From: Shardul Bankar When a packet arrives with map_seq < ack_seq < end_seq, the beginning of the packet has already been acknowledged but the end contains new data. Currently the entire packet is dropped as "old data," forcing the sender to retransmit. Instead, skip the already-acked bytes by adjusting the skb offset and enqueue only the new portion. Update bytes_received and ack_seq to reflect the new data consumed. A previous attempt at this fix has been sent by Paolo Abeni [1], but had issues [2]: it also added a zero-window check and changed rcv_wnd_sent initialization, which caused test regressions. This version addresses only the partial packet handling without modifying receive window accounting. Fixes: ab174ad8ef76 ("mptcp: move ooo skbs into msk out of order queue.") Cc: stable@vger.kernel.org Link: https://lore.kernel.org/c9b426a4e163aa3c4fe8b80c79f1a610f47ae7d8.1763075056.git.pabeni@redhat.com [1] Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/600 [2] Signed-off-by: Shardul Bankar [pabeni@redhat.com: update map] Signed-off-by: Paolo Abeni Reviewed-by: Matthieu Baerts (NGI0) Signed-off-by: Matthieu Baerts (NGI0) --- v3: (Paolo) - update map_seq, too (AI tool) v2: (Shardul) - Drop the mptcp_try_coalesce() attempt for partial packets, since non-zero offset always prevents coalescing (Paolo). - https://lore.kernel.org/20260422143931.43281-1-shardul.b@mpiricsoftware.com v1: (Shardul) - https://lore.kernel.org/20260422120954.8877-1-shardul.b@mpiricsoftware.com v0: (Paolo) - https://lore.kernel.org/mptcp/c9b426a4e163aa3c4fe8b80c79f1a610f47ae7d8.1763075056.git.pabeni@redhat.com --- net/mptcp/protocol.c | 24 +++++++++++++++++++----- 1 file changed, 19 insertions(+), 5 deletions(-) diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c index 4546a8b09884..859df49e16dc 100644 --- a/net/mptcp/protocol.c +++ b/net/mptcp/protocol.c @@ -397,12 +397,26 @@ static bool __mptcp_move_skb(struct sock *sk, struct sk_buff *skb) return false; } - /* old data, keep it simple and drop the whole pkt, sender - * will retransmit as needed, if needed. + /* Completely old data? */ + if (!after64(MPTCP_SKB_CB(skb)->end_seq, msk->ack_seq)) { + MPTCP_INC_STATS(sock_net(sk), MPTCP_MIB_DUPDATA); + mptcp_drop(sk, skb); + return false; + } + + /* Partial packet: map_seq < ack_seq < end_seq. + * Skip the already-acked bytes and enqueue the new data. */ - MPTCP_INC_STATS(sock_net(sk), MPTCP_MIB_DUPDATA); - mptcp_drop(sk, skb); - return false; + copy_len = MPTCP_SKB_CB(skb)->end_seq - msk->ack_seq; + MPTCP_SKB_CB(skb)->offset += msk->ack_seq - MPTCP_SKB_CB(skb)->map_seq; + MPTCP_SKB_CB(skb)->map_seq += msk->ack_seq - + MPTCP_SKB_CB(skb)->map_seq; + msk->bytes_received += copy_len; + WRITE_ONCE(msk->ack_seq, msk->ack_seq + copy_len); + + skb_set_owner_r(skb, sk); + __skb_queue_tail(&sk->sk_receive_queue, skb); + return true; } static void mptcp_stop_rtx_timer(struct sock *sk) -- 2.53.0 From: Li Xiasong When TCP option space is insufficient (e.g., when sending ADD_ADDR with an IPv6 address and port while tcp_timestamps is enabled), the original code jumped to out_unlock without clearing the addr_signal flag. This caused mptcp_pm_add_timer to keep rescheduling indefinitely, not sending ADD_ADDR, preventing subsequent addresses in the endpoint list from being announced. Handle this case