c
> >>+
> >>+include $(RTE_SDK)/mk/rte.extapp.mk
> >>diff --git a/examples/ipsec-secgw/esp.c b/examples/ipsec-secgw/esp.c
> >>new file mode 100644
> >>index 0000000..1be0621
> >>--- /dev/null
> >>+++ b/examples/ipsec-secgw/esp.c
> >>@@ -0,0 +1,256 @@
> >>+/*-
> >>+ * BSD LICENSE
> >>+ *
> >>+ * Copyright(c) 2010-2016 Intel Corporation. All rights reserved.
> >>+ * 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.
> >>+ */
> >>+
> >>+#include <stdint.h>
> >>+#include <stdlib.h>
> >>+#include <netinet/ip.h>
> >>+#include <sys/types.h>
> >>+#include <sys/stat.h>
> >>+#include <fcntl.h>
> >>+#include <unistd.h>
> >>+
> >>+#include <rte_memcpy.h>
> >>+#include <rte_crypto.h>
> >>+#include <rte_cryptodev.h>
> >>+#include <rte_random.h>
> >>+
> >>+#include "ipsec.h"
> >>+#include "esp.h"
> >>+#include "ipip.h"
> >>+
> >>+static inline void
> >>+random_iv_u64(uint64_t *buf, uint16_t n)
> >>+{
> >>+ unsigned left = n & 0x7;
> >>+ unsigned i;
> >>+
> >>+ IPSEC_ASSERT((n & 0x3) == 0);
> >>+
> >>+ for (i = 0; i < (n >> 3); i++)
> >>+ buf[i] = rte_rand();
> >>+
> >>+ if (left)
> >>+ *((uint32_t *)&buf[i]) = (uint32_t)lrand48();
> >>+}
> >>+
> >>+/* IPv4 Tunnel */
> >>+int
> >>+esp4_tunnel_inbound_pre_crypto(struct rte_mbuf *m, struct ipsec_sa *sa,
> >>+ struct rte_crypto_op *cop)
> >>+{
> >>+ uint16_t digest_len, esp_len, iphdr_len;
> >>+ int32_t payload_len;
> >>+
> >>+ IPSEC_ASSERT(m != NULL);
> >>+
> >>+ iphdr_len = sizeof(struct ip);
> >>+ esp_len = sizeof(struct esp_hdr) + sa->iv_len;
> >>+ digest_len = sa->digest_len;
> >>+ payload_len = rte_pktmbuf_pkt_len(m) - iphdr_len - esp_len - digest_len;
> >>+
> >>+ if ((payload_len & (sa->block_size - 1)) || (payload_len <= 0)) {
> >>+ IPSEC_LOG(DEBUG, IPSEC_ESP, "payload %d not a multiple of %u\n",
> >>+ payload_len, sa->block_size);
> >>+ return -EINVAL;
> >>+ }
> >>+
> >>+ cop->data.to_cipher.offset = iphdr_len + esp_len;
> >>+ cop->data.to_cipher.length = payload_len;
> >>+
> >>+ cop->data.to_hash.offset = iphdr_len;
> >>+ if (sa->auth_algo == RTE_CRYPTO_AUTH_AES_GCM)
> >>+ cop->data.to_hash.length = esp_len - sa->iv_len;
> >>+ else
> >>+ cop->data.to_hash.length = esp_len + payload_len;
> >>+
> >>+ cop->iv.data = rte_pktmbuf_mtod_offset(m, void*,
> >>+ iphdr_len + esp_len - sa->iv_len);
> >>+ cop->iv.phys_addr = rte_pktmbuf_mtophys_offset(m,
> >>+ iphdr_len + esp_len - sa->iv_len);
> >>+
> >>+ cop->iv.length = sa->iv_len;
> >>+
> >>+ cop->digest.data = rte_pktmbuf_mtod_offset(m, void*,
> >>+ iphdr_len + esp_len + payload_len);
> >>+ cop->digest.phys_addr = rte_pktmbuf_mtophys_offset(m,
> >>+ iphdr_len + esp_len + payload_len);
> >>+ cop->digest.length = digest_len;
> >>+
> >>+ return 0;
> >>+}
> >>+
> >>+int
> >>+esp4_tunnel_inbound_post_crypto(struct rte_mbuf *m, struct ipsec_sa *sa,
> >>+ struct rte_crypto_op *cop)
> >>+{
> >>+ uint16_t digest_len, esp_len, iphdr_len;
> >>+ uint8_t *nexthdr, *pad_len;
> >>+ uint8_t *padding;
> >>+ uint16_t i;
> >>+
> >>+ IPSEC_ASSERT(m != NULL);
> >>+
> >>+ iphdr_len = sizeof(struct ip);
> >>+ esp_len = sizeof(struct esp_hdr) + sa->iv_len;
> >>+ digest_len = sa->digest_len;
> >>+
> >>+ if (cop->status != RTE_CRYPTO_OP_STATUS_SUCCESS) {
> >>+ IPSEC_LOG(ERR, IPSEC_ESP, "Failed crypto op\n");
> >>+ return -1;
> >>+ }
> >>+
> >>+ nexthdr = rte_pktmbuf_mtod_offset(m, uint8_t*,
> >>+ rte_pktmbuf_pkt_len(m) - digest_len - 1);
> >>+ pad_len = nexthdr - 1;
> >>+
> >>+ padding = pad_len - *pad_len;
> >>+ for (i = 0; i < *pad_len; i++) {
> >>+ if (padding[i] != i) {
> >>+ IPSEC_LOG(ERR, IPSEC_ESP, "invalid pad_len field\n");
> >>+ return -EINVAL;
> >>+ }
> >>+ }
> >>+
> >>+ if (rte_pktmbuf_trim(m, *pad_len + 2 + digest_len)) {
> >>+ IPSEC_LOG(ERR, IPSEC_ESP,
> >>+ "failed to remove pad_len + digest\n");
> >>+ return -EINVAL;
> >>+ }
> >>+
> >>+ return ip4ip_inbound(m, iphdr_len + esp_len);
> >>+}
> >>+
> >>+int
> >>+esp4_tunnel_outbound_pre_crypto(struct rte_mbuf *m, struct ipsec_sa *sa,
> >>+ struct rte_crypto_op *cop)
> >>+{
> >>+ uint16_t digest_len, esp_len, payload_len, block_sz, pad_len;
> >>+ int32_t pad_payload_len;
> >>+ struct ip *ip;
> >>+ struct esp_hdr *esp;
> >>+ int i;
> >>+ char *padding;
> >>+
> >>+ rte_prefetch0(rte_pktmbuf_mtod(m, uint8_t *) - RTE_CACHE_LINE_SIZE);
> >>+ rte_prefetch0(rte_pktmbuf_mtod(m, uint8_t *));
> >>+
> >>+ IPSEC_ASSERT(m != NULL);
> >>+ IPSEC_ASSERT(sa != NULL);
> >>+
> >>+ /* Payload length */
> >>+ payload_len = rte_pktmbuf_pkt_len(m);
> >>+
> >>+ block_sz = sa->block_size;
> >>+
> >>+ /* Per RFC4303:
> >>+ * padded payload needs to be multiple of 4 bytes.
> >>+ * All application supported block sizes must be power of 2
> >>+ */
> >>+ pad_len = (payload_len + 2) & (block_sz - 1);
> >>+ pad_len = ((block_sz - pad_len) & (block_sz - 1)) + 2;
> >>+
> >>+ /* Padded payload length */
> >>+ pad_payload_len = payload_len + pad_len;
> >>+ /* ESP header length */
> >>+ esp_len = sizeof(struct esp_hdr) + sa->iv_len;
> >>+ /* Digest length */
> >>+ digest_len = sa->digest_len;
> >>+
> >>+ IPSEC_LOG(DEBUG, IPSEC_ESP,
> >>+ "pktlen %u, esp_len %u, digest_len %u, payload_len %u,"
> >>+ " pad_payload_len %u, block_sz %u, pad_len %u\n",
> >>+ rte_pktmbuf_pkt_len(m), esp_len, digest_len,
> >>+ payload_len, pad_payload_len, block_sz, pad_len);
> >>+
> >>+ /* Check maximum packet size */
> >>+ if (unlikely(esp_len + pad_payload_len + digest_len > IP_MAXPACKET)) {
> >>+ IPSEC_LOG(DEBUG, IPSEC_ESP, "ipsec packet is too big\n");
> >>+ return -EINVAL;
> >>+ }
> >>+
> >>+ padding = rte_pktmbuf_append(m, pad_len + digest_len);
> >>+
> >>+ IPSEC_ASSERT(padding != NULL);
> >>+
> >>+ ip = ip4ip_outbound(m, esp_len, sa->src, sa->dst);
> >>+
> >>+ esp = (struct esp_hdr *)(ip + 1);
> >>+ esp->spi = sa->spi;
> >>+
> >>+ esp->seq = htonl(sa->seq++);
> >>+
> >>+ IPSEC_LOG(DEBUG, IPSEC_ESP, "pktlen %u\n", rte_pktmbuf_pkt_len(m));
> >>+
> >>+ /* Fill pad_len using default sequential scheme */
> >>+ for (i = 0; i < pad_len - 2; i++)
> >>+ padding[i] = i + 1;
> >>+
> >>+ padding[pad_len - 2] = pad_len - 2;
> >>+ padding[pad_len - 1] = IPPROTO_IPIP;
> >>+ ip->ip_p = IPPROTO_ESP;
> >>+
> >>+ cop->data.to_cipher.offset = sizeof(struct ip) + esp_len;
> >>+ cop->data.to_cipher.length = pad_payload_len;
> >>+
> >>+ cop->data.to_hash.offset = sizeof(struct ip);
> >>+ cop->data.to_hash.length = esp_len + pad_payload_len;
> >>+
> >>+ cop->iv.data = rte_pktmbuf_mtod_offset(m, void*,
> >>+ sizeof(struct ip) + esp_len - sa->iv_len);
> >>+ cop->iv.phys_addr = rte_pktmbuf_mtophys_offset(m,
> >>+ sizeof(struct ip) + esp_len - sa->iv_len);
> >>+ cop->iv.length = sa->iv_len;
> >>+
> >>+ cop->digest.data = rte_pktmbuf_mtod_offset(m, void*,
> >>+ sizeof(struct ip) + esp_len + pad_payload_len);
> >>+ cop->digest.phys_addr = rte_pktmbuf_mtophys_offset(m,
> >>+ sizeof(struct ip) + esp_len + pad_payload_len);
> >>+ cop->digest.length = digest_len;
> >>+
> >>+ random_iv_u64((uint64_t *)cop->iv.data, cop->iv.length);
> >>+
> >>+ return 0;
> >>+}
> >>+
> >>+int
> >>+esp4_tunnel_outbound_post_crypto(struct rte_mbuf *m __rte_unused,
> >>+ struct ipsec_sa *sa __rte_unused,
> >>+ struct rte_crypto_op *cop)
> >>+{
> >>+ if (cop->status != RTE_CRYPTO_OP_STATUS_SUCCESS) {
> >>+ IPSEC_LOG(ERR, IPSEC_ESP, "Failed crypto op\n");
> >>+ return -1;
> >>+ }
> >>+
> >>+ return 0;
> >>+}
> >>diff --git a/examples/ipsec-secgw/esp.h b/examples/ipsec-secgw/esp.h
> >>new file mode 100644
> >>index 0000000..d7c8ba6
> >>--- /dev/null
> >>+++ b/examples/ipsec-secgw/esp.h
> >>@@ -0,0 +1,66 @@
> >>+/*-
> >>+ * BSD LICENSE
> >>+ *
> >>+ * Copyright(c) 2010-2016 Intel Corporation. All rights reserved.
> >>+ * 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.
> >>+ */
> >>+#ifndef __RTE_IPSEC_XFORM_ESP_H__
> >>+#define __RTE_IPSEC_XFORM_ESP_H__
> >>+
> >>+struct mbuf;
> >>+
> >>+/* RFC4303 */
> >>+struct esp_hdr {
> >>+ uint32_t spi;
> >>+ uint32_t seq;
> >>+ /* Payload */
> >>+ /* Padding */
> >>+ /* Pad Length */
> >>+ /* Next Header */
> >>+ /* Integrity Check Value - ICV */
> >>+};
> >>+
> >>+/* IPv4 Tunnel */
> >>+int
> >>+esp4_tunnel_inbound_pre_crypto(struct rte_mbuf *m, struct ipsec_sa *sa,
> >>+ struct rte_crypto_op *cop);
> >>+
> >>+int
> >>+esp4_tunnel_inbound_post_crypto(struct rte_mbuf *m, struct ipsec_sa *sa,
> >>+ struct rte_crypto_op *cop);
> >>+
> >>+int
> >>+esp4_tunnel_outbound_pre_crypto(struct rte_mbuf *m, struct ipsec_sa *sa,
> >>+ struct rte_crypto_op *cop);
> >>+
> >>+int
> >>+esp4_tunnel_outbound_post_crypto(struct rte_mbuf *m, struct ipsec_sa *sa,
> >>+ struct rte_crypto_op *cop);
> >>+
> >>+#endif /* __RTE_IPSEC_XFORM_ESP_H__ */
> >>diff --git a/examples/ipsec-secgw/ipip.h b/examples/ipsec-secgw/ipip.h
> >>new file mode 100644
> >>index 0000000..16c5dfb
> >>--- /dev/null
> >>+++ b/examples/ipsec-secgw/ipip.h
> >>@@ -0,0 +1,100 @@
> >>+/*-
> >>+ * BSD LICENSE
> >>+ *
> >>+ * Copyright(c) 2010-2016 Intel Corporation. All rights reserved.
> >>+ * 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.
> >>+ */
> >>+
> >>+#ifndef __IPIP_H__
> >>+#define __IPIP_H__
> >>+
> >>+#include <stdint.h>
> >>+#include <netinet/in.h>
> >>+#include <netinet/ip.h>
> >>+
> >>+#include <rte_mbuf.h>
> >>+
> >>+static inline struct ip *
> >>+ip4ip_outbound(struct rte_mbuf *m, uint32_t offset, uint32_t src, uint32_t dst)
> >>+{
> >>+ struct ip *inip, *outip;
> >>+
> >>+ inip = rte_pktmbuf_mtod(m, struct ip*);
> >>+
> >>+ IPSEC_ASSERT(inip->ip_v == IPVERSION || inip->ip_v == IPV6_VERSION);
> >>+
> >>+ offset += sizeof(struct ip);
> >>+
> >>+ outip = (struct ip *)rte_pktmbuf_prepend(m, offset);
> >>+
> >>+ IPSEC_ASSERT(outip != NULL);
> >>+
> >>+ /* Per RFC4301 5.1.2.1 */
> >>+ outip->ip_len = htons(rte_pktmbuf_data_len(m));
> >>+ outip->ip_v = IPVERSION;
> >>+ outip->ip_hl = 5;
> >>+ outip->ip_tos = inip->ip_tos;
> >>+
> >>+ outip->ip_id = 0;
> >>+ outip->ip_off = 0;
> >>+
> >>+ outip->ip_ttl = IPDEFTTL;
> >>+
> >>+ outip->ip_dst.s_addr = dst;
> >>+ outip->ip_src.s_addr = src;
> >>+
> >>+ return outip;
> >>+}
> >>+
> >>+static inline int
> >>+ip4ip_inbound(struct rte_mbuf *m, uint32_t offset)
> >>+{
> >>+ struct ip *inip;
> >>+ struct ip *outip;
> >>+
> >>+ outip = rte_pktmbuf_mtod(m, struct ip*);
> >>+
> >>+ IPSEC_ASSERT(outip->ip_v == IPVERSION);
> >>+
> >>+ offset += sizeof(struct ip);
> >>+ inip = (struct ip *)rte_pktmbuf_adj(m, offset);
> >>+ IPSEC_ASSERT(inip->ip_v == IPVERSION || inip->ip_v == IPV6_VERSION);
> >>+
> >>+ /* Check packet is still bigger than IP header (inner) */
> >>+ IPSEC_ASSERT(rte_pktmbuf_pkt_len(m) > sizeof(struct ip));
> >>+
> >>+ /* RFC4301 5.1.2.1 Note 6 */
> >>+ if ((inip->ip_tos & htons(IPTOS_ECN_ECT0 | IPTOS_ECN_ECT1)) &&
> >>+ ((outip->ip_tos & htons(IPTOS_ECN_CE)) == IPTOS_ECN_CE))
> >>+ inip->ip_tos |= htons(IPTOS_ECN_CE);
> >>+
> >>+ return 0;
> >>+}
> >>+
> >>+#endif /* __IPIP_H__ */
> >>diff --git a/examples/ipsec-secgw/ipsec-secgw.c b/examples/ipsec-secgw/ipsec-secgw.c
> >>new file mode 100644
> >>index 0000000..d2e8972
> >>--- /dev/null
> >>+++ b/examples/ipsec-secgw/ipsec-secgw.c
> >>@@ -0,0 +1,1218 @@
> >>+/*-
> >>+ * BSD LICENSE
> >>+ *
> >>+ * Copyright(c) 2010-2016 Intel Corporation. All rights reserved.
