Re: [PATCH net-next v3 12/12] selftests: net: Add a test for BIG TCP in UDP tunnels

[Willem de Bruijn <willemdebruijn.kernel@gmail.com>]

Alice Mikityanska wrote:
> On Thu, 16 Apr 2026 at 14:06, Willem de Bruijn
> <willemdebruijn.kernel@gmail.com> wrote:
> >
> > Alice Mikityanska wrote:
> > > From: Alice Mikityanska <alice@isovalent.com>
> > >
> > > The test sets up VXLAN and GENEVE tunnels over IPv4 and IPv6 and runs
> > > IPv4 and IPv6 traffic through them with BIG TCP enabled. It checks that
> > > a non-negligible amount of big aggregated packets are seen in tcpdump.
> > >
> > > Signed-off-by: Alice Mikityanska <alice@isovalent.com>
> > > ---
> > >  tools/testing/selftests/net/Makefile          |   1 +
> > >  .../testing/selftests/net/big_tcp_tunnels.sh  | 145 ++++++++++++++++++
> > >  2 files changed, 146 insertions(+)
> > >  create mode 100755 tools/testing/selftests/net/big_tcp_tunnels.sh
> > >
> > > diff --git a/tools/testing/selftests/net/Makefile b/tools/testing/selftests/net/Makefile
> > > index cab74ebdaced..c8ea9d4bb94f 100644
> > > --- a/tools/testing/selftests/net/Makefile
> > > +++ b/tools/testing/selftests/net/Makefile
> > > @@ -13,6 +13,7 @@ TEST_PROGS := \
> > >       arp_ndisc_untracked_subnets.sh \
> > >       bareudp.sh \
> > >       big_tcp.sh \
> > > +     big_tcp_tunnels.sh \
> > >       bind_bhash.sh \
> > >       bpf_offload.py \
> > >       bridge_vlan_dump.sh \
> > > diff --git a/tools/testing/selftests/net/big_tcp_tunnels.sh b/tools/testing/selftests/net/big_tcp_tunnels.sh
> > > new file mode 100755
> > > index 000000000000..b819911519ac
> > > --- /dev/null
> > > +++ b/tools/testing/selftests/net/big_tcp_tunnels.sh
> > > @@ -0,0 +1,145 @@
> > > +#!/usr/bin/env bash
> > > +# SPDX-License-Identifier: GPL-2.0
> > > +#
> > > +# Testing for IPv4 and IPv6 BIG TCP over VXLAN and GENEVE tunnels.
> > > +
> > > +SERVER_NS=$(mktemp -u server-XXXXXXXX)
> > > +SERVER_IP4="192.168.1.1"
> > > +SERVER_IP6="2001:db8::1:1"
> > > +SERVER_IP4_TUN="192.168.2.1"
> > > +SERVER_IP6_TUN="2001:db8::2:1"
> > > +
> > > +CLIENT_NS=$(mktemp -u client-XXXXXXXX)
> > > +CLIENT_IP4="192.168.1.2"
> > > +CLIENT_IP6="2001:db8::1:2"
> > > +CLIENT_IP4_TUN="192.168.2.2"
> > > +CLIENT_IP6_TUN="2001:db8::2:2"
> > > +
> > > +PACKETS_THRESHOLD=10000
> > > +
> > > +# Kselftest framework requirement - SKIP code is 4.
> > > +ksft_skip=4
> > > +
> > > +setup() {
> > > +     ip netns add "$SERVER_NS"
> > > +     ip netns add "$CLIENT_NS"
> > > +     ip -netns "$SERVER_NS" link add link1 type veth peer name link0 netns "$CLIENT_NS"
> > > +
> > > +     ip -netns "$CLIENT_NS" link set link0 up
> > > +     ip -netns "$CLIENT_NS" addr replace "$CLIENT_IP4/24" dev link0
> > > +     ip -netns "$CLIENT_NS" addr replace "$CLIENT_IP6/112" dev link0 nodad
> > > +     ip -netns "$CLIENT_NS" link set link0 \
> > > +             gso_max_size 196608 gso_ipv4_max_size 196608 \
> > > +             gro_max_size 196608 gro_ipv4_max_size 196608
> > > +     ip -netns "$SERVER_NS" link set link1 up
> > > +     ip -netns "$SERVER_NS" addr replace "$SERVER_IP4/24" dev link1
> > > +     ip -netns "$SERVER_NS" addr replace "$SERVER_IP6/112" dev link1 nodad
> > > +     ip -netns "$SERVER_NS" link set link1 \
> > > +             gso_max_size 196608 gso_ipv4_max_size 196608 \
> > > +             gro_max_size 196608 gro_ipv4_max_size 196608
> > > +
> > > +     ip netns exec "$SERVER_NS" netserver >/dev/null
> > > +}
