On Wed, Oct 7, 2026, at 04:47, Haishuang Yan wrote: > On slow machines, do_test() raises initcwnd on the route into the tunnel > with "ip route change". For IPv6 traffic this fails with: > > RTNETLINK answers: No such file or directory
Thanks for the fix! It's already merged, but anyway: Tested-by: Alice Mikityanska <[email protected]> > The IPv6 prefix route of the tunnel address has metric 256, while > "ip -6 route change" without a metric looks for a route with the default > metric 1024, so nothing is changed. The test still passes, but without > the larger initcwnd it is meant to use. > > Pass the metric of the prefix route. > > Fixes: 1d62b83fb751 ("selftests: net: Fix slow configurations in > big_tcp_tunnels.sh") > Signed-off-by: Haishuang Yan <[email protected]> > --- > tools/testing/selftests/net/big_tcp_tunnels.sh | 3 ++- > 1 file changed, 2 insertions(+), 1 deletion(-) > > diff --git a/tools/testing/selftests/net/big_tcp_tunnels.sh > b/tools/testing/selftests/net/big_tcp_tunnels.sh > index cc0875e52fb9..fb43fed824e3 100755 > --- a/tools/testing/selftests/net/big_tcp_tunnels.sh > +++ b/tools/testing/selftests/net/big_tcp_tunnels.sh > @@ -145,7 +145,8 @@ do_test() { > route change 192.168.2.0/24 dev tun0 initcwnd 100 > else > ip -netns "$CLIENT_NS" -6 \ > - route change 2001:db8::2:0/112 dev tun0 initcwnd 100 > + route change 2001:db8::2:0/112 dev tun0 initcwnd > 100 \ > + metric 256 > fi > else > ip netns exec "$CLIENT_NS" \ > -- > 2.43.0
