pktgen.txt 11.8 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
1 2


3
                  HOWTO for the linux packet generator
Linus Torvalds's avatar
Linus Torvalds committed
4 5
                  ------------------------------------

6 7
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
8 9 10
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.
Linus Torvalds's avatar
Linus Torvalds committed
11 12 13 14 15 16 17 18

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]


19
For monitoring and control pktgen creates:
Linus Torvalds's avatar
Linus Torvalds committed
20 21 22 23 24
	/proc/net/pktgen/pgctrl
	/proc/net/pktgen/kpktgend_X
        /proc/net/pktgen/ethX


25 26 27
Tuning NIC for max performance
==============================

28
The default NIC settings are (likely) not tuned for pktgen's artificial
29 30 31 32 33 34 35
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
36
than the CPU's L1/L2 cache, 2) because it allows more queueing in the
37 38
NIC HW layer (which is bad for bufferbloat).

39
One should hesitate to conclude that packets/descriptors in the HW
40
TX ring cause delay.  Drivers usually delay cleaning up the
41 42
ring-buffers for various performance reasons, and packets stalling
the TX ring might just be waiting for cleanup.
43

44 45
This cleanup issue is specifically the case for the driver ixgbe
(Intel 82599 chip).  This driver (ixgbe) combines TX+RX ring cleanups,
46 47 48
and the cleanup interval is affected by the ethtool --coalesce setting
of parameter "rx-usecs".

49
For ixgbe use e.g. "30" resulting in approx 33K interrupts/sec (1/30*10^6):
50 51 52
 # ethtool -C ethX rx-usecs 30


53 54 55 56 57 58 59 60 61 62 63 64
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.
Linus Torvalds's avatar
Linus Torvalds committed
65

66 67 68
The two basic thread commands are:
 * add_device DEVICE@NAME -- adds a single device
 * rem_device_all         -- remove all associated devices
Linus Torvalds's avatar
Linus Torvalds committed
69

70
When adding a device to a thread, a corresponding procfile is created
71 72 73 74
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
75
with multi queue NICs, the device naming scheme is extended with "@":
76 77 78 79
 device@something

The part after "@" can be anything, but it is custom to use the thread
number.
Linus Torvalds's avatar
Linus Torvalds committed
80 81 82 83

Viewing devices
===============

84 85 86
The Params section holds configured information.  The Current section
holds running statistics.  The Result is printed after a run or after
interruption.  Example:
Linus Torvalds's avatar
Linus Torvalds committed
87

88
/proc/net/pktgen/eth4@0
Linus Torvalds's avatar
Linus Torvalds committed
89

90 91
 Params: count 100000  min_pkt_size: 60  max_pkt_size: 60
     frags: 0  delay: 0  clone_skb: 64  ifname: eth4@0
Linus Torvalds's avatar
Linus Torvalds committed
92
     flows: 0 flowlen: 0
93 94 95 96 97 98 99 100 101 102 103 104 105 106
     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
Linus Torvalds's avatar
Linus Torvalds committed
107
     flows: 0
108 109 110
 Result: OK: 15430(c15405+d25) usec, 100000 (60byte,0frags)
  6480562pps 3110Mb/sec (3110669760bps) errors: 0

Linus Torvalds's avatar
Linus Torvalds committed
111

112 113
Configuring devices
===================
114 115
This is done via the /proc interface, and most easily done via pgset
as defined in the sample scripts.
Linus Torvalds's avatar
Linus Torvalds committed
116 117 118 119 120

Examples:

 pgset "clone_skb 1"     sets the number of copies of the same packet
 pgset "clone_skb 0"     use single SKB for all transmits
121 122 123
 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
Linus Torvalds's avatar
Linus Torvalds committed
124 125 126
 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
127
                         for continuous sends until explicitly stopped.
Linus Torvalds's avatar
Linus Torvalds committed
128 129 130 131 132 133 134 135 136 137 138 139 140 141 142

 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

143 144 145 146 147
 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

148
 pgset "src_mac_count 1" Sets the number of MACs we'll range through.
Linus Torvalds's avatar
Linus Torvalds committed
149 150 151 152 153 154
                         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
155 156 157 158 159
                         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,
160
                              MPLS_RND, VID_RND, SVID_RND
161
                              FLOW_SEQ,
162 163
                              QUEUE_MAP_RND # queue map random
                              QUEUE_MAP_CPU # queue map mirrors smp_processor_id()
164 165 166
                              UDPCSUM,
                              IPSEC # IPsec encapsulation (needs CONFIG_XFRM)
                              NODE_ALLOC # node specific memory allocation
167
                              NO_TIMESTAMP # disable timestamping
168

169
 pgset spi SPI_VALUE     Set specific SA used to transform packet.
Linus Torvalds's avatar
Linus Torvalds committed
170 171 172 173 174 175 176 177 178

 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.

