[Feature]add MT2731_MP2_MR2_SVN388 baseline version

Change-Id: Ief04314834b31e27effab435d3ca8ba33b499059
diff --git a/src/kernel/linux/v4.14/Documentation/networking/pktgen.txt b/src/kernel/linux/v4.14/Documentation/networking/pktgen.txt
new file mode 100644
index 0000000..2c4e335
--- /dev/null
+++ b/src/kernel/linux/v4.14/Documentation/networking/pktgen.txt
@@ -0,0 +1,391 @@
+
+
+                  HOWTO for the linux packet generator
+                  ------------------------------------
+
+Enable CONFIG_NET_PKTGEN to compile and build pktgen either in-kernel
+or as a module.  A module is preferred; modprobe pktgen if needed.  Once
+running, pktgen creates a thread for each CPU with affinity to that CPU.
+Monitoring and controlling is done via /proc.  It is easiest to select a
+suitable sample script and configure that.
+
+On a dual CPU:
+
+ps aux | grep pkt
+root       129  0.3  0.0     0    0 ?        SW    2003 523:20 [pktgen/0]
+root       130  0.3  0.0     0    0 ?        SW    2003 509:50 [pktgen/1]
+
+
+For monitoring and control pktgen creates:
+	/proc/net/pktgen/pgctrl
+	/proc/net/pktgen/kpktgend_X
+        /proc/net/pktgen/ethX
+
+
+Tuning NIC for max performance
+==============================
+
+The default NIC settings are (likely) not tuned for pktgen's artificial
+overload type of benchmarking, as this could hurt the normal use-case.
+
+Specifically increasing the TX ring buffer in the NIC:
+ # ethtool -G ethX tx 1024
+
+A larger TX ring can improve pktgen's performance, while it can hurt
+in the general case, 1) because the TX ring buffer might get larger
+than the CPU's L1/L2 cache, 2) because it allows more queueing in the
+NIC HW layer (which is bad for bufferbloat).
+
+One should hesitate to conclude that packets/descriptors in the HW
+TX ring cause delay.  Drivers usually delay cleaning up the
+ring-buffers for various performance reasons, and packets stalling
+the TX ring might just be waiting for cleanup.
+
+This cleanup issue is specifically the case for the driver ixgbe
+(Intel 82599 chip).  This driver (ixgbe) combines TX+RX ring cleanups,
+and the cleanup interval is affected by the ethtool --coalesce setting
+of parameter "rx-usecs".
+
+For ixgbe use e.g. "30" resulting in approx 33K interrupts/sec (1/30*10^6):
+ # ethtool -C ethX rx-usecs 30
+
+
+Kernel threads
+==============
+Pktgen creates a thread for each CPU with affinity to that CPU.
+Which is controlled through procfile /proc/net/pktgen/kpktgend_X.
+
+Example: /proc/net/pktgen/kpktgend_0
+
+ Running:
+ Stopped: eth4@0
+ Result: OK: add_device=eth4@0
+
+Most important are the devices assigned to the thread.
+
+The two basic thread commands are:
+ * add_device DEVICE@NAME -- adds a single device
+ * rem_device_all         -- remove all associated devices
+
+When adding a device to a thread, a corresponding procfile is created
+which is used for configuring this device. Thus, device names need to
+be unique.
+
+To support adding the same device to multiple threads, which is useful
+with multi queue NICs, the device naming scheme is extended with "@":
+ device@something
+
+The part after "@" can be anything, but it is custom to use the thread
+number.
+
+Viewing devices
+===============
+
+The Params section holds configured information.  The Current section
+holds running statistics.  The Result is printed after a run or after
+interruption.  Example:
+
+/proc/net/pktgen/eth4@0
+
+ Params: count 100000  min_pkt_size: 60  max_pkt_size: 60
+     frags: 0  delay: 0  clone_skb: 64  ifname: eth4@0
+     flows: 0 flowlen: 0
+     queue_map_min: 0  queue_map_max: 0
+     dst_min: 192.168.81.2  dst_max:
+     src_min:   src_max:
+     src_mac: 90:e2:ba:0a:56:b4 dst_mac: 00:1b:21:3c:9d:f8
+     udp_src_min: 9  udp_src_max: 109  udp_dst_min: 9  udp_dst_max: 9
+     src_mac_count: 0  dst_mac_count: 0
+     Flags: UDPSRC_RND  NO_TIMESTAMP  QUEUE_MAP_CPU
+ Current:
+     pkts-sofar: 100000  errors: 0
+     started: 623913381008us  stopped: 623913396439us idle: 25us
+     seq_num: 100001  cur_dst_mac_offset: 0  cur_src_mac_offset: 0
+     cur_saddr: 192.168.8.3  cur_daddr: 192.168.81.2
+     cur_udp_dst: 9  cur_udp_src: 42
+     cur_queue_map: 0
+     flows: 0
+ Result: OK: 15430(c15405+d25) usec, 100000 (60byte,0frags)
+  6480562pps 3110Mb/sec (3110669760bps) errors: 0
+
+
+Configuring devices
+===================
+This is done via the /proc interface, and most easily done via pgset
+as defined in the sample scripts.
