From f5da7c45188eea71394bf445655cae2df88a7788 Mon Sep 17 00:00:00 2001 From: Jakub Kicinski Date: Mon, 17 Feb 2025 15:29:05 -0800 Subject: tcp: adjust rcvq_space after updating scaling ratio Since commit under Fixes we set the window clamp in accordance to newly measured rcvbuf scaling_ratio. If the scaling_ratio decreased significantly we may put ourselves in a situation where windows become smaller than rcvq_space, preventing tcp_rcv_space_adjust() from increasing rcvbuf. The significant decrease of scaling_ratio is far more likely since commit 697a6c8cec03 ("tcp: increase the default TCP scaling ratio"), which increased the "default" scaling ratio from ~30% to 50%. Hitting the bad condition depends a lot on TCP tuning, and drivers at play. One of Meta's workloads hits it reliably under following conditions: - default rcvbuf of 125k - sender MTU 1500, receiver MTU 5000 - driver settles on scaling_ratio of 78 for the config above. Initial rcvq_space gets calculated as TCP_INIT_CWND * tp->advmss (10 * 5k = 50k). Once we find out the true scaling ratio and MSS we clamp the windows to 38k. Triggering the condition also depends on the message sequence of this workload. I can't repro the problem with simple iperf or TCP_RR-style tests. Fixes: a2cbb1603943 ("tcp: Update window clamping condition") Reviewed-by: Eric Dumazet Reviewed-by: Neal Cardwell Link: https://patch.msgid.link/20250217232905.3162187-1-kuba@kernel.org Signed-off-by: Jakub Kicinski --- net/ipv4/tcp_input.c | 12 +++++++++--- 1 file changed, 9 insertions(+), 3 deletions(-) (limited to 'net/ipv4/tcp_input.c') diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index eb82e01da911..98b8cc740392 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -243,9 +243,15 @@ static void tcp_measure_rcv_mss(struct sock *sk, const struct sk_buff *skb) do_div(val, skb->truesize); tcp_sk(sk)->scaling_ratio = val ? val : 1; - if (old_ratio != tcp_sk(sk)->scaling_ratio) - WRITE_ONCE(tcp_sk(sk)->window_clamp, - tcp_win_from_space(sk, sk->sk_rcvbuf)); + if (old_ratio != tcp_sk(sk)->scaling_ratio) { + struct tcp_sock *tp = tcp_sk(sk); + + val = tcp_win_from_space(sk, sk->sk_rcvbuf); + tcp_set_window_clamp(sk, val); + + if (tp->window_clamp < tp->rcvq_space.space) + tp->rcvq_space.space = tp->window_clamp; + } } icsk->icsk_ack.rcv_mss = min_t(unsigned int, len, tcp_sk(sk)->advmss); -- cgit v1.2.3-59-g8ed1b From 9b6412e6979f6f9e0632075f8f008937b5cd4efd Mon Sep 17 00:00:00 2001 From: Sabrina Dubroca Date: Mon, 17 Feb 2025 11:23:35 +0100 Subject: tcp: drop secpath at the same time as we currently drop dst Xiumei reported hitting the WARN in xfrm6_tunnel_net_exit while running tests that boil down to: - create a pair of netns - run a basic TCP test over ipcomp6 - delete the pair of netns The xfrm_state found on spi_byaddr was not deleted at the time we delete the netns, because we still have a reference on it. This lingering reference comes from a secpath (which holds a ref on the xfrm_state), which is still attached to an skb. This skb is not leaked, it ends up on sk_receive_queue and then gets defer-free'd by skb_attempt_defer_free. The problem happens when we defer freeing an skb (push it on one CPU's defer_list), and don't flush that list before the netns is deleted. In that case, we still have a reference on the xfrm_state that we don't expect at this point. We already drop the skb's dst in the TCP receive path when it's no longer needed, so let's also drop the secpath. At this point, tcp_filter has already called into the LSM hooks that may require the secpath, so it should not be needed anymore. However, in some of those places, the MPTCP extension has just been attached to the skb, so we cannot simply drop all extensions. Fixes: 68822bdf76f1 ("net: generalize