by clearing the ADD_ADDR signal and skipping the matching ADD_ADDR retransmission entry. The skip path cancels the matching timer (with id check) and advances PM state progression, preserving forward progress to subsequent PM work. This cancellation is inherently best-effort. A concurrent add_timer callback may already be running and may acquire pm.lock before the cancel path updates entry state. In that case, one final ADD_ADDR transmit attempt can still be executed. Once the cancel path sets entry->retrans_times to ADD_ADDR_RETRANS_MAX, the callback-side retrans_times check suppresses further ADD_ADDR retransmissions. Note that when an ADD_ADDR is being prepared, a pure-ACK is queued. On the output side, it means that it is fine to skip non-pure-ACK packets, when drop_other_suboptions is set: a pure-ACK will be processed soon after. Fixes: 00cfd77b9063 ("mptcp: retransmit ADD_ADDR when timeout") Cc: stable@vger.kernel.org Signed-off-by: Li Xiasong Reviewed-by: Matthieu Baerts (NGI0) Signed-off-by: Matthieu Baerts (NGI0) --- v2: add a note for sashiko-nipa's: it's a false positive. --- net/mptcp/pm.c | 56 ++++++++++++++++++++++++++++++++++++++++++++++---------- 1 file changed, 46 insertions(+), 10 deletions(-) diff --git a/net/mptcp/pm.c b/net/mptcp/pm.c index 3c152bf66cd5..3e770c7407e1 100644 --- a/net/mptcp/pm.c +++ b/net/mptcp/pm.c @@ -364,7 +364,13 @@ static void mptcp_pm_add_timer(struct timer_list *timer) spin_lock_bh(&msk->pm.lock); - if (!mptcp_pm_should_add_signal_addr(msk)) { + /* The cancel path (mptcp_pm_del_add_timer()) can race with this + * callback. Once cancel updates retrans_times to MAX, suppress further + * retransmissions here. If this callback acquires pm.lock first, one + * final transmit attempt is still possible. + */ + if (entry->retrans_times < ADD_ADDR_RETRANS_MAX && + !mptcp_pm_should_add_signal_addr(msk)) { pr_debug("retransmit ADD_ADDR id=%d\n", entry->addr.id); mptcp_pm_announce_addr(msk, &entry->addr, false); mptcp_pm_add_addr_send_ack(msk); @@ -414,8 +420,12 @@ mptcp_pm_del_add_timer(struct mptcp_sock *msk, /* Note: entry might have been removed by another thread. * We hold rcu_read_lock() to ensure it is not freed under us. */ - if (stop_timer) - sk_stop_timer_sync(sk, &entry->add_timer); + if (stop_timer) { + if (check_id) + sk_stop_timer(sk, &entry->add_timer); + else + sk_stop_timer_sync(sk, &entry->add_timer); + } rcu_read_unlock(); return entry; @@ -882,6 +892,7 @@ bool mptcp_pm_add_addr_signal(struct mptcp_sock *msk, const struct sk_buff *skb, struct mptcp_addr_info *addr, bool *echo, bool *drop_other_suboptions) { + bool skip_add_addr = false; int ret = false; u8 add_addr; u8 family; @@ -903,24 +914,49 @@ bool mptcp_pm_add_addr_signal(struct mptcp_sock *msk, const struct sk_buff *skb, } *echo = mptcp_pm_should_add_signal_echo(msk); - port = !!(*echo ? msk->pm.remote.port : msk->pm.local.port); - - family = *echo ? msk->pm.remote.family : msk->pm.local.family; - if (remaining < mptcp_add_addr_len(family, *echo, port)) - goto out_unlock; - if (*echo) { *addr = msk->pm.remote; add_addr = msk->pm.addr_signal & ~BIT(MPTCP_ADD_ADDR_ECHO); + port = !!msk->pm.remote.port; + family = msk->pm.remote.family; } else { *addr = msk->pm.local; add_addr = msk->pm.addr_signal & ~BIT(MPTCP_ADD_ADDR_SIGNAL); + port = !!msk->pm.local.port; + family = msk->pm.local.family; } - WRITE_ONCE(msk->pm.addr_signal, add_addr); + + if (remaining < mptcp_add_addr_len(family, *echo, port)) { + struct net *net = sock_net((struct sock *)msk); + + if (!