> >>+ * 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.
> >>+ */
> >>+
> >>+#include <stdio.h>
> >>+#include <stdlib.h>
> >>+#include <stdint.h>
> >>+#include <inttypes.h>
> >>+#include <sys/types.h>
> >>+#include <string.h>
> >>+#include <sys/queue.h>
> >>+#include <stdarg.h>
> >>+#include <errno.h>
> >>+#include <getopt.h>
> >>+
> >>+#include <rte_common.h>
> >>+#include <rte_byteorder.h>
> >>+#include <rte_log.h>
> >>+#include <rte_eal.h>
> >>+#include <rte_launch.h>
> >>+#include <rte_atomic.h>
> >>+#include <rte_cycles.h>
> >>+#include <rte_prefetch.h>
> >>+#include <rte_lcore.h>
> >>+#include <rte_per_lcore.h>
> >>+#include <rte_branch_prediction.h>
> >>+#include <rte_interrupts.h>
> >>+#include <rte_pci.h>
> >>+#include <rte_random.h>
> >>+#include <rte_debug.h>
> >>+#include <rte_ether.h>
> >>+#include <rte_ethdev.h>
> >>+#include <rte_mempool.h>
> >>+#include <rte_mbuf.h>
> >>+#include <rte_acl.h>
> >>+#include <rte_lpm.h>
> >>+#include <rte_cryptodev.h>
> >>+#include <rte_cryptodev.h>
> >>+
> >>+#include "ipsec.h"
> >>+
> >>+#define RTE_LOGTYPE_IPSEC RTE_LOGTYPE_USER1
> >>+
> >>+#define MAX_JUMBO_PKT_LEN 9600
> >>+
> >>+#define MEMPOOL_CACHE_SIZE 256
> >>+
> >>+#define NB_MBUF (32000)
> >>+
> >>+#define CDEV_MP_NB_OBJS 2048
> >>+#define CDEV_MP_CACHE_SZ 64
> >>+#define MAX_QUEUE_PAIRS 1
> >>+
> >>+#define OPTION_CONFIG "config"
> >>+#define OPTION_EP0 "ep0"
> >>+#define OPTION_EP1 "ep1"
> >>+#define OPTION_CDEV_TYPE "cdev"
> >>+
> >>+#define MAX_PKT_BURST 32
> >>+#define BURST_TX_DRAIN_US 100 /* TX drain every ~100us */
> >>+
> >>+#define NB_SOCKETS 4
> >>+
> >>+/* Configure how many packets ahead to prefetch, when reading packets */
> >>+#define PREFETCH_OFFSET 3
> >>+
> >>+#define MAX_RX_QUEUE_PER_LCORE 16
> >>+
> >>+#define MAX_LCORE_PARAMS 1024
> >>+
> >>+#define UNPROTECTED_PORT(port) (unprotected_port_mask & (1 << portid))
> >>+
> >>+/*
> >>+ * Configurable number of RX/TX ring descriptors
> >>+ */
> >>+#define IPSEC_SECGW_RX_DESC_DEFAULT 128
> >>+#define IPSEC_SECGW_TX_DESC_DEFAULT 512
> >>+static uint16_t nb_rxd = IPSEC_SECGW_RX_DESC_DEFAULT;
> >>+static uint16_t nb_txd = IPSEC_SECGW_TX_DESC_DEFAULT;
> >>+
> >>+#if RTE_BYTE_ORDER != RTE_LITTLE_ENDIAN
> >>+#define __BYTES_TO_UINT64(a, b, c, d, e, f, g, h) \
> >>+ (((uint64_t)((a) & 0xff) << 56) | \
> >>+ ((uint64_t)((b) & 0xff) << 48) | \
> >>+ ((uint64_t)((c) & 0xff) << 40) | \
> >>+ ((uint64_t)((d) & 0xff) << 32) | \
> >>+ ((uint64_t)((e) & 0xff) << 24) | \
> >>+ ((uint64_t)((f) & 0xff) << 16) | \
> >>+ ((uint64_t)((g) & 0xff) << 8) | \
> >>+ ((uint64_t)(h) & 0xff))
> >>+#else
> >>+#define __BYTES_TO_UINT64(a, b, c, d, e, f, g, h) \
> >>+ (((uint64_t)((h) & 0xff) << 56) | \
> >>+ ((uint64_t)((g) & 0xff) << 48) | \
> >>+ ((uint64_t)((f) & 0xff) << 40) | \
> >>+ ((uint64_t)((e) & 0xff) << 32) | \
> >>+ ((uint64_t)((d) & 0xff) << 24) | \
> >>+ ((uint64_t)((c) & 0xff) << 16) | \
> >>+ ((uint64_t)((b) & 0xff) << 8) | \
> >>+ ((uint64_t)(a) & 0xff))
> >>+#endif
> >>+#define ETHADDR(a, b, c, d, e, f) (__BYTES_TO_UINT64(a, b, c, d, e, f, 0, 0))
> >>+
> >>+#define ETHADDR_TO_UINT64(addr) __BYTES_TO_UINT64( \
> >>+ addr.addr_bytes[0], addr.addr_bytes[1], \
> >>+ addr.addr_bytes[2], addr.addr_bytes[3], \
> >>+ addr.addr_bytes[4], addr.addr_bytes[5], \
> >>+ 0, 0)
> >>+
> >>+/* port/source ethernet addr and destination ethernet addr */
> >>+struct ethaddr_info {
> >>+ uint64_t src, dst;
> >>+};
> >>+
> >>+struct ethaddr_info ethaddr_tbl[RTE_MAX_ETHPORTS] = {
> >>+ { 0, ETHADDR(0x00, 0x16, 0x3e, 0x7e, 0x94, 0x9a) },
> >>+ { 0, ETHADDR(0x00, 0x16, 0x3e, 0x22, 0xa1, 0xd9) },
> >>+ { 0, ETHADDR(0x00, 0x16, 0x3e, 0x08, 0x69, 0x26) },
> >>+ { 0, ETHADDR(0x00, 0x16, 0x3e, 0x49, 0x9e, 0xdd) }
> >>+};
> >>+
> >>+/* mask of enabled ports */
> >>+static uint32_t enabled_port_mask;
> >>+static uint32_t unprotected_port_mask;
> >>+static int promiscuous_on = 1;
> >>+static int numa_on = 1; /**< NUMA is enabled by default. */
> >>+static int ep = -1; /**< Endpoint configuration (0 or 1) */
> >>+static unsigned cdev_type = -1;
> >>+static unsigned nb_lcores;
> >>+
> >>+struct buffer {
> >>+ uint16_t len;
> >>+ struct rte_mbuf *m_table[MAX_PKT_BURST];
> >>+};
> >>+
> >>+struct lcore_rx_queue {
> >>+ uint8_t port_id;
> >>+ uint8_t queue_id;
> >>+} __rte_cache_aligned;
> >>+
> >>+struct lcore_params {
> >>+ uint8_t port_id;
> >>+ uint8_t queue_id;
> >>+ uint8_t lcore_id;
> >>+} __rte_cache_aligned;
> >>+
> >>+static struct lcore_params lcore_params_array[MAX_LCORE_PARAMS];
> >>+
> >>+static struct lcore_params *lcore_params;
> >>+static uint16_t nb_lcore_params;
> >>+
> >>+struct lcore_conf {
> >>+ uint16_t nb_rx_queue;
> >>+ struct lcore_rx_queue rx_queue_list[MAX_RX_QUEUE_PER_LCORE];
> >>+ uint16_t tx_queue_id[RTE_MAX_ETHPORTS];
> >>+ struct buffer tx_mbufs[RTE_MAX_ETHPORTS];
> >>+ uint16_t cdev;
> >>+ uint16_t cdev_q;
> >>+} __rte_cache_aligned;
> >>+
> >>+static struct lcore_conf lcore_conf[RTE_MAX_LCORE];
> >>+
> >>+static struct rte_eth_conf port_conf = {
> >>+ .rxmode = {
> >>+ .mq_mode = ETH_MQ_RX_RSS,
> >>+ .max_rx_pkt_len = ETHER_MAX_LEN,
> >>+ .split_hdr_size = 0,
> >>+ .header_split = 0, /**< Header Split disabled */
> >>+ .hw_ip_checksum = 1, /**< IP checksum offload enabled */
> >>+ .hw_vlan_filter = 0, /**< VLAN filtering disabled */
> >>+ .jumbo_frame = 0, /**< Jumbo Frame Support disabled */
> >>+ .hw_strip_crc = 0, /**< CRC stripped by hardware */
> >>+ },
> >>+ .rx_adv_conf = {
> >>+ .rss_conf = {
> >>+ .rss_key = NULL,
> >>+ .rss_hf = ETH_RSS_IP | ETH_RSS_UDP |
> >>+ ETH_RSS_TCP | ETH_RSS_SCTP,
> >>+ },
> >>+ },
> >>+ .txmode = {
> >>+ .mq_mode = ETH_MQ_TX_NONE,
> >>+ },
> >>+};
> >>+
> >>+static struct socket_ctx socket_ctx[NB_SOCKETS];
> >>+
> >>+struct traffic_type {
> >>+ const uint8_t *data[MAX_PKT_BURST * 2];
> >>+ struct rte_mbuf *pkts[MAX_PKT_BURST * 2];
> >>+ uint32_t res[MAX_PKT_BURST * 2];
> >>+ uint32_t num;
> >>+};
> >>+
> >>+struct ipsec_traffic {
> >>+ struct traffic_type ipsec4;
> >>+ struct traffic_type ipv4;
> >>+};
> >>+
> >>+static inline void
> >>+prepare_one_packet(struct rte_mbuf *pkt, struct ipsec_traffic *t)
> >>+{
> >>+ uint8_t *nlp;
> >>+
> >>+ if (RTE_ETH_IS_IPV4_HDR(pkt->packet_type)) {
> >>+ rte_pktmbuf_adj(pkt, ETHER_HDR_LEN);
> >>+ nlp = rte_pktmbuf_mtod_offset(pkt, uint8_t *,
> >>+ offsetof(struct ip, ip_p));
> >>+ if (*nlp == IPPROTO_ESP)
> >>+ t->ipsec4.pkts[(t->ipsec4.num)++] = pkt;
> >>+ else {
> >>+ t->ipv4.data[t->ipv4.num] = nlp;
> >>+ t->ipv4.pkts[(t->ipv4.num)++] = pkt;
> >>+ }
> >>+ } else {
> >>+ /* Unknown/Unsupported type, drop the packet */
> >>+ rte_pktmbuf_free(pkt);
> >>+ }
> >>+}
> >>+
> >>+static inline void
> >>+prepare_traffic(struct rte_mbuf **pkts, struct ipsec_traffic *t,
> >>+ uint16_t nb_pkts)
> >>+{
> >>+ int i;
> >>+
> >>+ t->ipsec4.num = 0;
> >>+ t->ipv4.num = 0;
> >>+
> >>+ for (i = 0; i < (nb_pkts - PREFETCH_OFFSET); i++) {
> >>+ rte_prefetch0(rte_pktmbuf_mtod(pkts[i + PREFETCH_OFFSET],
> >>+ void *));
> >>+ prepare_one_packet(pkts[i], t);
> >>+ }
> >>+ /* Process left packets */
> >>+ for (; i < nb_pkts; i++)
> >>+ prepare_one_packet(pkts[i], t);
> >>+}
> >>+
> >>+static inline void
> >>+prepare_tx_pkt(struct rte_mbuf *pkt, uint8_t port)
> >>+{
> >>+ pkt->ol_flags |= PKT_TX_IP_CKSUM | PKT_TX_IPV4;
> >>+ pkt->l3_len = sizeof(struct ip);
> >>+ pkt->l2_len = ETHER_HDR_LEN;
> >>+
> >>+ uint64_t *ethhdr = (uint64_t *)rte_pktmbuf_prepend(pkt, ETHER_HDR_LEN);
> >>+#if RTE_BYTE_ORDER == RTE_LITTLE_ENDIAN
> >>+ ethhdr[0] = ethaddr_tbl[port].dst |
> >>+ ((ethaddr_tbl[port].src & 0xffff) << 48);
> >>+ ethhdr[1] = ethaddr_tbl[port].src >> 16 |
> >>+ ((uint64_t)rte_cpu_to_be_16(ETHER_TYPE_IPv4) << 32) |
> >>+ (ethhdr[1] & (0xffffUL << 48));
> >>+#else
> >>+ ethhdr[0] = ethaddr_tbl[port].dst |
> >>+ (ethaddr_tbl[port].src >> 48 & 0xffff);
> >>+ ethhdr[1] = ethaddr_tbl[port].src << 16 | ((ETHER_TYPE_IPv4)UL << 16) |
> >>+ (ethhdr[1] & 0xffffUL);
> >>+#endif
> >>+}
> >>+
> >>+static inline void
> >>+prepare_tx_burst(struct rte_mbuf *pkts[], uint16_t nb_pkts, uint8_t port)
> >>+{
> >>+ int i;
> >>+ const int prefetch_offset = 2;
> >>+
> >>+ for (i = 0; i < (nb_pkts - prefetch_offset); i++) {
> >>+ rte_prefetch0(pkts[i + prefetch_offset]->cacheline1);
> >>+ prepare_tx_pkt(pkts[i], port);
> >>+ }
> >>+ /* Process left packets */
> >>+ for (; i < nb_pkts; i++)
> >>+ prepare_tx_pkt(pkts[i], port);
> >>+}
> >>+
> >>+/* Send burst of packets on an output interface */
> >>+static inline int
> >>+send_burst(struct lcore_conf *qconf, uint16_t n, uint8_t port)
> >>+{
> >>+ struct rte_mbuf **m_table;
> >>+ int ret;
> >>+ uint16_t queueid;
> >>+
> >>+ queueid = qconf->tx_queue_id[port];
> >>+ m_table = (struct rte_mbuf **)qconf->tx_mbufs[port].m_table;
> >>+
> >>+ prepare_tx_burst(m_table, n, port);
> >>+
> >>+ ret = rte_eth_tx_burst(port, queueid, m_table, n);
> >>+ if (unlikely(ret < n)) {
> >>+ do {
> >>+ rte_pktmbuf_free(m_table[ret]);
> >>+ } while (++ret < n);
> >>+ }
> >>+
> >>+ return 0;
> >>+}
> >>+
> >>+/* Enqueue a single packet, and send burst if queue is filled */
> >>+static inline int
> >>+send_single_packet(struct rte_mbuf *m, uint8_t port)
> >>+{
> >>+ uint32_t lcore_id;
> >>+ uint16_t len;
> >>+ struct lcore_conf *qconf;
> >>+
> >>+ lcore_id = rte_lcore_id();
> >>+
> >>+ qconf = &lcore_conf[lcore_id];
> >>+ len = qconf->tx_mbufs[port].len;
> >>+ qconf->tx_mbufs[port].m_table[len] = m;
> >>+ len++;
> >>+
> >>+ /* enough pkts to be sent */
> >>+ if (unlikely(len == MAX_PKT_BURST)) {
> >>+ send_burst(qconf, MAX_PKT_BURST, port);
> >>+ len = 0;
> >>+ }
> >>+
> >>+ qconf->tx_mbufs[port].len = len;
> >>+ return 0;
> >>+}
> >>+
> >>+static inline void
> >>+process_pkts_inbound(struct socket_ctx *ctx, struct ipsec_traffic *traffic)
> >>+{
> >>+ struct sa_ctx *sa_ctx = ctx->sa_ipv4_in;
> >>+ struct sp_ctx *sp_ctx = ctx->sp_ipv4_in;
> >>+ struct rte_mbuf *m;
> >>+ uint16_t idx, nb_pkts_in, i, j;
> >>+ uint32_t sa_idx, res;
> >>+ struct ipsec_ctx ipsec_ctx;
> >>+ struct lcore_conf *qconf;
> >>+
> >>+ qconf = &lcore_conf[rte_lcore_id()];
> >>+ ipsec_ctx.cdev = qconf->cdev;
> >>+ ipsec_ctx.queue = qconf->cdev_q;
> >>+ ipsec_ctx.sa_ctx = sa_ctx;
> >>+
> >>+ nb_pkts_in = ipsec_inbound(&ipsec_ctx, traffic->ipsec4.pkts,
> >>+ traffic->ipsec4.num, MAX_PKT_BURST);
> >>+
> >>+ /* SP/ACL Inbound check ipsec and ipv4 */
> >>+ for (i = 0; i < nb_pkts_in; i++) {
> >>+ idx = traffic->ipv4.num++;
> >>+ m = traffic->ipsec4.pkts[i];
> >>+ traffic->ipv4.pkts[idx] = m;
> >>+ traffic->ipv4.data[idx] = rte_pktmbuf_mtod_offset(m,
> >>+ uint8_t *, offsetof(struct ip, ip_p));
> >>+ }
> >>+
> >>+ rte_acl_classify((struct rte_acl_ctx *)sp_ctx, traffic->ipv4.data,
> >>+ traffic->ipv4.res, traffic->ipv4.num,
> >>+ DEFAULT_MAX_CATEGORIES);
> >>+
> >>+ j = 0;
> >>+ for (i = 0; i < traffic->ipv4.num - nb_pkts_in; i++) {
> >>+ m = traffic->ipv4.pkts[i];
> >>+ res = traffic->ipv4.res[i];
> >>+ if (res & ~BYPASS) {
> >>+ rte_pktmbuf_free(m);
> >>+ continue;
> >>+ }
> >>+ traffic->ipv4.pkts[j++] = m;
> >>+ }
> >>+ /* Check return SA SPI matches pkt SPI */
> >>+ for ( ; i < traffic->ipv4.num; i++) {
> >>+ m = traffic->ipv4.pkts[i];
> >>+ sa_idx = traffic->ipv4.res[i] & PROTECT_MASK;
> >>+ if (sa_idx == 0 || !inbound_sa_check(sa_ctx, m, sa_idx)) {
> >>+ rte_pktmbuf_free(m);
> >>+ continue;
> >>+ }
> >>+ traffic->ipv4.pkts[j++] = m;
> >>+ }
> >>+ traffic->ipv4.num = j;
> >>+}
> >>+
> >>+static inline void
> >>+process_pkts_outbound(struct socket_ctx *ctx, struct ipsec_traffic *traffic)
> >>+{
> >>+ struct sa_ctx *sa_ctx = ctx->sa_ipv4_out;
> >>+ struct sp_ctx *sp_ctx = ctx->sp_ipv4_out;
> >>+ struct rte_mbuf *m;
> >>+ uint16_t idx, nb_pkts_out, i, j;
> >>+ uint32_t sa_idx, res;
> >>+ struct ipsec_ctx ipsec_ctx;
> >>+ struct lcore_conf *qconf;
> >>+
> >>+ qconf = &lcore_conf[rte_lcore_id()];
> >>+ ipsec_ctx.cdev = qconf->cdev;
> >>+ ipsec_ctx.queue = qconf->cdev_q;
> >>+ ipsec_ctx.sa_ctx = sa_ctx;
> >>+
> >>+ rte_acl_classify((struct rte_acl_ctx *)sp_ctx, traffic->ipv4.data,
> >>+ traffic->ipv4.res, traffic->ipv4.num,
> >>+ DEFAULT_MAX_CATEGORIES);
> >IMO, an option for single SA based outbound processing would be useful
> >measuring performance bottlenecks with SA lookup.
