selftests: rtnetlink: Use more sensible tos values

Using tos 0x1 with 'ip route get <IPv4 address> ...' doesn't test much
of the tos option handling: 0x1 just sets an ECN bit, which is cleared
by inet_rtm_getroute() before doing the fib lookup. Let's use 0x10
instead, which is actually taken into account in the route lookup (and
is less surprising for the reader).

For consistency, use 0x10 for the IPv6 route lookup too (IPv6 currently
doesn't clear ECN bits, but might do so in the future).

Signed-off-by: Guillaume Nault <gnault@redhat.com>
Link: https://lore.kernel.org/r/d61119e68d01ba7ef3ba50c1345a5123a11de123.1643815297.git.gnault@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>

diff --git a/tools/testing/selftests/net/rtnetlink.sh b/tools/testing/selftests/net/rtnetlink.sh
index c9ce3df..0900c54 100755 (executable)
--- a/tools/testing/selftests/net/rtnetlink.sh
+++ b/tools/testing/selftests/net/rtnetlink.sh
@@ -216,9 +216,9 @@ kci_test_route_get()
        check_err $?
        ip route get fe80::1 dev "$devdummy" > /dev/null
        check_err $?
-       ip route get 127.0.0.1 from 127.0.0.1 oif lo tos 0x1 mark 0x1 > /dev/null
+       ip route get 127.0.0.1 from 127.0.0.1 oif lo tos 0x10 mark 0x1 > /dev/null
        check_err $?
-       ip route get ::1 from ::1 iif lo oif lo tos 0x1 mark 0x1 > /dev/null
+       ip route get ::1 from ::1 iif lo oif lo tos 0x10 mark 0x1 > /dev/null
        check_err $?
        ip addr add dev "$devdummy" 10.23.7.11/24
        check_err $?