*drop_other_suboptions) + goto out_unlock; + + if (*echo) { + MPTCP_INC_STATS(net, MPTCP_MIB_ECHOADDTXDROP); + } else { + skip_add_addr = true; + MPTCP_INC_STATS(net, MPTCP_MIB_ADDADDRTXDROP); + } + goto drop_signal_mark; + } + ret = true; +drop_signal_mark: + WRITE_ONCE(msk->pm.addr_signal, add_addr); + out_unlock: spin_unlock_bh(&msk->pm.lock); + + /* On pure-ACK option-space exhaustion, stop retrying this ADD_ADDR: + * clear the signal bit, cancel the matching retransmission timer, and + * let the PM state machine progress. + */ + if (skip_add_addr) { + mptcp_pm_del_add_timer(msk, addr, true); + mptcp_pm_subflow_established(msk); + } return ret; } -- 2.53.0 From: Li Xiasong Extend add_addr_ports_tests with IPv6 signaling cases that exercise ADD_ADDR tx-space shortage when tcp_timestamps are enabled. Add one case to verify PM still progresses to later signal endpoints after the first one is dropped. This covers both failure accounting and the non-blocking behavior of the announce list after a tx-space drop on pure ACK. Signed-off-by: Li Xiasong Reviewed-by: Matthieu Baerts (NGI0) Signed-off-by: Matthieu Baerts (NGI0) --- To: Shuah Khan Cc: linux-kselftest@vger.kernel.org --- tools/testing/selftests/net/mptcp/mptcp_join.sh | 31 +++++++++++++++++++++++++ 1 file changed, 31 insertions(+) diff --git a/tools/testing/selftests/net/mptcp/mptcp_join.sh b/tools/testing/selftests/net/mptcp/mptcp_join.sh index beec41f6662a..5acd12021e6e 100755 --- a/tools/testing/selftests/net/mptcp/mptcp_join.sh +++ b/tools/testing/selftests/net/mptcp/mptcp_join.sh @@ -1828,6 +1828,22 @@ chk_add_tx_nr() fi } +chk_add_drop_tx_nr() +{ + local drop_tx_nr=$1 + local count + + print_check "add addr tx drop" + count=$(mptcp_lib_get_counter ${ns1} "MPTcpExtAddAddrTxDrop") + if [ -z "$count" ]; then + print_skip + elif [ "$count" != "$drop_tx_nr" ]; then + fail_test "got $count ADD_ADDR drop[s] TX, expected $drop_tx_nr" + else + print_ok + fi +} + chk_rm_nr() { local rm_addr_nr=$1 @@ -3278,6 +3294,21 @@ add_addr_ports_tests() chk_mpc_endp_attempt ${retl} 1 fi + + # first signal address drops, second one still progresses + if reset "signal addr list progresses after tx drop"; then + pm_nl_set_limits $ns1 0 2 + pm_nl_set_limits $ns2 1 0 + ip netns exec $ns1 sysctl -q net.ipv4.tcp_timestamps=1 + ip netns exec $ns2 sysctl -q net.ipv4.tcp_timestamps=1 + + pm_nl_add_endpoint $ns1 dead:beef:2::1 flags signal port 10100 + pm_nl_add_endpoint $ns1 dead:beef:3::1 flags signal + run_tests $ns1 $ns2 dead:beef:1::1 + chk_add_drop_tx_nr 1 + chk_add_tx_nr 1 1 + chk_add_nr 1 1 0 + fi } bind_tests() -- 2.53.0 From: Paolo Abeni If the MPTCP socket fallback to TCP before the MP handshake completion, the IASN remain 0, and the rcv_wnd_sent field is not explicitly initialized, just incremented over time with the data transfer. At disconnect time such value is not cleared. If the next connection falls back to TCP before the MP handshake completion, the data transfer will keep incrementing the receive window end sequence starting from the last value used in the previous connection: the announced window will be