> >
>
> Hi Jerin,
>
> Are you suggesting to have an option so we basically encrypt all traffic
> using
> a single SA bypassing the SP/ACL ?
Yes. Basicaly an option to bypass "rte_acl_classify" if its for single
SA use case.
>
> Currently the ACL entry contains the SA index in the SA table, so there is
> no SA
> lookup on outbound.
> That might have to change a bit if we were to add IKE support.
>
> >>+ j = 0;
> >>+ for (i = 0; i < traffic->ipv4.num; i++) {
> >>+ m = traffic->ipv4.pkts[i];
> >>+ res = traffic->ipv4.res[i];
> >>+ sa_idx = res & PROTECT_MASK;
> >>+ if (res & DISCARD) {
> >>+ rte_pktmbuf_free(m);
> >>+ } else if (sa_idx != 0) {
> >>+ traffic->ipsec4.res[traffic->ipsec4.num] = sa_idx;
> >>+ traffic->ipsec4.pkts[traffic->ipsec4.num++] = m;
> >>+ } else /* BYPASS */
> >>+ traffic->ipv4.pkts[j++] = m;
> >>+ }
> >>+ traffic->ipv4.num = j;
> >>+
> >>+ nb_pkts_out = ipsec_outbound(&ipsec_ctx, traffic->ipsec4.pkts,
> >>+ traffic->ipsec4.res, traffic->ipsec4.num,
> >>+ MAX_PKT_BURST);
> >>+
> >>+ for (i = 0; i < nb_pkts_out; i++) {
> >>+ idx = traffic->ipv4.num++;
> >>+ m = traffic->ipsec4.pkts[i];
> >>+ traffic->ipv4.pkts[idx] = m;
> >>+ }
> >>+}
> >>+
> >>+static inline void
> >>+route_pkts(struct rt_ctx *rt_ctx, struct rte_mbuf *pkts[], uint8_t nb_pkts)
> >>+{
> >>+ uint16_t hop[MAX_PKT_BURST * 2];
> >>+ uint32_t dst_ip[MAX_PKT_BURST * 2];
> >>+ uint16_t i, offset;
> >>+
> >>+ if (nb_pkts == 0)
> >>+ return;
> >>+
> >>+ for (i = 0; i < nb_pkts; i++) {
> >>+ offset = offsetof(struct ip, ip_dst);
> >>+ dst_ip[i] = *rte_pktmbuf_mtod_offset(pkts[i],
> >>+ uint32_t *, offset);
> >>+ dst_ip[i] = rte_be_to_cpu_32(dst_ip[i]);
> >>+ }
> >>+
> >>+ rte_lpm_lookup_bulk((struct rte_lpm *)rt_ctx, dst_ip, hop, nb_pkts);
> >>+
> >>+ for (i = 0; i < nb_pkts; i++) {
> >>+ if ((hop[i] & RTE_LPM_LOOKUP_SUCCESS) == 0) {
> >>+ rte_pktmbuf_free(pkts[i]);
> >>+ continue;
> >>+ }
> >>+ send_single_packet(pkts[i], hop[i] & 0xff);
> >>+ }
> >>+}
> >>+
> >>+static inline void
> >>+process_pkts(struct socket_ctx *ctx, struct rte_mbuf **pkts,
> >>+ uint8_t nb_pkts, uint8_t portid)
> >>+{
> >>+ struct ipsec_traffic traffic;
> >>+
> >>+ prepare_traffic(pkts, &traffic, nb_pkts);
> >>+
> >>+ if (UNPROTECTED_PORT(portid))
> >>+ process_pkts_inbound(ctx, &traffic);
> >>+ else
> >>+ process_pkts_outbound(ctx, &traffic);
> >>+
> >>+ route_pkts(ctx->rt_ipv4, traffic.ipv4.pkts, traffic.ipv4.num);
> >>+}
> >>+
> >>+static inline void
> >>+drain_buffers(struct lcore_conf *qconf)
> >>+{
> >>+ struct buffer *buf;
> >>+ unsigned portid;
> >>+
> >>+ for (portid = 0; portid < RTE_MAX_ETHPORTS; portid++) {
> >>+ buf = &qconf->tx_mbufs[portid];
> >>+ if (buf->len == 0)
> >>+ continue;
> >>+ send_burst(qconf, buf->len, portid);
> >>+ buf->len = 0;
> >>+ }
> >>+}
> >>+
> >>+/* main processing loop */
> >>+static int
> >>+main_loop(__attribute__((unused)) void *dummy)
> >>+{
> >>+ struct rte_mbuf *pkts[MAX_PKT_BURST];
> >>+ unsigned lcore_id;
> >>+ uint64_t prev_tsc, diff_tsc, cur_tsc;
> >>+ int i, nb_rx;
> >>+ uint8_t portid, queueid;
> >>+ struct lcore_conf *qconf;
> >>+ int socket_id;
> >>+ const uint64_t drain_tsc = (rte_get_tsc_hz() + US_PER_S - 1)
> >>+ / US_PER_S * BURST_TX_DRAIN_US;
> >>+ struct socket_ctx *ctx;
> >>+ struct lcore_rx_queue *rxql;
> >>+
> >>+ prev_tsc = 0;
> >>+ lcore_id = rte_lcore_id();
> >>+ qconf = &lcore_conf[lcore_id];
> >>+ rxql = qconf->rx_queue_list;
> >>+ socket_id = rte_lcore_to_socket_id(lcore_id);
> >>+
> >>+ ctx = &socket_ctx[socket_id];
> >>+
> >>+ if (qconf->nb_rx_queue == 0) {
> >>+ RTE_LOG(INFO, IPSEC, "lcore %u has nothing to do\n", lcore_id);
> >>+ return 0;
> >>+ }
> >>+
> >>+ RTE_LOG(INFO, IPSEC, "entering main loop on lcore %u\n", lcore_id);
> >>+
> >>+ for (i = 0; i < qconf->nb_rx_queue; i++) {
> >>+ portid = rxql[i].port_id;
> >>+ queueid = rxql[i].queue_id;
> >>+ RTE_LOG(INFO, IPSEC,
> >>+ " -- lcoreid=%u portid=%hhu rxqueueid=%hhu\n",
> >>+ lcore_id, portid, queueid);
> >>+ }
> >>+
> >>+ while (1) {
> >>+ cur_tsc = rte_rdtsc();
> >>+
> >>+ /* TX queue buffer drain */
> >>+ diff_tsc = cur_tsc - prev_tsc;
> >>+
> >>+ if (unlikely(diff_tsc > drain_tsc)) {
> >>+ drain_buffers(qconf);
> >>+ prev_tsc = cur_tsc;
> >>+ }
> >>+
> >>+ /* Read packet from RX queues */
> >>+ for (i = 0; i < qconf->nb_rx_queue; ++i) {
> >>+ portid = rxql[i].port_id;
> >>+ queueid = rxql[i].queue_id;
> >>+ nb_rx = rte_eth_rx_burst(portid, queueid,
> >>+ pkts, MAX_PKT_BURST);
> >>+
> >>+ if (nb_rx > 0)
> >>+ process_pkts(ctx, pkts, nb_rx, portid);
> >>+ }
> >>+ }
> >>+}
> >>+
> >>+static int
> >>+check_params(void)
> >>+{
> >>+ uint8_t lcore, portid, nb_ports;
> >>+ uint16_t i;
> >>+ int socket_id;
> >>+
> >>+ if (lcore_params == NULL) {
> >>+ printf("Error: No port/queue/core mappings\n");
> >>+ return -1;
> >>+ }
> >>+
> >>+ nb_ports = rte_eth_dev_count();
> >>+ if (nb_ports > RTE_MAX_ETHPORTS)
> >>+ nb_ports = RTE_MAX_ETHPORTS;
> >>+
> >>+ for (i = 0; i < nb_lcore_params; ++i) {
> >>+ lcore = lcore_params[i].lcore_id;
> >>+ if (!rte_lcore_is_enabled(lcore)) {
> >>+ printf("error: lcore %hhu is not enabled in "
> >>+ "lcore mask\n", lcore);
> >>+ return -1;
> >>+ }
> >>+ socket_id = rte_lcore_to_socket_id(lcore);
> >>+ if (socket_id != 0 && numa_on == 0) {
> >>+ printf("warning: lcore %hhu is on socket %d "
> >>+ "with numa off\n",
> >>+ lcore, socket_id);
> >>+ }
> >>+ portid = lcore_params[i].port_id;
> >>+ if ((enabled_port_mask & (1 << portid)) == 0) {
> >>+ printf("port %u is not enabled in port mask\n", portid);
> >>+ return -1;
> >>+ }
> >>+ if (portid >= nb_ports) {
> >>+ printf("port %u is not present on the board\n", portid);
> >>+ return -1;
> >>+ }
> >>+ }
> >>+ return 0;
> >>+}
> >>+
> >>+static uint8_t
> >>+get_port_nb_rx_queues(const uint8_t port)
> >>+{
> >>+ int queue = -1;
> >>+ uint16_t i;
> >>+
> >>+ for (i = 0; i < nb_lcore_params; ++i) {
> >>+ if (lcore_params[i].port_id == port &&
> >>+ lcore_params[i].queue_id > queue)
> >>+ queue = lcore_params[i].queue_id;
> >>+ }
> >>+ return (uint8_t)(++queue);
> >>+}
> >>+
> >>+static int
> >>+init_lcore_rx_queues(void)
> >>+{
> >>+ uint16_t i, nb_rx_queue;
> >>+ uint8_t lcore;
> >>+
> >>+ for (i = 0; i < nb_lcore_params; ++i) {
> >>+ lcore = lcore_params[i].lcore_id;
> >>+ nb_rx_queue = lcore_conf[lcore].nb_rx_queue;
> >>+ if (nb_rx_queue >= MAX_RX_QUEUE_PER_LCORE) {
> >>+ printf("error: too many queues (%u) for lcore: %u\n",
> >>+ nb_rx_queue + 1, lcore);
> >>+ return -1;
> >>+ }
> >>+ lcore_conf[lcore].rx_queue_list[nb_rx_queue].port_id =
> >>+ lcore_params[i].port_id;
> >>+ lcore_conf[lcore].rx_queue_list[nb_rx_queue].queue_id =
> >>+ lcore_params[i].queue_id;
> >>+ lcore_conf[lcore].nb_rx_queue++;
> >>+ }
> >>+ return 0;
> >>+}
> >>+
> >>+/* display usage */
> >>+static void
> >>+print_usage(const char *prgname)
> >>+{
> >>+ printf("%s [EAL options] -- -p PORTMASK -P -u PORTMASK"
> >>+ " --"OPTION_CONFIG" (port,queue,lcore)[,(port,queue,lcore]"
> >>+ " --EP0|--EP1 --cdev AESNI|QAT\n"
> >>+ " -p PORTMASK: hexadecimal bitmask of ports to configure\n"
> >>+ " -P : enable promiscuous mode\n"
> >>+ " -u PORTMASK: hexadecimal bitmask of unprotected ports\n"
> >>+ " --"OPTION_CONFIG": (port,queue,lcore): "
> >>+ "rx queues configuration\n"
> >>+ " --cdev AESNI | QAT\n"
> >>+ " --EP0: Configure as Endpoint 0\n"
> >>+ " --EP1: Configure as Endpoint 1\n", prgname);
> >>+}
> >>+
> >>+static int
> >>+parse_portmask(const char *portmask)
> >>+{
> >>+ char *end = NULL;
> >>+ unsigned long pm;
> >>+
> >>+ /* parse hexadecimal string */
> >>+ pm = strtoul(portmask, &end, 16);
> >>+ if ((portmask[0] == '\0') || (end == NULL) || (*end != '\0'))
