Message ID | tencent_501B9947760682D23ADE286C65020C801307@qq.com (mailing list archive) |
---|---|
State | Changes Requested |
Delegated to: | Geliang Tang |
Headers | show |
Series | [1/3] mptcp: fix invalid addr occupy 'add_addr_accepted' | expand |
Hi Gang, On 09/01/2025 14:32, Gang Yan wrote: > From: Gang Yan <yangang@kylinos.cn> > > In ADD_ADDR timeout test cases, the syn_tx MIB counter is not an > accurate number. Please explain why this counter is not an accurate number. When you say that, it sounds like there is something to fix in the kernel, but not in the tests. (It might help to squash this patch with the next one, or to rephrase it to say that a future test will do some retransmissions) > This patch introduces '+n' to 'join_syn_tx' > to relax the strict condition. (...) Cheers, Matt
diff --git a/tools/testing/selftests/net/mptcp/mptcp_join.sh b/tools/testing/selftests/net/mptcp/mptcp_join.sh index c07e2bd3a315..6b0705d10840 100755 --- a/tools/testing/selftests/net/mptcp/mptcp_join.sh +++ b/tools/testing/selftests/net/mptcp/mptcp_join.sh @@ -1361,13 +1361,20 @@ chk_join_tx_nr() local create=${join_create_err:-0} local bind=${join_bind_err:-0} local connect=${join_connect_err:-0} + local allow_multi_syn_tx=0 local rc=${KSFT_PASS} local count + if [[ "${join_syn_tx}" = "+"* ]]; then + allow_multi_syn_tx=1 + syn_tx=${join_syn_tx:1} + fi + count=$(mptcp_lib_get_counter ${ns2} "MPTcpExtMPJoinSynTx") if [ -z "$count" ]; then rc=${KSFT_SKIP} - elif [ "$count" != "$syn_tx" ]; then + elif { [ "$count" != $syn_tx ] && [ $allow_multi_syn_tx -eq 0 ]; } || + { [ "$count" -lt $syn_tx ] && [ $allow_multi_syn_tx -eq 1 ]; }; then rc=${KSFT_FAIL} print_check "syn tx" fail_test "got $count JOIN[s] syn tx expected $syn_tx"