> > > +
> > > +setup_tunnel() {
> > > +     if [ "$2" = 4 ]; then
> > > +             SERVER_IP="$SERVER_IP4"
> > > +             CLIENT_IP="$CLIENT_IP4"
> > > +             echo "Setting up ${1^^} over IPv4"
> > > +     else
> > > +             SERVER_IP="$SERVER_IP6"
> > > +             CLIENT_IP="$CLIENT_IP6"
> > > +             echo "Setting up ${1^^} over IPv6"
> > > +     fi
> > > +
> > > +     if [ "$1" = vxlan ]; then
> > > +             ip -netns "$CLIENT_NS" link add tun0 type vxlan \
> > > +                     id 5001 remote "$SERVER_IP" local "$CLIENT_IP" dev link0 dstport 4789
> > > +     else
> > > +             ip -netns "$CLIENT_NS" link add tun0 type geneve \
> > > +                     id 5001 remote "$SERVER_IP"
> > > +     fi
> > > +     ip -netns "$CLIENT_NS" link set tun0 up
> > > +     ip -netns "$CLIENT_NS" addr replace "$CLIENT_IP4_TUN/24" dev tun0
> > > +     ip -netns "$CLIENT_NS" addr replace "$CLIENT_IP6_TUN/112" dev tun0 nodad
> > > +     ip -netns "$CLIENT_NS" link set tun0 \
> > > +             gso_max_size 196608 gso_ipv4_max_size 196608 \
> > > +             gro_max_size 196608 gro_ipv4_max_size 196608
> > > +     if [ "$1" = vxlan ]; then
> > > +             ip -netns "$SERVER_NS" link add tun1 type vxlan \
> > > +                     id 5001 remote "$CLIENT_IP" local "$SERVER_IP" dev link1 dstport 4789
> > > +     else
> > > +             ip -netns "$SERVER_NS" link add tun1 type geneve \
> > > +                     id 5001 remote "$CLIENT_IP"
> > > +     fi
> > > +     ip -netns "$SERVER_NS" link set tun1 up
> > > +     ip -netns "$SERVER_NS" addr replace "$SERVER_IP4_TUN/24" dev tun1
> > > +     ip -netns "$SERVER_NS" addr replace "$SERVER_IP6_TUN/112" dev tun1 nodad
> > > +     ip -netns "$SERVER_NS" link set tun1 \
> > > +             gso_max_size 196608 gso_ipv4_max_size 196608 \
> > > +             gro_max_size 196608 gro_ipv4_max_size 196608
> > > +}
> > > +
> > > +cleanup_tunnel() {
> > > +     ip -netns "$CLIENT_NS" link del tun0
> > > +     ip -netns "$SERVER_NS" link del tun1
> > > +}
> > > +
> > > +cleanup() {
> > > +     ip netns exec "$SERVER_NS" killall netserver
> > > +     ip netns del "$SERVER_NS"
> > > +     ip netns del "$CLIENT_NS"
> > > +}
> > > +
> > > +do_test() {
> > > +     exec 3< <(ip netns exec "$SERVER_NS" tcpdump -nn -i link1 greater 65536 2> /dev/null)
> > > +     TCPDUMP_SERVER_PID="$!"
> > > +     exec 4< <(wc -l <&3)
> > > +     exec 5< <(ip netns exec "$CLIENT_NS" tcpdump -nn -i link0 greater 65536 2> /dev/null)
> > > +     TCPDUMP_CLIENT_PID="$!"
> > > +     exec 6< <(wc -l <&5)
> > > +
> > > +     if [ "$1" = 4 ]; then
> > > +             SERVER_IP="$SERVER_IP4_TUN"
> > > +             echo "Running IPv4 traffic in the tunnel"
> > > +     else
> > > +             SERVER_IP="$SERVER_IP6_TUN"
> > > +             echo "Running IPv6 traffic in the tunnel"
> > > +     fi
> > > +
> > > +     ip netns exec "$CLIENT_NS" netperf -t TCP_STREAM -l 5 -H "$SERVER_IP" -- \
> > > +             -r 80000:80000 > /dev/null
> >
> > is -r valid for TCP_STREAM
> 
> Yes, it controls how much data is sent in one send() syscall. If I set
> a smaller size, the kernel will still try to aggregate the packets,
> but there will be much fewer BIG TCP packets.

Not according to the care & feeding doc, where it is an RR option.
Maybe it is used in practice. Maybe also depends on whether using
omni or bsd mode. But just a note.