unrelated from the actual receiver buffer size and likely too big. Address the issue zeroing the field at disconnect time. Fixes: b29fcfb54cd7 ("mptcp: full disconnect implementation") Cc: stable@vger.kernel.org Signed-off-by: Paolo Abeni Reviewed-by: Matthieu Baerts (NGI0) Signed-off-by: Matthieu Baerts (NGI0) --- net/mptcp/protocol.c | 1 + 1 file changed, 1 insertion(+) diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c index 859df49e16dc..a72a6ad6ee8b 100644 --- a/net/mptcp/protocol.c +++ b/net/mptcp/protocol.c @@ -3487,6 +3487,7 @@ static int mptcp_disconnect(struct sock *sk, int flags) /* for fallback's sake */ WRITE_ONCE(msk->ack_seq, 0); + atomic64_set(&msk->rcv_wnd_sent, 0); WRITE_ONCE(sk->sk_shutdown, 0); sk_error_report(sk); -- 2.53.0 From: Gang Yan Add __mptcp_subflow_set_rcvbuf() helper to write the subflow sk_rcvbuf, but also to call the recently added tcp_set_rcvbuf() helper to update window_clamp. This is needed because the window clap is updated when scaling_ratio changes, in tcp_measure_rcv_mss(). Until scaling_ratio changes, the subflow is stuck with the old window clamp which may be based on a small initial buffer. Use this new helper in both mptcp_sol_socket_sync_intval() (setsockopt path) and sync_socket_options() (new subflow creation path). Note that this patch depends on commit b025461303d8 ("tcp: update window_clamp when SO_RCVBUF is set"): it fixes the issue on TCP side, but the same fix is needed on MPTCP side as well. Fixes: a2cbb1603943 ("tcp: Update window clamping condition") Cc: stable@vger.kernel.org Closes: https://github.com/multipath-tcp/mptcp_net-next/issues/619 Signed-off-by: Gang Yan Reviewed-by: Matthieu Baerts (NGI0) Signed-off-by: Matthieu Baerts (NGI0) --- v2: remove 'inline' keyword (NIPA) + update Fixes tag (Jakub) --- net/mptcp/sockopt.c | 10 ++++++++-- 1 file changed, 8 insertions(+), 2 deletions(-) diff --git a/net/mptcp/sockopt.c b/net/mptcp/sockopt.c index 1cf608e7357b..87b5796d0135 100644 --- a/net/mptcp/sockopt.c +++ b/net/mptcp/sockopt.c @@ -67,6 +67,12 @@ static int mptcp_get_int_option(struct mptcp_sock *msk, sockptr_t optval, return 0; } +static void __mptcp_subflow_set_rcvbuf(struct sock *ssk, int val) +{ + WRITE_ONCE(ssk->sk_rcvbuf, val); + tcp_set_rcvbuf(ssk, val); +} + static void mptcp_sol_socket_sync_intval(struct mptcp_sock *msk, int optname, int val) { struct mptcp_subflow_context *subflow; @@ -100,7 +106,7 @@ static void mptcp_sol_socket_sync_intval(struct mptcp_sock *msk, int optname, in case SO_RCVBUF: case SO_RCVBUFFORCE: ssk->sk_userlocks |= SOCK_RCVBUF_LOCK; - WRITE_ONCE(ssk->sk_rcvbuf, sk->sk_rcvbuf); + __mptcp_subflow_set_rcvbuf(ssk, sk->sk_rcvbuf); break; case SO_MARK: if (READ_ONCE(ssk->sk_mark) != sk->sk_mark) { @@ -1560,7 +1566,7 @@ static void sync_socket_options(struct mptcp_sock *msk, struct sock *ssk) mptcp_subflow_ctx(ssk)->cached_sndbuf = sk->sk_sndbuf; } if (sk->sk_userlocks & SOCK_RCVBUF_LOCK) - WRITE_ONCE(ssk->sk_rcvbuf, sk->sk_rcvbuf); + __mptcp_subflow_set_rcvbuf(ssk, sk->sk_rcvbuf); } if (sock_flag(sk, SOCK_LINGER)) { -- 2.53.0 Using the format specifier +%s%3N with GNU date is honoured, and only prints 3 digits of the nanoseconds portion of the seconds since epoch, which corresponds to the milliseconds. The uutils implementation of date currently does not honour this, and