179 180 181 182 183 184
 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,
185
					 that's done automatically. If you do
186 187 188 189 190 191 192 193 194
					 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!)

195 196 197 198 199 200 201 202 203 204 205 206 207 208 209
 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)

Linus Torvalds's avatar
Linus Torvalds committed
210 211
 pgset stop    	          aborts injection. Also, ^C aborts generator.

212 213
 pgset "rate 300M"        set rate to 300 Mb/s
 pgset "ratep 1000000"    set rate to 1Mpps
Linus Torvalds's avatar
Linus Torvalds committed
214

215 216 217 218
 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".

219 220
Sample scripts
==============
Linus Torvalds's avatar
Linus Torvalds committed
221

222 223
A collection of tutorial scripts and helpers for pktgen is in the
samples/pktgen directory. The helper parameters.sh file support easy
224
and consistent parameter parsing across the sample scripts.
225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244

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:
Linus Torvalds's avatar
Linus Torvalds committed
245 246 247 248 249 250 251 252 253 254

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
===================
255 256 257
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.
Linus Torvalds's avatar
Linus Torvalds committed
258

259 260 261
Plus using the device flag QUEUE_MAP_CPU, which maps the SKBs TX queue
to the running threads CPU (directly from smp_processor_id()).

262 263
Enable IPsec
============
264 265
Default IPsec transformation with ESP encapsulation plus transport mode
can be enabled by simply setting:
266 267 268 269 270

pgset "flag IPSEC"
pgset "flows 1"

To avoid breaking existing testbed scripts for using AH type and tunnel mode,
271
you can use "pgset spi SPI_VALUE" to specify which transformation mode
272 273
to employ.

Linus Torvalds's avatar
Linus Torvalds committed
274 275 276 277 278 279 280 281

Current commands and configuration options
==========================================

** Pgcontrol commands:

start
stop
282
reset
Linus Torvalds's avatar
Linus Torvalds committed
283 284 285 286 287 288 289 290 291 292 293

** Thread commands:

add_device
rem_device_all


** Device commands:

count
clone_skb
294
burst
Linus Torvalds's avatar
Linus Torvalds committed
295 296 297 298 299 300 301 302
debug

frags
delay

src_mac_count
dst_mac_count

303
pkt_size
Linus Torvalds's avatar
Linus Torvalds committed
304 305 306
min_pkt_size
max_pkt_size

307 308 309 310 311 312 313
queue_map_min
queue_map_max
skb_priority

tos           (ipv4)
traffic_class (ipv6)

314 315
mpls

Linus Torvalds's avatar
Linus Torvalds committed
316 317 318 319 320 321
udp_src_min
udp_src_max

udp_dst_min
udp_dst_max

322 323
node

Linus Torvalds's avatar
Linus Torvalds committed
324 325 326 327 328 329 330
flag
  IPSRC_RND
  IPDST_RND
  UDPSRC_RND
  UDPDST_RND
  MACSRC_RND
  MACDST_RND
331 332 333 334 335 336 337 338 339
  TXSIZE_RND
  IPV6
  MPLS_RND
  VID_RND
  SVID_RND
  FLOW_SEQ
  QUEUE_MAP_RND
  QUEUE_MAP_CPU
  UDPCSUM
340
  IPSEC
341
  NODE_ALLOC
342
  NO_TIMESTAMP
Linus Torvalds's avatar
Linus Torvalds committed
343

344 345
spi (ipsec)

Linus Torvalds's avatar
Linus Torvalds committed
346 347 348 349 350 351 352 353 354 355 356 357
dst_min
dst_max

src_min
src_max

dst_mac
src_mac

clear_counters

src6
358 359 360
dst6
dst6_max
dst6_min
Linus Torvalds's avatar
Linus Torvalds committed
361 362 363 364

flows
flowlen

365 366 367
rate
ratep

368 369
xmit_mode <start_xmit|netif_receive>

370 371 372 373 374 375 376 377
vlan_cfi
vlan_id
vlan_p

svlan_cfi
svlan_id
svlan_p

378

Linus Torvalds's avatar
Linus Torvalds committed
379 380 381 382 383 384 385 386 387 388 389 390
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.


391
Good luck with the linux net-development.