> >>+ return -1;
> >>+
> >>+ if (pm == 0)
> >>+ return -1;
> >>+
> >>+ return pm;
> >>+}
> >>+
> >>+static int
> >>+parse_config(const char *q_arg)
> >>+{
> >>+ char s[256];
> >>+ const char *p, *p0 = q_arg;
> >>+ char *end;
> >>+ enum fieldnames {
> >>+ FLD_PORT = 0,
> >>+ FLD_QUEUE,
> >>+ FLD_LCORE,
> >>+ _NUM_FLD
> >>+ };
> >>+ unsigned long int_fld[_NUM_FLD];
> >>+ char *str_fld[_NUM_FLD];
> >>+ int i;
> >>+ unsigned size;
> >>+
> >>+ nb_lcore_params = 0;
> >>+
> >>+ while ((p = strchr(p0, '(')) != NULL) {
> >>+ ++p;
> >>+ p0 = strchr(p, ')');
> >>+ if (p0 == NULL)
> >>+ return -1;
> >>+
> >>+ size = p0 - p;
> >>+ if (size >= sizeof(s))
> >>+ return -1;
> >>+
> >>+ snprintf(s, sizeof(s), "%.*s", size, p);
> >>+ if (rte_strsplit(s, sizeof(s), str_fld, _NUM_FLD, ',') !=
> >>+ _NUM_FLD)
> >>+ return -1;
> >>+ for (i = 0; i < _NUM_FLD; i++) {
> >>+ errno = 0;
> >>+ int_fld[i] = strtoul(str_fld[i], &end, 0);
> >>+ if (errno != 0 || end == str_fld[i] || int_fld[i] > 255)
> >>+ return -1;
> >>+ }
> >>+ if (nb_lcore_params >= MAX_LCORE_PARAMS) {
> >>+ printf("exceeded max number of lcore params: %hu\n",
> >>+ nb_lcore_params);
> >>+ return -1;
> >>+ }
> >>+ lcore_params_array[nb_lcore_params].port_id =
> >>+ (uint8_t)int_fld[FLD_PORT];
> >>+ lcore_params_array[nb_lcore_params].queue_id =
> >>+ (uint8_t)int_fld[FLD_QUEUE];
> >>+ lcore_params_array[nb_lcore_params].lcore_id =
> >>+ (uint8_t)int_fld[FLD_LCORE];
> >>+ ++nb_lcore_params;
> >>+ }
> >>+ lcore_params = lcore_params_array;
> >>+ return 0;
> >>+}
> >>+
> >>+#define __STRNCMP(name, opt) (!strncmp(name, opt, sizeof(opt)))
> >>+static int
> >>+parse_args_long_options(struct option *lgopts, int option_index)
> >>+{
> >>+ int ret = -1;
> >>+ const char *optname = lgopts[option_index].name;
> >>+
> >>+ if (__STRNCMP(optname, OPTION_CONFIG)) {
> >>+ ret = parse_config(optarg);
> >>+ if (ret)
> >>+ printf("invalid config\n");
> >>+ }
> >>+
> >>+ if (__STRNCMP(optname, OPTION_EP0)) {
> >>+ printf("endpoint 0\n");
> >>+ ep = 0;
> >>+ ret = 0;
> >>+ }
> >>+
> >>+ if (__STRNCMP(optname, OPTION_EP1)) {
> >>+ printf("endpoint 1\n");
> >>+ ep = 1;
> >>+ ret = 0;
> >>+ }
> >>+
> >>+ if (__STRNCMP(optname, OPTION_CDEV_TYPE)) {
> >>+ if (__STRNCMP(optarg, "AESNI")) {
> >>+ cdev_type = RTE_CRYPTODEV_AESNI_MB_PMD;
> >>+ ret = 0;
> >>+ } else if (__STRNCMP(optarg, "QAT")) {
> >>+ cdev_type = RTE_CRYPTODEV_QAT_PMD;
> >>+ ret = 0;
> >>+ }
> >>+ }
> >>+
> >>+ return ret;
> >>+}
> >>+#undef __STRNCMP
> >>+
> >>+static int
> >>+parse_args(int argc, char **argv)
> >>+{
> >>+ int opt, ret;
> >>+ char **argvopt;
> >>+ int option_index;
> >>+ char *prgname = argv[0];
> >>+ static struct option lgopts[] = {
> >>+ {OPTION_CONFIG, 1, 0, 0},
> >>+ {OPTION_EP0, 0, 0, 0},
> >>+ {OPTION_EP1, 0, 0, 0},
> >>+ {OPTION_CDEV_TYPE, 1, 0, 0},
> >>+ {NULL, 0, 0, 0}
> >>+ };
> >>+
> >>+ argvopt = argv;
> >>+
> >>+ while ((opt = getopt_long(argc, argvopt, "p:Pu:",
> >>+ lgopts, &option_index)) != EOF) {
> >>+
> >>+ switch (opt) {
> >>+ case 'p':
> >>+ enabled_port_mask = parse_portmask(optarg);
> >>+ if (enabled_port_mask == 0) {
> >>+ printf("invalid portmask\n");
> >>+ print_usage(prgname);
> >>+ return -1;
> >>+ }
> >>+ break;
> >>+ case 'P':
> >>+ printf("Promiscuous mode selected\n");
> >>+ promiscuous_on = 1;
> >>+ break;
> >>+ case 'u':
> >>+ unprotected_port_mask = parse_portmask(optarg);
> >>+ if (unprotected_port_mask == 0) {
> >>+ printf("invalid unprotected portmask\n");
> >>+ print_usage(prgname);
> >>+ return -1;
> >>+ }
> >>+ break;
> >>+ case 0:
> >>+ if (parse_args_long_options(lgopts, option_index)) {
> >>+ print_usage(prgname);
> >>+ return -1;
> >>+ }
> >>+ break;
> >>+ default:
> >>+ print_usage(prgname);
> >>+ return -1;
> >>+ }
> >>+ }
> >>+
> >>+ if (optind >= 0)
> >>+ argv[optind-1] = prgname;
> >>+
> >>+ ret = optind-1;
> >>+ optind = 0; /* reset getopt lib */
> >>+ return ret;
> >>+}
> >>+
> >>+static void
> >>+print_ethaddr(const char *name, const struct ether_addr *eth_addr)
> >>+{
> >>+ char buf[ETHER_ADDR_FMT_SIZE];
> >>+ ether_format_addr(buf, ETHER_ADDR_FMT_SIZE, eth_addr);
> >>+ printf("%s%s", name, buf);
> >>+}
> >>+
> >>+/* Check the link status of all ports in up to 9s, and print them finally */
> >>+static void
> >>+check_all_ports_link_status(uint8_t port_num, uint32_t port_mask)
> >>+{
> >>+#define CHECK_INTERVAL 100 /* 100ms */
> >>+#define MAX_CHECK_TIME 90 /* 9s (90 * 100ms) in total */
> >>+ uint8_t portid, count, all_ports_up, print_flag = 0;
> >>+ struct rte_eth_link link;
> >>+
> >>+ printf("\nChecking link status");
> >>+ fflush(stdout);
> >>+ for (count = 0; count <= MAX_CHECK_TIME; count++) {
> >>+ all_ports_up = 1;
> >>+ for (portid = 0; portid < port_num; portid++) {
> >>+ if ((port_mask & (1 << portid)) == 0)
> >>+ continue;
> >>+ memset(&link, 0, sizeof(link));
> >>+ rte_eth_link_get_nowait(portid, &link);
> >>+ /* print link status if flag set */
> >>+ if (print_flag == 1) {
> >>+ if (link.link_status)
> >>+ printf("Port %d Link Up - speed %u "
> >>+ "Mbps - %s\n", (uint8_t)portid,
> >>+ (unsigned)link.link_speed,
> >>+ (link.link_duplex == ETH_LINK_FULL_DUPLEX) ?
> >>+ ("full-duplex") : ("half-duplex\n"));
> >>+ else
> >>+ printf("Port %d Link Down\n",
> >>+ (uint8_t)portid);
> >>+ continue;
> >>+ }
> >>+ /* clear all_ports_up flag if any link down */
> >>+ if (link.link_status == 0) {
> >>+ all_ports_up = 0;
> >>+ break;
> >>+ }
> >>+ }
> >>+ /* after finally printing all link status, get out */
> >>+ if (print_flag == 1)
> >>+ break;
> >>+
> >>+ if (all_ports_up == 0) {
> >>+ printf(".");
> >>+ fflush(stdout);
> >>+ rte_delay_ms(CHECK_INTERVAL);
> >>+ }
> >>+
> >>+ /* set the print_flag if all ports up or timeout */
> >>+ if (all_ports_up == 1 || count == (MAX_CHECK_TIME - 1)) {
> >>+ print_flag = 1;
> >>+ printf("done\n");
> >>+ }
> >>+ }
> >>+}
> >>+
> >>+static uint16_t
> >>+find_next_cdev(unsigned cdev_type, uint16_t start_cdev_id)
> >>+{
> >>+ uint16_t cdev_id, cnt;
> >>+ struct rte_cryptodev_info info;
> >>+
> >>+ cnt = rte_cryptodev_count();
> >>+ for (cdev_id = start_cdev_id; cdev_id < cnt; cdev_id++) {
> >>+ rte_cryptodev_info_get(cdev_id, &info);
> >>+ if (info.dev_type == cdev_type)
> >>+ return cdev_id;
> >>+ }
> >>+
> >>+ return -1;
> >>+}
> >>+
> >>+static uint16_t
> >>+find_cdev_socket(int socket_id, unsigned cdev_type)
> >>+{
> >>+ uint16_t cdev_id, cnt;
> >>+ struct rte_cryptodev_info info;
> >>+ int cdev_socket;
> >>+
> >>+ cnt = rte_cryptodev_count();
> >>+ for (cdev_id = 0; cdev_id < cnt; cdev_id++) {
> >>+ rte_cryptodev_info_get(cdev_id, &info);
> >>+ cdev_socket = rte_cryptodev_socket_id(cdev_id);
> >>+ if ((info.dev_type == cdev_type) && (cdev_socket == socket_id))
> >>+ return cdev_id;
> >>+ }
> >>+
> >>+ return -1;
> >>+}
> >>+
> >>+static int
> >>+cryptodevs_init(void)
> >>+{
> >>+ struct rte_cryptodev_config dev_conf;
> >>+ struct rte_cryptodev_qp_conf qp_conf;
> >>+ struct lcore_conf *qconf;
> >>+ uint16_t cnt, lcore_id, cdev_id;
> >>+ int ret, i, socket_id;
> >>+
> >>+ if (cdev_type == RTE_CRYPTODEV_QAT_PMD) {
> >>+ cnt = rte_cryptodev_count_devtype(RTE_CRYPTODEV_QAT_PMD);
> >>+ printf("Found %u QAT devices\n", cnt);
> >>+ if (cnt < nb_lcores)
> >>+ rte_panic("Not enough QAT devices detected, "
> >>+ "need %u (1 per core), found %d\n",
> >>+ nb_lcores, cnt);
> >>+ } else if (cdev_type == RTE_CRYPTODEV_AESNI_MB_PMD) {
> >>+ printf("Initializing %u AESNI vdevs\n", rte_lcore_count());
> >Is it possible to remove PMD specific Initialization code from the
> >application and move driver layer so that new crypto pmd inclusion will
> >not have any change in this place.