+
+Examples:
+
+ pgset "clone_skb 1"     sets the number of copies of the same packet
+ pgset "clone_skb 0"     use single SKB for all transmits
+ pgset "burst 8"         uses xmit_more API to queue 8 copies of the same
+                         packet and update HW tx queue tail pointer once.
+                         "burst 1" is the default
+ pgset "pkt_size 9014"   sets packet size to 9014
+ pgset "frags 5"         packet will consist of 5 fragments
+ pgset "count 200000"    sets number of packets to send, set to zero
+                         for continuous sends until explicitly stopped.
+
+ pgset "delay 5000"      adds delay to hard_start_xmit(). nanoseconds
+
+ pgset "dst 10.0.0.1"    sets IP destination address
+                         (BEWARE! This generator is very aggressive!)
+
+ pgset "dst_min 10.0.0.1"            Same as dst
+ pgset "dst_max 10.0.0.254"          Set the maximum destination IP.
+ pgset "src_min 10.0.0.1"            Set the minimum (or only) source IP.
+ pgset "src_max 10.0.0.254"          Set the maximum source IP.
+ pgset "dst6 fec0::1"     IPV6 destination address
+ pgset "src6 fec0::2"     IPV6 source address
+ pgset "dstmac 00:00:00:00:00:00"    sets MAC destination address
+ pgset "srcmac 00:00:00:00:00:00"    sets MAC source address
+
+ pgset "queue_map_min 0" Sets the min value of tx queue interval
+ pgset "queue_map_max 7" Sets the max value of tx queue interval, for multiqueue devices
+                         To select queue 1 of a given device,
+                         use queue_map_min=1 and queue_map_max=1
+
+ pgset "src_mac_count 1" Sets the number of MACs we'll range through.
+                         The 'minimum' MAC is what you set with srcmac.
+
+ pgset "dst_mac_count 1" Sets the number of MACs we'll range through.
+                         The 'minimum' MAC is what you set with dstmac.
+
+ pgset "flag [name]"     Set a flag to determine behaviour.  Current flags
+                         are: IPSRC_RND # IP source is random (between min/max)
+                              IPDST_RND # IP destination is random
+                              UDPSRC_RND, UDPDST_RND,
+                              MACSRC_RND, MACDST_RND
+                              TXSIZE_RND, IPV6,
+                              MPLS_RND, VID_RND, SVID_RND
+                              FLOW_SEQ,
+                              QUEUE_MAP_RND # queue map random
+                              QUEUE_MAP_CPU # queue map mirrors smp_processor_id()
+                              UDPCSUM,
+                              IPSEC # IPsec encapsulation (needs CONFIG_XFRM)
+                              NODE_ALLOC # node specific memory allocation
+                              NO_TIMESTAMP # disable timestamping
+
+ pgset spi SPI_VALUE     Set specific SA used to transform packet.
+
+ pgset "udp_src_min 9"   set UDP source port min, If < udp_src_max, then
+                         cycle through the port range.
+
+ pgset "udp_src_max 9"   set UDP source port max.
+ pgset "udp_dst_min 9"   set UDP destination port min, If < udp_dst_max, then
+                         cycle through the port range.
+ pgset "udp_dst_max 9"   set UDP destination port max.
+
+ pgset "mpls 0001000a,0002000a,0000000a" set MPLS labels (in this example
+                                         outer label=16,middle label=32,
+					 inner label=0 (IPv4 NULL)) Note that
+					 there must be no spaces between the
+					 arguments. Leading zeros are required.
+					 Do not set the bottom of stack bit,
+					 that's done automatically. If you do
+					 set the bottom of stack bit, that
+					 indicates that you want to randomly
+					 generate that address and the flag
+					 MPLS_RND will be turned on. You
+					 can have any mix of random and fixed
+					 labels in the label stack.
+
+ pgset "mpls 0"		  turn off mpls (or any invalid argument works too!)
+
+ pgset "vlan_id 77"       set VLAN ID 0-4095
+ pgset "vlan_p 3"         set priority bit 0-7 (default 0)
+ pgset "vlan_cfi 0"       set canonical format identifier 0-1 (default 0)
+
+ pgset "svlan_id 22"      set SVLAN ID 0-4095
+ pgset "svlan_p 3"        set priority bit 0-7 (default 0)
+ pgset "svlan_cfi 0"      set canonical format identifier 0-1 (default 0)
+
+ pgset "vlan_id 9999"     > 4095 remove vlan and svlan tags
+ pgset "svlan 9999"       > 4095 remove svlan tag
+
+
+ pgset "tos XX"           set former IPv4 TOS field (e.g. "tos 28" for AF11 no ECN, default 00)
+ pgset "traffic_class XX" set former IPv6 TRAFFIC CLASS (e.g. "traffic_class B8" for EF no ECN, default 00)
+
+ pgset stop    	          aborts injection. Also, ^C aborts generator.