always prints all 9 digits. This is a known issue [1], but can be worked around by adapting this test to use nanoseconds instead of microseconds, and then divide it by 1e6. This fix is similar to what has been done on systemd side [2], and it is needed to run the selftests on Ubuntu 26.04, containing uutils 0.8.0. Note that the Fixes tag is there even if this patch doesn't fix an issue in the kernel selftests, but it is useful for those using uutils 0.8.0. Fixes: 048d19d444be ("mptcp: add basic kselftest for mptcp") Cc: stable@vger.kernel.org Link: https://github.com/uutils/coreutils/issues/11658 [1] Link: https://github.com/systemd/systemd/pull/41627 [2] Signed-off-by: Matthieu Baerts (NGI0) --- To: Shuah Khan Cc: linux-kselftest@vger.kernel.org --- tools/testing/selftests/net/mptcp/mptcp_connect.sh | 6 +++--- tools/testing/selftests/net/mptcp/mptcp_lib.sh | 10 +++++----- 2 files changed, 8 insertions(+), 8 deletions(-) diff --git a/tools/testing/selftests/net/mptcp/mptcp_connect.sh b/tools/testing/selftests/net/mptcp/mptcp_connect.sh index a6447f7a31fe..d158678fa6ab 100755 --- a/tools/testing/selftests/net/mptcp/mptcp_connect.sh +++ b/tools/testing/selftests/net/mptcp/mptcp_connect.sh @@ -401,7 +401,7 @@ do_transfer() mptcp_lib_wait_local_port_listen "${listener_ns}" "${port}" local start - start=$(date +%s%3N) + start=$(date +%s%N) ip netns exec ${connector_ns} \ ./mptcp_connect -t ${timeout_poll} -p $port -s ${cl_proto} \ $extra_args $connect_addr < "$cin" > "$cout" & @@ -423,7 +423,7 @@ do_transfer() fi local stop - stop=$(date +%s%3N) + stop=$(date +%s%N) if $capture; then sleep 1 @@ -439,7 +439,7 @@ do_transfer() fi local duration - duration=$((stop-start)) + duration=$(((stop-start) / 1000000)) printf "(duration %05sms) " "${duration}" if [ ${rets} -ne 0 ] || [ ${retc} -ne 0 ] || [ ${timeout_pid} -ne 0 ]; then mptcp_lib_pr_fail "client exit code $retc, server $rets" diff --git a/tools/testing/selftests/net/mptcp/mptcp_lib.sh b/tools/testing/selftests/net/mptcp/mptcp_lib.sh index 989a5975dcea..5ef6033775c8 100644 --- a/tools/testing/selftests/net/mptcp/mptcp_lib.sh +++ b/tools/testing/selftests/net/mptcp/mptcp_lib.sh @@ -28,7 +28,7 @@ declare -rx MPTCP_LIB_AF_INET6=10 MPTCP_LIB_SUBTESTS=() MPTCP_LIB_SUBTESTS_DUPLICATED=0 MPTCP_LIB_SUBTEST_FLAKY=0 -MPTCP_LIB_SUBTESTS_LAST_TS_MS= +MPTCP_LIB_SUBTESTS_LAST_TS_NS= MPTCP_LIB_TEST_COUNTER=0 MPTCP_LIB_TEST_FORMAT="%02u %-50s" MPTCP_LIB_IP_MPTCP=0 @@ -236,7 +236,7 @@ mptcp_lib_kversion_ge() { } mptcp_lib_subtests_last_ts_reset() { - MPTCP_LIB_SUBTESTS_LAST_TS_MS="$(date +%s%3N)" + MPTCP_LIB_SUBTESTS_LAST_TS_NS="$(date +%s%N)" } mptcp_lib_subtests_last_ts_reset @@ -255,7 +255,7 @@ __mptcp_lib_result_check_duplicated() { __mptcp_lib_result_add() { local result="${1}" local time="time=" - local ts_prev_ms + local ts_prev_ns shift local id=$((${#MPTCP_LIB_SUBTESTS[@]} + 1)) @@ -265,9 +265,9 @@ __mptcp_lib_result_add() { # not to add two '#' [[ "${*}" != *"#"* ]] && time="# ${time}" - ts_prev_ms="${MPTCP_LIB_SUBTESTS_LAST_TS_MS}" + ts_prev_ns="${MPTCP_LIB_SUBTESTS_LAST_TS_NS}" mptcp_lib_subtests_last_ts_reset - time+="$((MPTCP_LIB_SUBTESTS_LAST_TS_MS - ts_prev_ms))ms" + time+="$(((MPTCP_LIB_SUBTESTS_LAST_TS_NS - ts_prev_ns) / 1000000))ms" MPTCP_LIB_SUBTESTS+=("${result} ${id} - ${KSFT_TEST}: ${*} ${time}") } -- 2.53.0