>
> Yes, it should be possible now that we have:
> http://dpdk.org/ml/archives/dev/2016-January/032448.html
>
> I also plan to rework the code a bit to use the new cryptodev API:
> http://dpdk.org/ml/archives/dev/2016-January/032445.html
>
> >>+ for (i = 0; i < (int)rte_lcore_count(); i++) {
> >>+ ret = rte_eal_vdev_init(
> >>+ CRYPTODEV_NAME_AESNI_MB_PMD,
> >>+ NULL);
> >>+ if (ret < 0)
> >>+ rte_panic("Failed to create AESNI-MB vdev\n");
> >>+ }
> >>+ } else
> >>+ rte_panic("Need to set cryptodev type option --cdev\n");
> >>+
> >>+ cdev_id = 0;
> >>+ for (lcore_id = 0; lcore_id < RTE_MAX_LCORE; lcore_id++) {
> >>+ if (rte_lcore_is_enabled(lcore_id) == 0)
> >>+ continue;
> >>+
> >>+ socket_id = rte_lcore_to_socket_id(lcore_id);
> >>+ cdev_id = find_next_cdev(cdev_type, cdev_id);
> >>+ qconf = &lcore_conf[lcore_id];
> >>+ qconf->cdev = cdev_id;
> >>+ qconf->cdev_q = 0;
> >>+ if (cdev_type == RTE_CRYPTODEV_AESNI_MB_PMD) {
> >>+ dev_conf.socket_id = socket_id;
> >>+ printf("Initializing AESNI-MB device %u socket %u\n",
> >>+ cdev_id, socket_id);
> >>+ } else {
> >>+ dev_conf.socket_id = rte_cryptodev_socket_id(cdev_id);
> >>+ printf("Initialising QAT device %u\n", cdev_id);
> >>+ }
> >>+
> >>+ dev_conf.nb_queue_pairs = 1;
> >>+ dev_conf.session_mp.nb_objs = CDEV_MP_NB_OBJS;
> >>+ dev_conf.session_mp.cache_size = CDEV_MP_CACHE_SZ;
> >>+
> >>+ if (rte_cryptodev_configure(cdev_id, &dev_conf))
> >>+ rte_panic("Failed to initialize crypodev %u\n",
> >>+ cdev_id);
> >>+
> >>+ qp_conf.nb_descriptors = CDEV_MP_NB_OBJS;
> >>+ if (rte_cryptodev_queue_pair_setup(cdev_id, 0, &qp_conf,
> >>+ dev_conf.socket_id))
> >>+ rte_panic("Failed to setup queue %u for cdev_id %u\n",
> >>+ 0, cdev_id);
> >>+ cdev_id++;
> >>+ }
> >>+
> >>+ return 0;
> >>+}
> >>+
> >>+static void
> >>+port_init(uint8_t portid)
> >>+{
> >>+ struct rte_eth_dev_info dev_info;
> >>+ struct rte_eth_txconf *txconf;
> >>+ uint16_t nb_tx_queue, nb_rx_queue;
> >>+ uint16_t tx_queueid, rx_queueid, queue, lcore_id;
> >>+ int ret, socket_id;
> >>+ struct lcore_conf *qconf;
> >>+ struct ether_addr ethaddr;
> >>+
> >>+ rte_eth_dev_info_get(portid, &dev_info);
> >>+
> >>+ printf("Configuring device port %u:\n", portid);
> >>+
> >>+ rte_eth_macaddr_get(portid, ðaddr);
> >>+ ethaddr_tbl[portid].src = ETHADDR_TO_UINT64(ethaddr);
> >>+ print_ethaddr("Address: ", ðaddr);
> >>+ printf("\n");
> >>+
> >>+ nb_rx_queue = get_port_nb_rx_queues(portid);
> >>+ nb_tx_queue = nb_lcores;
> >>+
> >>+ if (nb_rx_queue > dev_info.max_rx_queues)
> >>+ rte_exit(EXIT_FAILURE, "Error: queue %u not available "
> >>+ "(max rx queue is %u)\n",
> >>+ nb_rx_queue, dev_info.max_rx_queues);
> >>+
> >>+ if (nb_tx_queue > dev_info.max_tx_queues)
> >>+ rte_exit(EXIT_FAILURE, "Error: queue %u not available "
> >>+ "(max tx queue is %u)\n",
> >>+ nb_tx_queue, dev_info.max_tx_queues);
> >>+
> >>+ printf("Creating queues: nb_rx_queue=%d nb_tx_queue=%u...\n",
> >>+ nb_rx_queue, nb_tx_queue);
> >>+
> >>+ ret = rte_eth_dev_configure(portid, nb_rx_queue, nb_tx_queue,
> >>+ &port_conf);
> >>+ if (ret < 0)
> >>+ rte_exit(EXIT_FAILURE, "Cannot configure device: "
> >>+ "err=%d, port=%d\n", ret, portid);
> >>+
> >>+ /* init one TX queue per lcore */
> >>+ tx_queueid = 0;
> >>+ for (lcore_id = 0; lcore_id < RTE_MAX_LCORE; lcore_id++) {
> >>+ if (rte_lcore_is_enabled(lcore_id) == 0)
> >>+ continue;
> >>+
> >>+ if (numa_on)
> >>+ socket_id = (uint8_t)rte_lcore_to_socket_id(lcore_id);
> >>+ else
> >>+ socket_id = 0;
> >>+
> >>+ /* init TX queue */
> >>+ printf("Setup txq=%u,%d,%d\n", lcore_id, tx_queueid, socket_id);
> >>+
> >>+ txconf = &dev_info.default_txconf;
> >>+ txconf->txq_flags = 0;
> >>+
> >>+ ret = rte_eth_tx_queue_setup(portid, tx_queueid, nb_txd,
> >>+ socket_id, txconf);
> >>+ if (ret < 0)
> >>+ rte_exit(EXIT_FAILURE, "rte_eth_tx_queue_setup: "
> >>+ "err=%d, port=%d\n", ret, portid);
> >>+
> >>+ qconf = &lcore_conf[lcore_id];
> >>+ qconf->tx_queue_id[portid] = tx_queueid;
> >>+ tx_queueid++;
> >>+
> >>+ /* init RX queues */
> >>+ for (queue = 0; queue < qconf->nb_rx_queue; ++queue) {
> >>+ if (portid != qconf->rx_queue_list[queue].port_id)
> >>+ continue;
> >>+
> >>+ rx_queueid = qconf->rx_queue_list[queue].queue_id;
> >>+
> >>+ printf("Setup rxq=%d,%d,%d\n", portid, rx_queueid,
> >>+ socket_id);
> >>+
> >>+ ret = rte_eth_rx_queue_setup(portid, rx_queueid,
> >>+ nb_rxd, socket_id, NULL,
> >>+ socket_ctx[socket_id].mbuf_pool);
> >>+ if (ret < 0)
> >>+ rte_exit(EXIT_FAILURE,
> >>+ "rte_eth_rx_queue_setup: err=%d, "
> >>+ "port=%d\n", ret, portid);
> >>+ }
> >>+ }
> >>+ printf("\n");
> >>+}
> >>+
> >>+static void
> >>+pool_init(struct socket_ctx *ctx, int socket_id, unsigned nb_mbuf)
> >>+{
> >>+ char s[64];
> >>+
> >>+ snprintf(s, sizeof(s), "mbuf_pool_%d", socket_id);
> >>+ ctx->mbuf_pool = rte_pktmbuf_pool_create(s, nb_mbuf,
> >>+ MEMPOOL_CACHE_SIZE, ipsec_metadata_size(),
> >>+ RTE_MBUF_DEFAULT_BUF_SIZE,
> >>+ socket_id);
> >>+ if (ctx->mbuf_pool == NULL)
> >>+ rte_exit(EXIT_FAILURE, "Cannot init mbuf pool on socket %d\n",
> >>+ socket_id);
> >>+ else
> >>+ printf("Allocated mbuf pool on socket %d\n", socket_id);
> >>+}
> >>+
> >>+int
> >>+main(int argc, char **argv)
> >>+{
> >>+ int ret;
> >>+ unsigned lcore_id, nb_ports;
> >>+ uint8_t portid, socket_id;
> >>+
> >>+ /* init EAL */
> >>+ ret = rte_eal_init(argc, argv);
> >>+ if (ret < 0)
> >>+ rte_exit(EXIT_FAILURE, "Invalid EAL parameters\n");
> >>+ argc -= ret;
> >>+ argv += ret;
> >>+
> >>+ /* parse application arguments (after the EAL ones) */
> >>+ ret = parse_args(argc, argv);
> >>+ if (ret < 0)
> >>+ rte_exit(EXIT_FAILURE, "Invalid parameters\n");
> >>+
> >>+ if (ep < 0)
> >>+ rte_exit(EXIT_FAILURE, "need to choose either EP0 or EP1\n");
> >>+
> >>+ if ((unprotected_port_mask & enabled_port_mask) !=
> >>+ unprotected_port_mask)
> >>+ rte_exit(EXIT_FAILURE, "Invalid unprotected portmask 0x%x\n",
> >>+ unprotected_port_mask);
> >>+
> >>+ nb_ports = rte_eth_dev_count();
> >>+ if (nb_ports > RTE_MAX_ETHPORTS)
> >>+ nb_ports = RTE_MAX_ETHPORTS;
> >>+
> >>+ if (check_params() < 0)
> >>+ rte_exit(EXIT_FAILURE, "check_params failed\n");
> >>+
> >>+ ret = init_lcore_rx_queues();
> >>+ if (ret < 0)
> >>+ rte_exit(EXIT_FAILURE, "init_lcore_rx_queues failed\n");
> >>+
> >>+ nb_lcores = rte_lcore_count();
> >>+
> >>+ cryptodevs_init();
> >>+
> >>+ /* Replicate each contex per socket */
> >>+ for (lcore_id = 0; lcore_id < RTE_MAX_LCORE; lcore_id++) {
> >>+ if (rte_lcore_is_enabled(lcore_id) == 0)
> >>+ continue;
> >>+
> >>+ if (numa_on)
> >>+ socket_id = (uint8_t)rte_lcore_to_socket_id(lcore_id);
> >>+ else
> >>+ socket_id = 0;
> >>+
> >>+ if (socket_ctx[socket_id].mbuf_pool)
> >>+ continue;
> >>+
> >>+ sa_init(&socket_ctx[socket_id], socket_id, ep,
> >>+ find_cdev_socket(socket_id, cdev_type));
> >>+
> >>+ sp_init(&socket_ctx[socket_id], socket_id, ep);
> >>+
> >>+ rt_init(&socket_ctx[socket_id], socket_id, ep);
> >>+
> >>+ pool_init(&socket_ctx[socket_id], socket_id, NB_MBUF);
> >>+ }
> >>+
> >>+ for (portid = 0; portid < nb_ports; portid++) {
> >>+ if ((enabled_port_mask & (1 << portid)) == 0)
> >>+ continue;
> >>+
> >>+ port_init(portid);
> >>+ }
> >>+
> >>+ /* start ports */
> >>+ for (portid = 0; portid < nb_ports; portid++) {
> >>+ if ((enabled_port_mask & (1 << portid)) == 0)
> >>+ continue;
> >>+
> >>+ /* Start device */
> >>+ ret = rte_eth_dev_start(portid);
> >>+ if (ret < 0)
> >>+ rte_exit(EXIT_FAILURE, "rte_eth_dev_start: "
> >>+ "err=%d, port=%d\n", ret, portid);
> >>+ /*
> >>+ * If enabled, put device in promiscuous mode.
> >>+ * This allows IO forwarding mode to forward packets
> >>+ * to itself through 2 cross-connected ports of the
> >>+ * target machine.
> >>+ */
> >>+ if (promiscuous_on)
> >>+ rte_eth_promiscuous_enable(portid);
> >>+ }
> >>+
> >>+ check_all_ports_link_status((uint8_t)nb_ports, enabled_port_mask);
> >>+
> >>+ /* launch per-lcore init on every lcore */
> >>+ rte_eal_mp_remote_launch(main_loop, NULL, CALL_MASTER);
> >>+ RTE_LCORE_FOREACH_SLAVE(lcore_id) {
> >>+ if (rte_eal_wait_lcore(lcore_id) < 0)
> >>+ return -1;
> >>+ }
> >>+
> >>+ return 0;
> >>+}
> >>diff --git a/examples/ipsec-secgw/ipsec.c b/examples/ipsec-secgw/ipsec.c
> >>new file mode 100644
> >>index 0000000..83e3326
> >>--- /dev/null
> >>+++ b/examples/ipsec-secgw/ipsec.c
> >>@@ -0,0 +1,138 @@
> >>+/*-
> >>+ * BSD LICENSE
> >>+ *
> >>+ * Copyright(c) 2010-2016 Intel Corporation. All rights reserved.
> >>+ * 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.
> >>+ */
> >>+
> >>+#include <netinet/in.h>
> >>+#include <netinet/ip.h>
> >>+
> >>+#include <rte_branch_prediction.h>
> >>+#include <rte_log.h>
> >>+#include <rte_crypto.h>
> >>+#include <rte_cryptodev.h>
> >>+#include <rte_mbuf.h>
> >>+
> >>+#include "ipsec.h"
> >>+
> >>+static inline uint16_t
> >>+ipsec_processing(uint8_t cdev_id, uint16_t qp_id, struct rte_mbuf *pkts[],
> >>+ struct ipsec_sa *sas[], uint16_t nb_pkts, uint16_t max_pkts)
> >>+{
> >>+ int ret, i, j;
> >>+ struct ipsec_mbuf_metadata *priv;
> >>+ struct rte_mbuf_offload *ol;
> >>+ struct ipsec_sa *sa;
> >>+
> >>+ j = 0;
> >>+ for (i = 0; i < nb_pkts; i++) {
> >>+ rte_prefetch0(sas[i]);
> >>+ rte_prefetch0(pkts[i]);
> >>+
> >>+ priv = get_priv(pkts[i]);
> >>+ ol = &priv->ol;
> >>+ sa = sas[i];
> >>+ priv->sa = sa;
> >>+
> >>+ IPSEC_ASSERT(sa != NULL);
> >>+
> >>+ __rte_pktmbuf_offload_reset(ol, RTE_PKTMBUF_OL_CRYPTO);
> >>+
> >>+ rte_crypto_op_attach_session(&ol->op.crypto,
> >>+ sa->crypto_session);
> >>+
> >>+ pkts[i]->offload_ops = ol;
> >>+
> >>+ ret = sa->pre_crypto(pkts[i], sa, &ol->op.crypto);
> >>+ if (unlikely(ret))
> >>+ rte_pktmbuf_free(pkts[i]);
> >>+ else
> >>+ pkts[j++] = pkts[i];
> >>+ }
> >>+ nb_pkts = j;
> >>+
> >>+ ret = rte_cryptodev_enqueue_burst(cdev_id, qp_id, pkts, nb_pkts);
> >>+ if (ret < nb_pkts) {
> >>+ IPSEC_LOG(DEBUG, IPSEC, "Cryptodev %u queue %u:"
> >>+ " enqueued %u packets (out of %u)\n",
> >>+ cdev_id, qp_id, ret, nb_pkts);
> >>+ for (i = ret; i < nb_pkts; i++)
> >>+ rte_pktmbuf_free(pkts[i]);
> >>+ }
> >>+
> >>+ nb_pkts = rte_cryptodev_dequeue_burst(cdev_id, qp_id, pkts, max_pkts);
> >>+ if ((nb_pkts != 0) && (nb_pkts < max_pkts))
> >>+ IPSEC_LOG(DEBUG, IPSEC, "Cryptodev %u queue %u:"
> >>+ " dequeued %u packets (out of %u)\n",
> >>+ cdev_id, qp_id, nb_pkts, max_pkts);
> >>+
> >>+ j = 0;
> >>+ for (i = 0; i < nb_pkts; i++) {
> >>+ rte_prefetch0(pkts[i]);
> >>+
> >>+ priv = get_priv(pkts[i]);
> >>+ ol = &priv->ol;
> >>+ sa = priv->sa;
> >>+ rte_prefetch0(sa);
> >>+
> >>+ IPSEC_ASSERT(sa != NULL);
> >>+
> >>+ ret = sa->post_crypto(pkts[i], sa, &ol->op.crypto);
> >>+ if (unlikely(ret))
> >>+ rte_pktmbuf_free(pkts[i]);
> >>+ else
> >>+ pkts[j++] = pkts[i];
> >>+ }
> >>+
> >>+ /* return packets */
> >>+ return j;
> >>+}
> >>+
> >>+uint16_t
> >>+ipsec_inbound(struct ipsec_ctx *ctx, struct rte_mbuf *pkts[],
> >>+ uint16_t nb_pkts, uint16_t len)
> >>+{
> >>+ struct ipsec_sa *sas[nb_pkts];
> >>+
> >>+ inbound_sa_lookup(ctx->sa_ctx, pkts, sas, nb_pkts);
> >>+
> >>+ return ipsec_processing(ctx->cdev, ctx->queue, pkts, sas, nb_pkts, len);
> >>+}
> >>+
> >>+uint16_t
> >>+ipsec_outbound(struct ipsec_ctx *ctx, struct rte_mbuf *pkts[],
> >>+ uint32_t sa_idx[], uint16_t nb_pkts, uint16_t len)
> >>+{
> >>+ struct ipsec_sa *sas[nb_pkts];
> >>+
> >>+ outbound_sa_lookup(ctx->sa_ctx, sa_idx, sas, nb_pkts);
> >>+
> >>+ return ipsec_processing(ctx->cdev, ctx->queue, pkts, sas, nb_pkts, len);
> >>+}
> >>diff --git a/examples/ipsec-secgw/ipsec.h b/examples/ipsec-secgw/ipsec.h
> >>new file mode 100644
> >>index 0000000..abf9d5f
> >>--- /dev/null
> >>+++ b/examples/ipsec-secgw/ipsec.h
> >>@@ -0,0 +1,184 @@
> >>+/*-
> >>+ * BSD LICENSE
> >>+ *
> >>+ * Copyright(c) 2010-2016 Intel Corporation. All rights reserved.
> >>+ * 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.