+
+ pgset "rate 300M"        set rate to 300 Mb/s
+ pgset "ratep 1000000"    set rate to 1Mpps
+
+ pgset "xmit_mode netif_receive"  RX inject into stack netif_receive_skb()
+				  Works with "burst" but not with "clone_skb".
+				  Default xmit_mode is "start_xmit".
+
+Sample scripts
+==============
+
+A collection of tutorial scripts and helpers for pktgen is in the
+samples/pktgen directory. The helper parameters.sh file support easy
+and consistent parameter parsing across the sample scripts.
+
+Usage example and help:
+ ./pktgen_sample01_simple.sh -i eth4 -m 00:1B:21:3C:9D:F8 -d 192.168.8.2
+
+Usage: ./pktgen_sample01_simple.sh [-vx] -i ethX
+  -i : ($DEV)       output interface/device (required)
+  -s : ($PKT_SIZE)  packet size
+  -d : ($DEST_IP)   destination IP
+  -m : ($DST_MAC)   destination MAC-addr
+  -t : ($THREADS)   threads to start
+  -c : ($SKB_CLONE) SKB clones send before alloc new SKB
+  -b : ($BURST)     HW level bursting of SKBs
+  -v : ($VERBOSE)   verbose
+  -x : ($DEBUG)     debug
+
+The global variables being set are also listed.  E.g. the required
+interface/device parameter "-i" sets variable $DEV.  Copy the
+pktgen_sampleXX scripts and modify them to fit your own needs.
+
+The old scripts:
+
+pktgen.conf-1-2                  # 1 CPU 2 dev
+pktgen.conf-1-1-rdos             # 1 CPU 1 dev w. route DoS 
+pktgen.conf-1-1-ip6              # 1 CPU 1 dev ipv6
+pktgen.conf-1-1-ip6-rdos         # 1 CPU 1 dev ipv6  w. route DoS
+pktgen.conf-1-1-flows            # 1 CPU 1 dev multiple flows.
+
+
+Interrupt affinity
+===================
+Note that when adding devices to a specific CPU it is a good idea to
+also assign /proc/irq/XX/smp_affinity so that the TX interrupts are bound
+to the same CPU.  This reduces cache bouncing when freeing skbs.
+
+Plus using the device flag QUEUE_MAP_CPU, which maps the SKBs TX queue
+to the running threads CPU (directly from smp_processor_id()).
+
+Enable IPsec
+============
+Default IPsec transformation with ESP encapsulation plus transport mode
+can be enabled by simply setting:
+
+pgset "flag IPSEC"
+pgset "flows 1"
+
+To avoid breaking existing testbed scripts for using AH type and tunnel mode,
+you can use "pgset spi SPI_VALUE" to specify which transformation mode
+to employ.
+
+
+Current commands and configuration options
+==========================================
+
+** Pgcontrol commands:
+
+start
+stop
+reset
+
+** Thread commands:
+
+add_device
+rem_device_all
+
+
+** Device commands:
+
+count
+clone_skb
+burst
+debug
+
+frags
+delay
+
+src_mac_count
+dst_mac_count
+
+pkt_size
+min_pkt_size
+max_pkt_size
+
+queue_map_min
+queue_map_max
+skb_priority
+
+tos           (ipv4)
+traffic_class (ipv6)
+
+mpls
+
+udp_src_min
+udp_src_max
+
+udp_dst_min
+udp_dst_max
+
+node
+
+flag
+  IPSRC_RND
+  IPDST_RND
+  UDPSRC_RND
+  UDPDST_RND
+  MACSRC_RND
+  MACDST_RND
+  TXSIZE_RND
+  IPV6
+  MPLS_RND
+  VID_RND
+  SVID_RND
+  FLOW_SEQ
+  QUEUE_MAP_RND
+  QUEUE_MAP_CPU
+  UDPCSUM
+  IPSEC
+  NODE_ALLOC
+  NO_TIMESTAMP
+
+spi (ipsec)
+
+dst_min
+dst_max
+
+src_min
+src_max
+
+dst_mac
+src_mac
+
+clear_counters
+
+src6
+dst6
+dst6_max
+dst6_min
+
+flows
+flowlen
+
+rate
+ratep
+
+xmit_mode <start_xmit|netif_receive>
+
+vlan_cfi
+vlan_id
+vlan_p
+
+svlan_cfi
+svlan_id
+svlan_p
+
+
+References:
+ftp://robur.slu.se/pub/Linux/net-development/pktgen-testing/
+ftp://robur.slu.se/pub/Linux/net-development/pktgen-testing/examples/
+
+Paper from Linux-Kongress in Erlangen 2004.
+ftp://robur.slu.se/pub/Linux/net-development/pktgen-testing/pktgen_paper.pdf
+
+Thanks to:
+Grant Grundler for testing on IA-64 and parisc, Harald Welte,  Lennert Buytenhek
+Stephen Hemminger, Andi Kleen, Dave Miller and many others.
+
+
+Good luck with the linux net-development.