1d73ba16ad
Add the pktgen samples script pktgen_sample03_burst_single_flow.sh
that demonstrates how to acheive maximum performance.
If correctly tuned[1] single CPU 10Gbit/s wirespeed small pkts is
possible[2] which is 14.88Mpps. The trick is to take advantage of the
"burst" feature introduced in commit 38b2cf2982 ("net: pktgen:
packet bursting via skb->xmit_more").
[1] http://netoptimizer.blogspot.dk/2014/06/pktgen-for-network-overload-testing.html
[2] http://netoptimizer.blogspot.dk/2014/10/unlocked-10gbps-tx-wirespeed-smallest.html
Signed-off-by: Jesper Dangaard Brouer <brouer@redhat.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
82 lines
2.5 KiB
Bash
Executable file
82 lines
2.5 KiB
Bash
Executable file
#!/bin/bash
|
|
#
|
|
# Script for max single flow performance
|
|
# - If correctly tuned[1], single CPU 10G wirespeed small pkts is possible[2]
|
|
#
|
|
# Using pktgen "burst" option (use -b $N)
|
|
# - To boost max performance
|
|
# - Avail since: kernel v3.18
|
|
# * commit 38b2cf2982dc73 ("net: pktgen: packet bursting via skb->xmit_more")
|
|
# - This avoids writing the HW tailptr on every driver xmit
|
|
# - The performance boost is impressive, see commit and blog [2]
|
|
#
|
|
# Notice: On purpose generates a single (UDP) flow towards target,
|
|
# reason behind this is to only overload/activate a single CPU on
|
|
# target host. And no randomness for pktgen also makes it faster.
|
|
#
|
|
# Tuning see:
|
|
# [1] http://netoptimizer.blogspot.dk/2014/06/pktgen-for-network-overload-testing.html
|
|
# [2] http://netoptimizer.blogspot.dk/2014/10/unlocked-10gbps-tx-wirespeed-smallest.html
|
|
#
|
|
basedir=`dirname $0`
|
|
source ${basedir}/functions.sh
|
|
root_check_run_with_sudo "$@"
|
|
|
|
# Parameter parsing via include
|
|
source ${basedir}/parameters.sh
|
|
# Set some default params, if they didn't get set
|
|
[ -z "$DEST_IP" ] && DEST_IP="198.18.0.42"
|
|
[ -z "$DST_MAC" ] && DST_MAC="90:e2:ba:ff:ff:ff"
|
|
[ -z "$BURST" ] && BURST=32
|
|
[ -z "$CLONE_SKB" ] && CLONE_SKB="100000"
|
|
|
|
# Base Config
|
|
DELAY="0" # Zero means max speed
|
|
COUNT="0" # Zero means indefinitely
|
|
|
|
# General cleanup everything since last run
|
|
pg_ctrl "reset"
|
|
|
|
# Threads are specified with parameter -t value in $THREADS
|
|
for ((thread = 0; thread < $THREADS; thread++)); do
|
|
dev=${DEV}@${thread}
|
|
|
|
# Add remove all other devices and add_device $dev to thread
|
|
pg_thread $thread "rem_device_all"
|
|
pg_thread $thread "add_device" $dev
|
|
|
|
# Base config
|
|
pg_set $dev "flag QUEUE_MAP_CPU"
|
|
pg_set $dev "count $COUNT"
|
|
pg_set $dev "clone_skb $CLONE_SKB"
|
|
pg_set $dev "pkt_size $PKT_SIZE"
|
|
pg_set $dev "delay $DELAY"
|
|
pg_set $dev "flag NO_TIMESTAMP"
|
|
|
|
# Destination
|
|
pg_set $dev "dst_mac $DST_MAC"
|
|
pg_set $dev "dst $DEST_IP"
|
|
|
|
# Setup burst, for easy testing -b 0 disable bursting
|
|
# (internally in pktgen default and minimum burst=1)
|
|
if [[ ${BURST} -ne 0 ]]; then
|
|
pg_set $dev "burst $BURST"
|
|
else
|
|
info "$dev: Not using burst"
|
|
fi
|
|
done
|
|
|
|
# Run if user hits control-c
|
|
function control_c() {
|
|
# Print results
|
|
for ((thread = 0; thread < $THREADS; thread++)); do
|
|
dev=${DEV}@${thread}
|
|
echo "Device: $dev"
|
|
cat /proc/net/pktgen/$dev | grep -A2 "Result:"
|
|
done
|
|
}
|
|
# trap keyboard interrupt (Ctrl-C)
|
|
trap control_c SIGINT
|
|
|
|
echo "Running... ctrl^C to stop" >&2
|
|
pg_ctrl "start"
|