> >>+ */
> >>+
> >>+#ifndef __IPSEC_H__
> >>+#define __IPSEC_H__
> >>+
> >>+#include <stdint.h>
> >>+#include <netinet/in.h>
> >>+#include <netinet/ip.h>
> >>+
> >>+#include <rte_mbuf_offload.h>
> >>+#include <rte_byteorder.h>
> >>+#include <rte_ip.h>
> >>+
> >>+#define RTE_LOGTYPE_IPSEC RTE_LOGTYPE_USER1
> >>+#define RTE_LOGTYPE_IPSEC_ESP RTE_LOGTYPE_USER2
> >>+#define RTE_LOGTYPE_IPSEC_IPIP RTE_LOGTYPE_USER3
> >>+
> >>+#ifdef IPSEC_DEBUG
> >>+#define IPSEC_ASSERT(exp) \
> >>+if (!(exp)) { \
> >>+ rte_panic("line%d\tassert \"" #exp "\" failed\n", __LINE__); \
> >>+}
> >>+
> >>+#define IPSEC_LOG RTE_LOG
> >>+#else
> >>+#define IPSEC_ASSERT(exp) do {} while (0)
> >>+#define IPSEC_LOG(...) do {} while (0)
> >>+#endif /* IPSEC_DEBUG */
> >>+
> >>+#define MAX_DIGEST_SIZE 32 /* Bytes -- 256 bits */
> >>+
> >>+#define uint32_t_to_char(ip, a, b, c, d) do {\
> >>+ *a = (unsigned char)(ip >> 24 & 0xff);\
> >>+ *b = (unsigned char)(ip >> 16 & 0xff);\
> >>+ *c = (unsigned char)(ip >> 8 & 0xff);\
> >>+ *d = (unsigned char)(ip & 0xff);\
> >>+ } while (0)
> >>+
> >>+#define DEFAULT_MAX_CATEGORIES 1
> >>+
> >>+#define IPSEC_SA_MAX_ENTRIES (64) /* must be power of 2, max 2 power 30 */
> >>+#define SPI2IDX(spi) (spi & (IPSEC_SA_MAX_ENTRIES - 1))
> >>+#define INVALID_SPI (0)
> >>+
> >>+#define DISCARD (0x80000000)
> >>+#define BYPASS (0x40000000)
> >>+#define PROTECT_MASK (0x3fffffff)
> >>+#define PROTECT(sa_idx) (SPI2IDX(sa_idx) & PROTECT_MASK) /* SA idx 30 bits */
> >>+
> >>+#define IPSEC_XFORM_MAX 2
> >>+
> >>+struct rte_crypto_xform;
> >>+struct ipsec_xform;
> >>+struct rte_cryptodev_session;
> >>+struct rte_mbuf;
> >>+
> >>+struct replay {
> >>+ uint32_t seq_h;
> >>+};
> >>+
> >>+struct lifetime {
> >>+ uint64_t bytes;
> >>+ uint64_t seconds;
> >>+};
> >>+
> >>+struct ipsec_sa;
> >>+
> >>+typedef int (*ipsec_xform_fn)(struct rte_mbuf *m, struct ipsec_sa *sa,
> >>+ struct rte_crypto_op *cop);
> >>+
> >>+struct ipsec_sa {
> >>+ uint32_t spi;
> >>+ uint32_t seq;
> >>+ uint32_t src;
> >>+ uint32_t dst;
> >>+ struct rte_cryptodev_session *crypto_session;
> >>+ struct rte_crypto_xform *xforms;
> >>+ enum rte_crypto_cipher_algorithm cipher_algo;
> >>+ enum rte_crypto_auth_algorithm auth_algo;
> >>+ uint16_t digest_len;
> >>+ uint16_t iv_len;
> >>+ uint16_t block_size;
> >>+ ipsec_xform_fn pre_crypto;
> >>+ ipsec_xform_fn post_crypto;
> >>+ uint32_t flags;
> >>+ /* does not apply if no automated SA management (IKE) support */
> >>+ struct lifetime current;
> >>+ struct lifetime soft;
> >>+ struct lifetime hard;
> >>+ struct replay replay;
> >>+} __rte_cache_aligned;
> >>+
> >>+struct ipsec_mbuf_metadata {
> >>+ struct ipsec_sa *sa;
> >>+ struct rte_mbuf_offload ol;
> >>+};
> >>+
> >>+struct ipsec_ctx {
> >>+ uint16_t cdev;
> >>+ uint16_t queue;
> >>+ struct sa_ctx *sa_ctx;
> >>+};
> >>+
> >>+struct socket_ctx {
> >>+ struct sa_ctx *sa_ipv4_in;
> >>+ struct sa_ctx *sa_ipv4_out;
> >>+ struct sp_ctx *sp_ipv4_in;
> >>+ struct sp_ctx *sp_ipv4_out;
> >>+ struct rt_ctx *rt_ipv4;
> >>+ struct rte_mempool *mbuf_pool;
> >>+};
> >>+
> >>+uint16_t
> >>+ipsec_inbound(struct ipsec_ctx *ctx, struct rte_mbuf *pkts[],
> >>+ uint16_t nb_pkts, uint16_t len);
> >>+
> >>+uint16_t
> >>+ipsec_outbound(struct ipsec_ctx *ctx, struct rte_mbuf *pkts[],
> >>+ uint32_t sa_idx[], uint16_t nb_pkts, uint16_t len);
> >>+
> >>+static inline uint16_t
> >>+ipsec_metadata_size(void)
> >>+{
> >>+ return sizeof(struct ipsec_mbuf_metadata);
> >>+}
> >>+
> >>+static inline struct ipsec_mbuf_metadata *
> >>+get_priv(struct rte_mbuf *m)
> >>+{
> >>+ return RTE_PTR_ADD(m, sizeof(struct rte_mbuf));
> >>+}
> >>+
> >>+int
> >>+inbound_sa_check(struct sa_ctx *sa_ctx, struct rte_mbuf *m, uint32_t sa_idx);
> >>+
> >>+void
> >>+inbound_sa_lookup(struct sa_ctx *sa_ctx, struct rte_mbuf *pkts[],
> >>+ struct ipsec_sa *sa[], uint16_t nb_pkts);
> >>+
> >>+void
> >>+outbound_sa_lookup(struct sa_ctx *sa_ctx, uint32_t sa_idx[],
> >>+ struct ipsec_sa *sa[], uint16_t nb_pkts);
> >>+
> >>+void
> >>+sp_init(struct socket_ctx *ctx, int socket_id, unsigned ep);
> >>+
> >>+void
> >>+sa_init(struct socket_ctx *ctx, int socket_id, unsigned ep, uint16_t cdev_id);
> >>+
> >>+void
> >>+rt_init(struct socket_ctx *ctx, int socket_id, unsigned ep);
> >>+
> >>+#endif /* __IPSEC_H__ */
> >>diff --git a/examples/ipsec-secgw/rt.c b/examples/ipsec-secgw/rt.c
> >>new file mode 100644
> >>index 0000000..82064b2
> >>--- /dev/null
> >>+++ b/examples/ipsec-secgw/rt.c
> >>@@ -0,0 +1,131 @@
> >>+/*-
> >>+ * BSD LICENSE
> >>+ *
> >>+ * Copyright(c) 2010-2016 Intel Corporation. All rights reserved.
> >>+ * 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.
> >>+ */
> >>+
> >>+/*
> >>+ * Routing Table (RT)
> >>+ */
> >>+#include <rte_lpm.h>
> >>+#include <rte_errno.h>
> >>+
> >>+#include "ipsec.h"
> >>+
> >>+#define RT_IPV4_MAX_RULES 64
> >>+
> >>+struct ipv4_route {
> >>+ uint32_t ip;
> >>+ uint8_t depth;
> >>+ uint8_t if_out;
> >>+};
> >>+
> >>+/* In the default routing table we have:
> >>+ * ep0 protected ports 0 and 1, and unprotected ports 2 and 3.
> >>+ */
> >>+static struct ipv4_route rt_ipv4_ep0[] = {
> >>+ { IPv4(172, 16, 2, 5), 32, 0 },
> >>+ { IPv4(172, 16, 2, 6), 32, 0 },
> >>+ { IPv4(172, 16, 2, 7), 32, 1 },
> >>+ { IPv4(172, 16, 2, 8), 32, 1 },
> >>+
> >>+ { IPv4(192, 168, 115, 0), 24, 2 },
> >>+ { IPv4(192, 168, 116, 0), 24, 2 },
> >>+ { IPv4(192, 168, 117, 0), 24, 3 },
> >>+ { IPv4(192, 168, 118, 0), 24, 3 }
> >>+};
> >>+
> >>+/* In the default routing table we have:
> >>+ * ep1 protected ports 0 and 1, and unprotected ports 2 and 3.
> >>+ */
> >>+static struct ipv4_route rt_ipv4_ep1[] = {
> >>+ { IPv4(172, 16, 1, 5), 32, 2 },
> >>+ { IPv4(172, 16, 1, 6), 32, 2 },
> >>+ { IPv4(172, 16, 1, 7), 32, 3 },
> >>+ { IPv4(172, 16, 1, 8), 32, 3 },
> >>+
> >>+ { IPv4(192, 168, 105, 0), 24, 0 },
> >>+ { IPv4(192, 168, 106, 0), 24, 0 },
> >>+ { IPv4(192, 168, 107, 0), 24, 1 },
> >>+ { IPv4(192, 168, 108, 0), 24, 1 }
> >>+};
> >>+
> >>+void
> >>+rt_init(struct socket_ctx *ctx, int socket_id, unsigned ep)
> >>+{
> >>+ char name[PATH_MAX];
> >>+ unsigned i;
> >>+ int ret;
> >>+ struct rte_lpm *lpm;
> >>+ struct ipv4_route *rt;
> >>+ char a, b, c, d;
> >>+ unsigned nb_routes;
> >>+
> >>+ if (ctx == NULL)
> >>+ rte_exit(EXIT_FAILURE, "NULL context.\n");
> >>+
> >>+ if (ctx->rt_ipv4 != NULL)
> >>+ rte_exit(EXIT_FAILURE, "Routing Table for socket %u already "
> >>+ "initialized\n", socket_id);
> >>+
> >>+ printf("Creating Routing Table (RT) context with %u max routes\n",
> >>+ RT_IPV4_MAX_RULES);
> >>+
> >>+ if (ep == 0) {
> >>+ rt = rt_ipv4_ep0;
> >>+ nb_routes = RTE_DIM(rt_ipv4_ep0);
> >>+ } else if (ep == 1) {
> >>+ rt = rt_ipv4_ep1;
> >>+ nb_routes = RTE_DIM(rt_ipv4_ep1);
> >>+ } else
> >>+ rte_exit(EXIT_FAILURE, "Invalid EP value %u. Only 0 or 1 "
> >>+ "supported.\n", ep);
> >>+
> >>+ /* create the LPM table */
> >>+ snprintf(name, sizeof(name), "%s_%u", "rt_ipv4", socket_id);
> >>+ lpm = rte_lpm_create(name, socket_id, RT_IPV4_MAX_RULES, 0);
> >>+ if (lpm == NULL)
> >>+ rte_exit(EXIT_FAILURE, "Unable to create LPM table "
> >>+ "on socket %d\n", socket_id);
> >>+
> >>+ /* populate the LPM table */
> >>+ for (i = 0; i < nb_routes; i++) {
> >>+ ret = rte_lpm_add(lpm, rt[i].ip, rt[i].depth, rt[i].if_out);
> >>+ if (ret < 0)
> >>+ rte_exit(EXIT_FAILURE, "Unable to add entry num %u to "
> >>+ "LPM table on socket %d\n", i, socket_id);
> >>+
> >>+ uint32_t_to_char(rt[i].ip, &a, &b, &c, &d);
> >>+ printf("LPM: Adding route %hhu.%hhu.%hhu.%hhu/%hhu (%hhu)\n",
> >>+ a, b, c, d, rt[i].depth, rt[i].if_out);
> >>+ }
> >>+
> >>+ ctx->rt_ipv4 = (struct rt_ctx *)lpm;
> >>+}
> >>diff --git a/examples/ipsec-secgw/sa.c b/examples/ipsec-secgw/sa.c
> >>new file mode 100644
> >>index 0000000..5ea7592
> >>--- /dev/null
> >>+++ b/examples/ipsec-secgw/sa.c
> >>@@ -0,0 +1,391 @@
> >>+/*-
> >>+ * BSD LICENSE
> >>+ *
> >>+ * Copyright(c) 2010-2016 Intel Corporation. All rights reserved.
> >>+ * 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.
> >>+ */
> >>+
> >>+/*
> >>+ * Security Associations
> >>+ */
> >>+#include <netinet/ip.h>
> >>+
> >>+#include <rte_memzone.h>
> >>+#include <rte_crypto.h>
> >>+#include <rte_cryptodev.h>
> >>+#include <rte_byteorder.h>
> >>+#include <rte_errno.h>
> >>+
> >>+#include "ipsec.h"
> >>+#include "esp.h"
> >>+
> >>+/* SAs EP0 Outbound */
> >>+const struct ipsec_sa sa_ep0_out[] = {
> >>+ { 5, 0, IPv4(172, 16, 1, 5), IPv4(172, 16, 2, 5),
> >>+ NULL, NULL,
> >>+ RTE_CRYPTO_CIPHER_AES_CBC, RTE_CRYPTO_AUTH_SHA1_HMAC,
> >>+ 12, 16, 16,
> >>+ esp4_tunnel_outbound_pre_crypto,
> >>+ esp4_tunnel_outbound_post_crypto,
> >>+ 0, { 0, 0 }, { 0, 0 }, { 0, 0 }, { 0 } },
> >>+ { 6, 0, IPv4(172, 16, 1, 6), IPv4(172, 16, 2, 6),
> >>+ NULL, NULL,
> >>+ RTE_CRYPTO_CIPHER_AES_CBC, RTE_CRYPTO_AUTH_SHA1_HMAC,
> >>+ 12, 16, 16,
> >>+ esp4_tunnel_outbound_pre_crypto,
> >>+ esp4_tunnel_outbound_post_crypto,
> >>+ 0, { 0, 0 }, { 0, 0 }, { 0, 0 }, { 0 } },
> >>+ { 7, 0, IPv4(172, 16, 1, 7), IPv4(172, 16, 2, 7),
> >>+ NULL, NULL,
> >>+ RTE_CRYPTO_CIPHER_AES_CBC, RTE_CRYPTO_AUTH_SHA1_HMAC,
> >>+ 12, 16, 16,
> >>+ esp4_tunnel_outbound_pre_crypto,
> >>+ esp4_tunnel_outbound_post_crypto,
> >>+ 0, { 0, 0 }, { 0, 0 }, { 0, 0 }, { 0 } },
> >>+ { 8, 0, IPv4(172, 16, 1, 8), IPv4(172, 16, 2, 8),
> >>+ NULL, NULL,
> >>+ RTE_CRYPTO_CIPHER_AES_CBC, RTE_CRYPTO_AUTH_SHA1_HMAC,
> >IMO, a SA with NULL crypto (rfc2410) operation will be useful for ESP
> >packet processing debugging and measuring the performance impact with
> >the crypto block in fast path.
> >
>
> I couldn't agree more.