skb freeing deferral to per-cpu lists") Reported-by: Xiumei Mu Signed-off-by: Sabrina Dubroca Reviewed-by: Eric Dumazet Link: https://patch.msgid.link/5055ba8f8f72bdcb602faa299faca73c280b7735.1739743613.git.sd@queasysnail.net Signed-off-by: Paolo Abeni --- include/net/tcp.h | 14 ++++++++++++++ net/ipv4/tcp_fastopen.c | 4 ++-- net/ipv4/tcp_input.c | 8 ++++---- net/ipv4/tcp_ipv4.c | 2 +- 4 files changed, 21 insertions(+), 7 deletions(-) (limited to 'net/ipv4/tcp_input.c') diff --git a/include/net/tcp.h b/include/net/tcp.h index 5b2b04835688..930cda5b5eb9 100644 --- a/include/net/tcp.h +++ b/include/net/tcp.h @@ -41,6 +41,7 @@ #include #include #include +#include #include #include @@ -683,6 +684,19 @@ void tcp_fin(struct sock *sk); void tcp_check_space(struct sock *sk); void tcp_sack_compress_send_ack(struct sock *sk); +static inline void tcp_cleanup_skb(struct sk_buff *skb) +{ + skb_dst_drop(skb); + secpath_reset(skb); +} + +static inline void tcp_add_receive_queue(struct sock *sk, struct sk_buff *skb) +{ + DEBUG_NET_WARN_ON_ONCE(skb_dst(skb)); + DEBUG_NET_WARN_ON_ONCE(secpath_exists(skb)); + __skb_queue_tail(&sk->sk_receive_queue, skb); +} + /* tcp_timer.c */ void tcp_init_xmit_timers(struct sock *); static inline void tcp_clear_xmit_timers(struct sock *sk) diff --git a/net/ipv4/tcp_fastopen.c b/net/ipv4/tcp_fastopen.c index 0f523cbfe329..32b28fc21b63 100644 --- a/net/ipv4/tcp_fastopen.c +++ b/net/ipv4/tcp_fastopen.c @@ -178,7 +178,7 @@ void tcp_fastopen_add_skb(struct sock *sk, struct sk_buff *skb) if (!skb) return; - skb_dst_drop(skb); + tcp_cleanup_skb(skb); /* segs_in has been initialized to 1 in tcp_create_openreq_child(). * Hence, reset segs_in to 0 before calling tcp_segs_in() * to avoid double counting. Also, tcp_segs_in() expects @@ -195,7 +195,7 @@ void tcp_fastopen_add_skb(struct sock *sk, struct sk_buff *skb) TCP_SKB_CB(skb)->tcp_flags &= ~TCPHDR_SYN; tp->rcv_nxt = TCP_SKB_CB(skb)->end_seq; - __skb_queue_tail(&sk->sk_receive_queue, skb); + tcp_add_receive_queue(sk, skb); tp->syn_data_acked = 1; /* u64_stats_update_begin(&tp->syncp) not needed here, diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 98b8cc740392..0cbf81bf3d45 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -4976,7 +4976,7 @@ static void tcp_ofo_queue(struct sock *sk) tcp_rcv_nxt_update(tp, TCP_SKB_CB(skb)->end_seq); fin = TCP_SKB_CB(skb)->tcp_flags & TCPHDR_FIN; if (!eaten) - __skb_queue_tail(&sk->sk_receive_queue, skb); + tcp_add_receive_queue(sk, skb); else kfree_skb_partial(skb, fragstolen); @@ -5168,7 +5168,7 @@ static int __must_check tcp_queue_rcv(struct sock *sk, struct sk_buff *skb, skb, fragstolen)) ? 1 : 0; tcp_rcv_nxt_update(tcp_sk(sk), TCP_SKB_CB(skb)->end_seq); if (!eaten) { - __skb_queue_tail(&sk->sk_receive_queue, skb); + tcp_add_receive_queue(sk, skb); skb_set_owner_r(skb, sk); } return eaten; @@ -5251,7 +5251,7 @@ static void tcp_data_queue(struct sock *sk, struct sk_buff *skb) __kfree_skb(skb); return; } - skb_dst_drop(skb); + tcp_cleanup_skb(skb); __skb_pull(skb, tcp_hdr(skb)->doff * 4); reason = SKB_DROP_REASON_NOT_SPECIFIED; @@ -6232,7 +6232,7 @@ void tcp_rcv_established(struct sock *sk, struct sk_buff *skb) NET_INC_STATS(sock_net(sk), LINUX_MIB_TCPHPHITS); /* Bulk data transfer: receiver */ - skb_dst_drop(skb); + tcp_cleanup_skb(skb); __skb_pull(skb, tcp_header_len); eaten = tcp_queue_rcv(sk, skb, &fragstolen); diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c index cc2b5194a18d..2632844d2c35 100644 --- a/net/ipv4/tcp_ipv4.c +++ b/net/ipv4/tcp_ipv4.c @@ -2027,7 +2027,7 @@ bool tcp_add_backlog(struct sock *sk, struct sk_buff *skb, */ skb_condense(skb); - skb_dst_drop(skb); + tcp_cleanup_skb(skb); if (unlikely(tcp_checksum_complete(skb))) { bh_unlock_sock(sk); -- cgit v1.2.3-59-g8ed1b