Squash to "selftests: mptcp: join: check re-re-adding ID 0 endp"

This will make "delete and re-add" MPTCP Join subtest longer.

This test is manipulating the endpoints while the connection runs in the
background. When all the manipulations are done, it stops (kills) the
connection. It is then fine to make the connection run for a longer time
by sending smaller chunks per batch, as it is anyway stopped when the
operations are over.

This modification is needed, because this test is doing more operations
now, and this causes the Netdev CI to fail, as reported by Jakub [1]. In
the logs [2], we can see that there is no more subflows for the
connections, plus we don't see the "read: Connection reset by peer"
message, telling us the connection has been killed. (Note that this
message is removed in "selftests: mptcp: join: mute errors when ran in
the background" patch, queued for net-next.)

We didn't see this error on our side, probably because we disable
CONFIG_SLUB_DEBUG_ON when using the debug kernel config, plus the Netdev
CI can be quite busy from time to time.

Link: https://lore.kernel.org/netdev/[email protected]/ [1]
Link: https://netdev-3.bots.linux.dev/vmksft-mptcp-dbg/results/745102/1-mptcp-join-sh/stdout [2]
Signed-off-by: Matthieu Baerts (NGI0) <[email protected]>

tools/testing/selftests/net/mptcp/mptcp_join.sh

@@ -3685,7 +3685,7 @@ endpoint_tests()
 		pm_nl_set_limits $ns2 0 3
 		pm_nl_add_endpoint $ns2 10.0.1.2 id 1 dev ns2eth1 flags subflow
 		pm_nl_add_endpoint $ns2 10.0.2.2 id 2 dev ns2eth2 flags subflow
-		{ test_linkfail=4 speed=20 \
+		{ test_linkfail=4 speed=5 \
 			run_tests $ns1 $ns2 10.0.1.1 & } 2>/dev/null
 		local tests_pid=$!