>
> I'll add support for NULL crypto in the app with dependency on previously
> mentioned cryptodev capabilities patch set and:
> http://dpdk.org/ml/archives/dev/2016-January/032458.html
>
> Sergio
>
> >>+ 12, 16, 16,
> >>+ esp4_tunnel_outbound_pre_crypto,
> >>+ esp4_tunnel_outbound_post_crypto,
> >>+ 0, { 0, 0 }, { 0, 0 }, { 0, 0 }, { 0 } }
> >>+};
> >>+
> >>+/* SAs EP0 Inbound */
> >>+const struct ipsec_sa sa_ep0_in[] = {
> >>+ { 5, 0, IPv4(172, 16, 2, 5), IPv4(172, 16, 1, 5),
> >>+ NULL, NULL,
> >>+ RTE_CRYPTO_CIPHER_AES_CBC, RTE_CRYPTO_AUTH_SHA1_HMAC,
> >>+ 12, 16, 16,
> >>+ esp4_tunnel_inbound_pre_crypto,
> >>+ esp4_tunnel_inbound_post_crypto,
> >>+ 0, { 0, 0 }, { 0, 0 }, { 0, 0 }, { 0 } },
> >>+ { 6, 0, IPv4(172, 16, 2, 6), IPv4(172, 16, 1, 6),
> >>+ NULL, NULL,
> >>+ RTE_CRYPTO_CIPHER_AES_CBC, RTE_CRYPTO_AUTH_SHA1_HMAC,
> >>+ 12, 16, 16,
> >>+ esp4_tunnel_inbound_pre_crypto,
> >>+ esp4_tunnel_inbound_post_crypto,
> >>+ 0, { 0, 0 }, { 0, 0 }, { 0, 0 }, { 0 } },
> >>+ { 7, 0, IPv4(172, 16, 2, 7), IPv4(172, 16, 1, 7),
> >>+ NULL, NULL,
> >>+ RTE_CRYPTO_CIPHER_AES_CBC, RTE_CRYPTO_AUTH_SHA1_HMAC,
> >>+ 12, 16, 16,
> >>+ esp4_tunnel_inbound_pre_crypto,
> >>+ esp4_tunnel_inbound_post_crypto,
> >>+ 0, { 0, 0 }, { 0, 0 }, { 0, 0 }, { 0 } },
> >>+ { 8, 0, IPv4(172, 16, 2, 8), IPv4(172, 16, 1, 8),
> >>+ NULL, NULL,
> >>+ RTE_CRYPTO_CIPHER_AES_CBC, RTE_CRYPTO_AUTH_SHA1_HMAC,
> >>+ 12, 16, 16,
> >>+ esp4_tunnel_inbound_pre_crypto,
> >>+ esp4_tunnel_inbound_post_crypto,
> >>+ 0, { 0, 0 }, { 0, 0 }, { 0, 0 }, { 0 } }
> >>+};
> >>+
> >>+/* SAs EP1 Outbound */
> >>+const struct ipsec_sa sa_ep1_out[] = {
> >>+ { 5, 0, IPv4(172, 16, 2, 5), IPv4(172, 16, 1, 5),
> >>+ NULL, NULL,
> >>+ RTE_CRYPTO_CIPHER_AES_CBC, RTE_CRYPTO_AUTH_SHA1_HMAC,
> >>+ 12, 16, 16,
> >>+ esp4_tunnel_outbound_pre_crypto,
> >>+ esp4_tunnel_outbound_post_crypto,
> >>+ 0, { 0, 0 }, { 0, 0 }, { 0, 0 }, { 0 } },
> >>+ { 6, 0, IPv4(172, 16, 2, 6), IPv4(172, 16, 1, 6),
> >>+ NULL, NULL,
> >>+ RTE_CRYPTO_CIPHER_AES_CBC, RTE_CRYPTO_AUTH_SHA1_HMAC,
> >>+ 12, 16, 16,
> >>+ esp4_tunnel_outbound_pre_crypto,
> >>+ esp4_tunnel_outbound_post_crypto,
> >>+ 0, { 0, 0 }, { 0, 0 }, { 0, 0 }, { 0 } },
> >>+ { 7, 0, IPv4(172, 16, 2, 7), IPv4(172, 16, 1, 7),
> >>+ NULL, NULL,
> >>+ RTE_CRYPTO_CIPHER_AES_CBC, RTE_CRYPTO_AUTH_SHA1_HMAC,
> >>+ 12, 16, 16,
> >>+ esp4_tunnel_outbound_pre_crypto,
> >>+ esp4_tunnel_outbound_post_crypto,
> >>+ 0, { 0, 0 }, { 0, 0 }, { 0, 0 }, { 0 } },
> >>+ { 8, 0, IPv4(172, 16, 2, 8), IPv4(172, 16, 1, 8),
> >>+ NULL, NULL,
> >>+ RTE_CRYPTO_CIPHER_AES_CBC, RTE_CRYPTO_AUTH_SHA1_HMAC,
> >>+ 12, 16, 16,
> >>+ esp4_tunnel_outbound_pre_crypto,
> >>+ esp4_tunnel_outbound_post_crypto,
> >>+ 0, { 0, 0 }, { 0, 0 }, { 0, 0 }, { 0 } }
> >>+};
> >>+
> >>+/* SAs EP1 Inbound */
> >>+const struct ipsec_sa sa_ep1_in[] = {
> >>+ { 5, 0, IPv4(172, 16, 1, 5), IPv4(172, 16, 2, 5),
> >>+ NULL, NULL,
> >>+ RTE_CRYPTO_CIPHER_AES_CBC, RTE_CRYPTO_AUTH_SHA1_HMAC,
> >>+ 12, 16, 16,
> >>+ esp4_tunnel_inbound_pre_crypto,
> >>+ esp4_tunnel_inbound_post_crypto,
> >>+ 0, { 0, 0 }, { 0, 0 }, { 0, 0 }, { 0 } },
> >>+ { 6, 0, IPv4(172, 16, 1, 6), IPv4(172, 16, 2, 6),
> >>+ NULL, NULL,
> >>+ RTE_CRYPTO_CIPHER_AES_CBC, RTE_CRYPTO_AUTH_SHA1_HMAC,
> >>+ 12, 16, 16,
> >>+ esp4_tunnel_inbound_pre_crypto,
> >>+ esp4_tunnel_inbound_post_crypto,
> >>+ 0, { 0, 0 }, { 0, 0 }, { 0, 0 }, { 0 } },
> >>+ { 7, 0, IPv4(172, 16, 1, 7), IPv4(172, 16, 2, 7),
> >>+ NULL, NULL,
> >>+ RTE_CRYPTO_CIPHER_AES_CBC, RTE_CRYPTO_AUTH_SHA1_HMAC,
> >>+ 12, 16, 16,
> >>+ esp4_tunnel_inbound_pre_crypto,
> >>+ esp4_tunnel_inbound_post_crypto,
> >>+ 0, { 0, 0 }, { 0, 0 }, { 0, 0 }, { 0 } },
> >>+ { 8, 0, IPv4(172, 16, 1, 8), IPv4(172, 16, 2, 8),
> >>+ NULL, NULL,
> >>+ RTE_CRYPTO_CIPHER_AES_CBC, RTE_CRYPTO_AUTH_SHA1_HMAC,
> >>+ 12, 16, 16,
> >>+ esp4_tunnel_inbound_pre_crypto,
> >>+ esp4_tunnel_inbound_post_crypto,
> >>+ 0, { 0, 0 }, { 0, 0 }, { 0, 0 }, { 0 } }
> >>+};
> >>+
> >>+static uint8_t cipher_key[256] = "sixteenbytes key";
> >>+
> >>+/* AES CBC xform */
> >>+const struct rte_crypto_xform aescbc_enc_xf = {
> >>+ NULL,
> >>+ RTE_CRYPTO_XFORM_CIPHER,
> >>+ .cipher = { RTE_CRYPTO_CIPHER_OP_ENCRYPT, RTE_CRYPTO_CIPHER_AES_CBC,
> >>+ .key = { cipher_key, 0, 16 } }
> >>+};
> >>+
> >>+const struct rte_crypto_xform aescbc_dec_xf = {
> >>+ NULL,
> >>+ RTE_CRYPTO_XFORM_CIPHER,
> >>+ .cipher = { RTE_CRYPTO_CIPHER_OP_DECRYPT, RTE_CRYPTO_CIPHER_AES_CBC,
> >>+ .key = { cipher_key, 0, 16 } }
> >>+};
> >>+
> >>+static uint8_t auth_key[256] = "twentybytes hash key";
> >>+
> >>+/* SHA1 HMAC xform */
> >>+const struct rte_crypto_xform sha1hmac_gen_xf = {
> >>+ NULL,
> >>+ RTE_CRYPTO_XFORM_AUTH,
> >>+ .auth = { RTE_CRYPTO_AUTH_OP_GENERATE, RTE_CRYPTO_AUTH_SHA1_HMAC,
> >>+ .key = { auth_key, 0, 20 }, 12, 0 }
> >>+};
> >>+
> >>+const struct rte_crypto_xform sha1hmac_verify_xf = {
> >>+ NULL,
> >>+ RTE_CRYPTO_XFORM_AUTH,
> >>+ .auth = { RTE_CRYPTO_AUTH_OP_VERIFY, RTE_CRYPTO_AUTH_SHA1_HMAC,
> >>+ .key = { auth_key, 0, 20 }, 12, 0 }
> >>+};
> >>+
> >>+struct sa_ctx {
> >>+ struct ipsec_sa sa[IPSEC_SA_MAX_ENTRIES];
> >>+ struct {
> >>+ struct rte_crypto_xform a;
> >>+ struct rte_crypto_xform b;
> >>+ } xf[IPSEC_SA_MAX_ENTRIES];
> >>+};
> >>+
> >>+static struct sa_ctx *
> >>+sa_ipv4_create(const char *name, int socket_id)
> >>+{
> >>+ char s[PATH_MAX];
> >>+ struct sa_ctx *sa_ctx;
> >>+ unsigned mz_size;
> >>+ const struct rte_memzone *mz;
> >>+
> >>+ snprintf(s, sizeof(s), "%s_%u", name, socket_id);
> >>+
> >>+ /* Create SA array table */
> >>+ printf("Creating SA context with %u maximum entries\n",
> >>+ IPSEC_SA_MAX_ENTRIES);
> >>+
> >>+ mz_size = sizeof(struct sa_ctx);
> >>+ mz = rte_memzone_reserve(s, mz_size, socket_id,
> >>+ RTE_MEMZONE_1GB | RTE_MEMZONE_SIZE_HINT_ONLY);
> >>+ if (mz == NULL) {
> >>+ printf("Failed to allocate SA DB memory\n");
> >>+ rte_errno = -ENOMEM;
> >>+ return NULL;
> >>+ }
> >>+
> >>+ sa_ctx = (struct sa_ctx *)mz->addr;
> >>+
> >>+ return sa_ctx;
> >>+}
> >>+
> >>+static int
> >>+sa_add_rules(struct sa_ctx *sa_ctx, const struct ipsec_sa entries[],
> >>+ unsigned nb_entries, uint16_t cdev_id, unsigned inbound)
> >>+{
> >>+ struct ipsec_sa *sa;
> >>+ unsigned i, idx;
> >>+
> >>+ for (i = 0; i < nb_entries; i++) {
> >>+ idx = SPI2IDX(entries[i].spi);
> >>+ sa = &sa_ctx->sa[idx];
> >>+ if (sa->spi != 0) {
> >>+ printf("Index %u already in use by SPI %u\n",
> >>+ idx, sa->spi);
> >>+ return -EINVAL;
> >>+ }
> >>+ *sa = entries[i];
> >>+ sa->src = rte_cpu_to_be_32(sa->src);
> >>+ sa->dst = rte_cpu_to_be_32(sa->dst);
> >>+ if (inbound) {
> >>+ sa_ctx->xf[idx].a = sha1hmac_verify_xf;
> >>+ sa_ctx->xf[idx].b = aescbc_dec_xf;
> >>+ } else { /* outbound */
> >>+ sa_ctx->xf[idx].a = aescbc_enc_xf;
> >>+ sa_ctx->xf[idx].b = sha1hmac_gen_xf;
> >>+ }
> >>+ sa_ctx->xf[idx].a.next = &sa_ctx->xf[idx].b;
> >>+ sa_ctx->xf[idx].b.next = NULL;
> >>+ sa->xforms = &sa_ctx->xf[idx].a;
> >>+
> >>+ sa->crypto_session = rte_cryptodev_session_create(cdev_id,
> >>+ sa->xforms);
> >>+ }
> >>+
> >>+ return 0;
> >>+}
> >>+
> >>+static inline int
> >>+sa_out_add_rules(struct sa_ctx *sa_ctx, const struct ipsec_sa entries[],
> >>+ unsigned nb_entries, uint16_t cdev_id)
> >>+{
> >>+ return sa_add_rules(sa_ctx, entries, nb_entries, cdev_id, 0);
> >>+}
> >>+
> >>+static inline int
> >>+sa_in_add_rules(struct sa_ctx *sa_ctx, const struct ipsec_sa entries[],
> >>+ unsigned nb_entries, uint16_t cdev_id)
> >>+{
> >>+ return sa_add_rules(sa_ctx, entries, nb_entries, cdev_id, 1);
> >>+}
> >>+
> >>+void
> >>+sa_init(struct socket_ctx *ctx, int socket_id, unsigned ep, uint16_t cdev_id)
> >>+{
> >>+ const struct ipsec_sa *sa_out_entries, *sa_in_entries;
> >>+ unsigned nb_out_entries, nb_in_entries;
> >>+ const char *name;
> >>+
> >>+ if (ctx == NULL)
> >>+ rte_exit(EXIT_FAILURE, "NULL context.\n");
> >>+
> >>+ if (ctx->sa_ipv4_in != NULL)
> >>+ rte_exit(EXIT_FAILURE, "Inbound SA DB for socket %u already "
> >>+ "initialized\n", socket_id);
> >>+
> >>+ if (ctx->sa_ipv4_out != NULL)
> >>+ rte_exit(EXIT_FAILURE, "Outbound SA DB for socket %u already "
> >>+ "initialized\n", socket_id);
> >>+
> >>+ if (ep == 0) {
> >>+ sa_out_entries = sa_ep0_out;
> >>+ nb_out_entries = RTE_DIM(sa_ep0_out);
> >>+ sa_in_entries = sa_ep0_in;
> >>+ nb_in_entries = RTE_DIM(sa_ep0_in);
> >>+ } else if (ep == 1) {
> >>+ sa_out_entries = sa_ep1_out;
> >>+ nb_out_entries = RTE_DIM(sa_ep1_out);
> >>+ sa_in_entries = sa_ep1_in;
> >>+ nb_in_entries = RTE_DIM(sa_ep1_in);
> >>+ } else
> >>+ rte_exit(EXIT_FAILURE, "Invalid EP value %u. "
> >>+ "Only 0 or 1 supported.\n", ep);
> >>+
> >>+ name = "sa_ipv4_in";
> >>+ ctx->sa_ipv4_in = sa_ipv4_create(name, socket_id);
> >>+ if (ctx->sa_ipv4_in == NULL)
> >>+ rte_exit(EXIT_FAILURE, "Error [%d] creating SA context %s "
> >>+ "in socket %d\n", rte_errno, name, socket_id);
> >>+
> >>+ name = "sa_ipv4_out";
> >>+ ctx->sa_ipv4_out = sa_ipv4_create(name, socket_id);
> >>+ if (ctx->sa_ipv4_out == NULL)
> >>+ rte_exit(EXIT_FAILURE, "Error [%d] creating SA context %s "
> >>+ "in socket %d\n", rte_errno, name, socket_id);
> >>+
> >>+ sa_in_add_rules(ctx->sa_ipv4_in, sa_in_entries,
> >>+ nb_in_entries, cdev_id);
> >>+
> >>+ sa_out_add_rules(ctx->sa_ipv4_out, sa_out_entries,
> >>+ nb_out_entries, cdev_id);
> >>+}
> >>+
> >>+int
> >>+inbound_sa_check(struct sa_ctx *sa_ctx, struct rte_mbuf *m, uint32_t sa_idx)
> >>+{
> >>+ struct ipsec_mbuf_metadata *priv;
> >>+
> >>+ priv = RTE_PTR_ADD(m, sizeof(struct rte_mbuf));
> >>+
> >>+ return (sa_ctx->sa[sa_idx].spi == priv->sa->spi);
> >>+}
> >>+
> >>+void
> >>+inbound_sa_lookup(struct sa_ctx *sa_ctx, struct rte_mbuf *pkts[],
> >>+ struct ipsec_sa *sa[], uint16_t nb_pkts)
> >>+{
> >>+ unsigned i;
> >>+ uint32_t *src, spi;
> >>+
> >>+ for (i = 0; i < nb_pkts; i++) {
> >>+ spi = rte_pktmbuf_mtod_offset(pkts[i], struct esp_hdr *,
> >>+ sizeof(struct ip))->spi;
> >>+ if (spi == INVALID_SPI)
> >>+ continue;
> >>+
> >>+ sa[i] = &sa_ctx->sa[SPI2IDX(spi)];
> >>+ if (spi != sa[i]->spi) {
> >>+ sa[i] = NULL;
> >>+ continue;
> >>+ }
> >>+
> >>+ src = rte_pktmbuf_mtod_offset(pkts[i], uint32_t *,
> >>+ offsetof(struct ip, ip_src));
> >>+ if ((sa[i]->src != *src) || (sa[i]->dst != *(src + 1)))
> >>+ sa[i] = NULL;
> >>+ }
> >>+}
> >>+
> >>+void
> >>+outbound_sa_lookup(struct sa_ctx *sa_ctx, uint32_t sa_idx[],
> >>+ struct ipsec_sa *sa[], uint16_t nb_pkts)
> >>+{
> >>+ unsigned i;
> >>+
> >>+ for (i = 0; i < nb_pkts; i++)
> >>+ sa[i] = &sa_ctx->sa[sa_idx[i]];
> >>+}
> >>diff --git a/examples/ipsec-secgw/sp.c b/examples/ipsec-secgw/sp.c
> >>new file mode 100644
> >>index 0000000..219e478
> >>--- /dev/null
> >>+++ b/examples/ipsec-secgw/sp.c
> >>@@ -0,0 +1,324 @@
> >>+/*-
> >>+ * BSD LICENSE
> >>+ *
> >>+ * Copyright(c) 2010-2016 Intel Corporation. All rights reserved.
> >>+ * 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.
> >>+ */
> >>+
> >>+/*
> >>+ * Security Policies
> >>+ */
> >>+#include <netinet/ip.h>
> >>+
> >>+#include <rte_acl.h>
> >>+
> >>+#include "ipsec.h"
> >>+
> >>+#define MAX_ACL_RULE_NUM 1000
> >>+
> >>+/*
> >>+ * Rule and trace formats definitions.
> >>+ */
> >>+enum {
> >>+ PROTO_FIELD_IPV4,
> >>+ SRC_FIELD_IPV4,
> >>+ DST_FIELD_IPV4,
> >>+ SRCP_FIELD_IPV4,
> >>+ DSTP_FIELD_IPV4,
> >>+ NUM_FIELDS_IPV4
> >>+};
> >>+
> >>+/*
> >>+ * That effectively defines order of IPV4 classifications:
> >>+ * - PROTO
> >>+ * - SRC IP ADDRESS
> >>+ * - DST IP ADDRESS
> >>+ * - PORTS (SRC and DST)
> >>+ */
> >>+enum {
> >>+ RTE_ACL_IPV4_PROTO,
> >>+ RTE_ACL_IPV4_SRC,
> >>+ RTE_ACL_IPV4_DST,
> >>+ RTE_ACL_IPV4_PORTS,
> >>+ RTE_ACL_IPV4_NUM
> >>+};
> >>+
> >>+struct rte_acl_field_def ipv4_defs[NUM_FIELDS_IPV4] = {
> >>+ {
> >>+ .type = RTE_ACL_FIELD_TYPE_BITMASK,
> >>+ .size = sizeof(uint8_t),
> >>+ .field_index = PROTO_FIELD_IPV4,
> >>+ .input_index = RTE_ACL_IPV4_PROTO,
> >>+ .offset = 0,
> >>+ },
> >>+ {
> >>+ .type = RTE_ACL_FIELD_TYPE_MASK,
> >>+ .size = sizeof(uint32_t),
> >>+ .field_index = SRC_FIELD_IPV4,
> >>+ .input_index = RTE_ACL_IPV4_SRC,
> >>+ .offset = offsetof(struct ip, ip_src) - offsetof(struct ip, ip_p)
> >>+ },
> >>+ {
> >>+ .type = RTE_ACL_FIELD_TYPE_MASK,
> >>+ .size = sizeof(uint32_t),
> >>+ .field_index = DST_FIELD_IPV4,
> >>+ .input_index = RTE_ACL_IPV4_DST,
> >>+ .offset = offsetof(struct ip, ip_dst) - offsetof(struct ip, ip_p)
> >>+ },
> >>+ {
> >>+ .type = RTE_ACL_FIELD_TYPE_RANGE,
> >>+ .size = sizeof(uint16_t),
> >>+ .field_index = SRCP_FIELD_IPV4,
> >>+ .input_index = RTE_ACL_IPV4_PORTS,
> >>+ .offset = sizeof(struct ip) - offsetof(struct ip, ip_p)
> >>+ },
> >>+ {
> >>+ .type = RTE_ACL_FIELD_TYPE_RANGE,
> >>+ .size = sizeof(uint16_t),
> >>+ .field_index = DSTP_FIELD_IPV4,
> >>+ .input_index = RTE_ACL_IPV4_PORTS,
> >>+ .offset = sizeof(struct ip) - offsetof(struct ip, ip_p) +
> >>+ sizeof(uint16_t)
> >>+ },
> >>+};
> >>+
> >>+RTE_ACL_RULE_DEF(acl4_rules, RTE_DIM(ipv4_defs));
> >>+
> >>+const struct acl4_rules acl4_rules_in[] = {
> >>+ {
> >>+ .data = {.userdata = PROTECT(5), .category_mask = 1, .priority = 1},
> >>+ /* destination IPv4 */
> >>+ .field[2] = {.value.u32 = IPv4(192, 168, 105, 0),
> >>+ .mask_range.u32 = 24,},
> >>+ /* source port */
> >>+ .field[3] = {.value.u16 = 0, .mask_range.u16 = 0xffff,},
> >>+ /* destination port */
> >>+ .field[4] = {.value.u16 = 0, .mask_range.u16 = 0xffff,}
> >>+ },
> >>+ {
> >>+ .data = {.userdata = PROTECT(6), .category_mask = 1, .priority = 2},
> >>+ /* destination IPv4 */
> >>+ .field[2] = {.value.u32 = IPv4(192, 168, 106, 0),
> >>+ .mask_range.u32 = 24,},
> >>+ /* source port */
> >>+ .field[3] = {.value.u16 = 0, .mask_range.u16 = 0xffff,},
> >>+ /* destination port */
> >>+ .field[4] = {.value.u16 = 0, .mask_range.u16 = 0xffff,}
> >>+ },
> >>+ {
> >>+ .data = {.userdata = PROTECT(7), .category_mask = 1, .priority = 3},
> >>+ /* destination IPv4 */
> >>+ .field[2] = {.value.u32 = IPv4(192, 168, 107, 0),
> >>+ .mask_range.u32 = 24,},
> >>+ /* source port */
> >>+ .field[3] = {.value.u16 = 0, .mask_range.u16 = 0xffff,},
> >>+ /* destination port */
> >>+ .field[4] = {.value.u16 = 0, .mask_range.u16 = 0xffff,}
> >>+ },
> >>+ {
> >>+ .data = {.userdata = PROTECT(8), .category_mask = 1, .priority = 4},
> >>+ /* destination IPv4 */
> >>+ .field[2] = {.value.u32 = IPv4(192, 168, 108, 0),
> >>+ .mask_range.u32 = 24,},
> >>+ /* source port */
> >>+ .field[3] = {.value.u16 = 0, .mask_range.u16 = 0xffff,},
> >>+ /* destination port */
> >>+ .field[4] = {.value.u16 = 0, .mask_range.u16 = 0xffff,}
> >>+ }
> >>+};
> >>+
> >>+const struct acl4_rules acl4_rules_out[] = {
> >>+ {
> >>+ .data = {.userdata = PROTECT(5), .category_mask = 1, .priority = 1},
> >>+ /* destination IPv4 */
> >>+ .field[2] = {.value.u32 = IPv4(192, 168, 115, 0),
> >>+ .mask_range.u32 = 24,},
> >>+ /* source port */
> >>+ .field[3] = {.value.u16 = 0, .mask_range.u16 = 0xffff,},
> >>+ /* destination port */
> >>+ .field[4] = {.value.u16 = 0, .mask_range.u16 = 0xffff,}
> >>+ },
> >>+ {
> >>+ .data = {.userdata = PROTECT(6), .category_mask = 1, .priority = 2},
> >>+ /* destination IPv4 */
> >>+ .field[2] = {.value.u32 = IPv4(192, 168, 116, 0),
> >>+ .mask_range.u32 = 24,},
> >>+ /* source port */
> >>+ .field[3] = {.value.u16 = 0, .mask_range.u16 = 0xffff,},
> >>+ /* destination port */
> >>+ .field[4] = {.value.u16 = 0, .mask_range.u16 = 0xffff,}
> >>+ },
> >>+ {
> >>+ .data = {.userdata = PROTECT(7), .category_mask = 1, .priority = 3},
> >>+ /* destination IPv4 */
> >>+ .field[2] = {.value.u32 = IPv4(192, 168, 117, 0),
> >>+ .mask_range.u32 = 24,},
> >>+ /* source port */
> >>+ .field[3] = {.value.u16 = 0, .mask_range.u16 = 0xffff,},
> >>+ /* destination port */
> >>+ .field[4] = {.value.u16 = 0, .mask_range.u16 = 0xffff,}
> >>+ },
> >>+ {
> >>+ .data = {.userdata = PROTECT(8), .category_mask = 1, .priority = 4},
> >>+ /* destination IPv4 */
> >>+ .field[2] = {.value.u32 = IPv4(192, 168, 118, 0),
> >>+ .mask_range.u32 = 24,},
> >>+ /* source port */
> >>+ .field[3] = {.value.u16 = 0, .mask_range.u16 = 0xffff,},
> >>+ /* destination port */
> >>+ .field[4] = {.value.u16 = 0, .mask_range.u16 = 0xffff,}
> >>+ }
> >>+};
> >>+
> >>+static void
> >>+print_one_ipv4_rule(const struct acl4_rules *rule, int extra)
> >>+{
> >>+ unsigned char a, b, c, d;
> >>+
> >>+ uint32_t_to_char(rule->field[SRC_FIELD_IPV4].value.u32,
> >>+ &a, &b, &c, &d);
> >>+ printf("%hhu.%hhu.%hhu.%hhu/%u ", a, b, c, d,
> >>+ rule->field[SRC_FIELD_IPV4].mask_range.u32);
> >>+ uint32_t_to_char(rule->field[DST_FIELD_IPV4].value.u32,
> >>+ &a, &b, &c, &d);
> >>+ printf("%hhu.%hhu.%hhu.%hhu/%u ", a, b, c, d,
> >>+ rule->field[DST_FIELD_IPV4].mask_range.u32);
> >>+ printf("%hu : %hu %hu : %hu 0x%hhx/0x%hhx ",
> >>+ rule->field[SRCP_FIELD_IPV4].value.u16,
> >>+ rule->field[SRCP_FIELD_IPV4].mask_range.u16,
> >>+ rule->field[DSTP_FIELD_IPV4].value.u16,
> >>+ rule->field[DSTP_FIELD_IPV4].mask_range.u16,
> >>+ rule->field[PROTO_FIELD_IPV4].value.u8,
> >>+ rule->field[PROTO_FIELD_IPV4].mask_range.u8);
> >>+ if (extra)
> >>+ printf("0x%x-0x%x-0x%x ",
> >>+ rule->data.category_mask,
> >>+ rule->data.priority,
> >>+ rule->data.userdata);
> >>+}
> >>+
> >>+static inline void
> >>+dump_ipv4_rules(const struct acl4_rules *rule, int num, int extra)
> >>+{
> >>+ int i;
> >>+
> >>+ for (i = 0; i < num; i++, rule++) {
> >>+ printf("\t%d:", i + 1);
> >>+ print_one_ipv4_rule(rule, extra);
> >>+ printf("\n");
> >>+ }
> >>+}
> >>+
> >>+static struct rte_acl_ctx *
> >>+acl4_init(const char *name, int socketid, const struct acl4_rules *rules,
> >>+ unsigned rules_nb)
> >>+{
> >>+ char s[PATH_MAX];
> >>+ struct rte_acl_param acl_param;
> >>+ struct rte_acl_config acl_build_param;
> >>+ struct rte_acl_ctx *ctx;
> >>+
> >>+ printf("Creating SP context with %u max rules\n", MAX_ACL_RULE_NUM);
> >>+
> >>+ memset(&acl_param, 0, sizeof(acl_param));
> >>+
> >>+ /* Create ACL contexts */
> >>+ snprintf(s, sizeof(s), "%s_%d", name, socketid);
> >>+
> >>+ printf("IPv4 %s entries [%u]:\n", s, rules_nb);
> >>+ dump_ipv4_rules(rules, rules_nb, 1);
> >>+
> >>+ acl_param.name = s;
> >>+ acl_param.socket_id = socketid;
> >>+ acl_param.rule_size = RTE_ACL_RULE_SZ(RTE_DIM(ipv4_defs));
> >>+ acl_param.max_rule_num = MAX_ACL_RULE_NUM;
> >>+
> >>+ ctx = rte_acl_create(&acl_param);
> >>+ if (ctx == NULL)
> >>+ rte_exit(EXIT_FAILURE, "Failed to create ACL context\n");
> >>+
> >>+ if (rte_acl_add_rules(ctx, (const struct rte_acl_rule *)rules,
> >>+ rules_nb) < 0)
> >>+ rte_exit(EXIT_FAILURE, "add rules failed\n");
> >>+
> >>+ /* Perform builds */
> >>+ memset(&acl_build_param, 0, sizeof(acl_build_param));
> >>+
> >>+ acl_build_param.num_categories = DEFAULT_MAX_CATEGORIES;
> >>+ acl_build_param.num_fields = rules_nb;
> >>+ memcpy(&acl_build_param.defs, ipv4_defs, sizeof(ipv4_defs));
> >>+
> >>+ if (rte_acl_build(ctx, &acl_build_param) != 0)
> >>+ rte_exit(EXIT_FAILURE, "Failed to build ACL trie\n");
> >>+
> >>+ rte_acl_dump(ctx);
> >>+
> >>+ return ctx;
> >>+}
> >>+
> >>+void
> >>+sp_init(struct socket_ctx *ctx, int socket_id, unsigned ep)
> >>+{
> >>+ const char *name;
> >>+ const struct acl4_rules *rules_out, *rules_in;
> >>+ unsigned nb_out_rules, nb_in_rules;
> >>+
> >>+ if (ctx == NULL)
> >>+ rte_exit(EXIT_FAILURE, "NULL context.\n");
> >>+
> >>+ if (ctx->sp_ipv4_in != NULL)
> >>+ rte_exit(EXIT_FAILURE, "Inbound SP DB for socket %u already "
> >>+ "initialized\n", socket_id);
> >>+
> >>+ if (ctx->sp_ipv4_out != NULL)
> >>+ rte_exit(EXIT_FAILURE, "Outbound SP DB for socket %u already "
> >>+ "initialized\n", socket_id);
> >>+
> >>+ if (ep == 0) {
> >>+ rules_out = acl4_rules_in;
> >>+ nb_out_rules = RTE_DIM(acl4_rules_in);
> >>+ rules_in = acl4_rules_out;
> >>+ nb_in_rules = RTE_DIM(acl4_rules_out);
> >>+ } else if (ep == 1) {
> >>+ rules_out = acl4_rules_out;
> >>+ nb_out_rules = RTE_DIM(acl4_rules_out);
> >>+ rules_in = acl4_rules_in;
> >>+ nb_in_rules = RTE_DIM(acl4_rules_in);
> >>+ } else
> >>+ rte_exit(EXIT_FAILURE, "Invalid EP value %u. "
> >>+ "Only 0 or 1 supported.\n", ep);
> >>+
> >>+ name = "sp_ipv4_in";
> >>+ ctx->sp_ipv4_in = (struct sp_ctx *)acl4_init(name, socket_id,
> >>+ rules_in, nb_in_rules);
> >>+
> >>+ name = "sp_ipv4_out";
> >>+ ctx->sp_ipv4_out = (struct sp_ctx *)acl4_init(name, socket_id,
> >>+ rules_out, nb_out_rules);
> >>+}
> >>--
> >>2.4.3
> >>
>
next prev parent reply other threads:[~2016-02-01 11:27 UTC|newest]
Thread overview: 15+ messages / expand[flat|nested] mbox.gz Atom feed top
2016-01-29 20:29 Sergio Gonzalez Monroy
2016-01-31 14:39 ` Jerin Jacob
2016-02-01 11:09 ` Sergio Gonzalez Monroy
2016-02-01 11:26 ` Jerin Jacob [this message]
2016-02-01 14:09 ` Thomas Monjalon
2016-03-09 23:54 ` Sergio Gonzalez Monroy
2016-03-10 3:41 ` Jerin Jacob
2016-02-24 13:32 ` Thomas Monjalon
2016-02-24 14:49 ` Sergio Gonzalez Monroy
2016-03-11 1:38 ` [dpdk-dev] [PATCH v2] " Sergio Gonzalez Monroy
2016-03-11 2:12 ` De Lara Guarch, Pablo
2016-03-11 2:12 ` [dpdk-dev] [PATCH v3] " Sergio Gonzalez Monroy
2016-03-11 10:01 ` De Lara Guarch, Pablo
2016-03-11 10:07 ` Thomas Monjalon
2016-03-11 10:02 ` Thomas Monjalon
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=20160201112646.GA10490@localhost.localdomain \
--to=jerin.jacob@caviumnetworks.com \
--cc=dev@dpdk.org \
--cc=sergio.gonzalez.monroy@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).