From 0f1e4d06591d0a7907c71f7b6d1c79f8a4de8098 Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Wed, 20 Jul 2022 09:50:19 -0700 Subject: tcp: Fix data-races around sysctl_tcp_workaround_signed_windows. While reading sysctl_tcp_workaround_signed_windows, it can be changed concurrently. Thus, we need to add READ_ONCE() to its readers. Fixes: 15d99e02baba ("[TCP]: sysctl to allow TCP window > 32767 sans wscale") Signed-off-by: Kuniyuki Iwashima Signed-off-by: David S. Miller --- net/ipv4/tcp_output.c | 4 ++-- 1 file changed, 2 insertions(+), 2 deletions(-) (limited to 'net/ipv4/tcp_output.c') diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index c38e07b50639..88f7d51e6691 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -230,7 +230,7 @@ void tcp_select_initial_window(const struct sock *sk, int __space, __u32 mss, * which we interpret as a sign the remote TCP is not * misinterpreting the window field as a signed quantity. */ - if (sock_net(sk)->ipv4.sysctl_tcp_workaround_signed_windows) + if (READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_workaround_signed_windows)) (*rcv_wnd) = min(space, MAX_TCP_WINDOW); else (*rcv_wnd) = min_t(u32, space, U16_MAX); @@ -285,7 +285,7 @@ static u16 tcp_select_window(struct sock *sk) * scaled window. */ if (!tp->rx_opt.rcv_wscale && - sock_net(sk)->ipv4.sysctl_tcp_workaround_signed_windows) + READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_workaround_signed_windows)) new_win = min(new_win, MAX_TCP_WINDOW); else new_win = min(new_win, (65535U << tp->rx_opt.rcv_wscale)); -- cgit v1.2.3 From 9fb90193fbd66b4c5409ef729fd081861f8b6351 Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Wed, 20 Jul 2022 09:50:20 -0700 Subject: tcp: Fix a data-race around sysctl_tcp_limit_output_bytes. While reading sysctl_tcp_limit_output_bytes, it can be changed concurrently. Thus, we need to add READ_ONCE() to its reader. Fixes: 46d3ceabd8d9 ("tcp: TCP Small Queues") Signed-off-by: Kuniyuki Iwashima Signed-off-by: David S. Miller --- net/ipv4/tcp_output.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net/ipv4/tcp_output.c') diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 88f7d51e6691..80c9bed73337 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -2507,7 +2507,7 @@ static bool tcp_small_queue_check(struct sock *sk, const struct sk_buff *skb, sk->sk_pacing_rate >> READ_ONCE(sk->sk_pacing_shift)); if (sk->sk_pacing_status == SK_PACING_NONE) limit = min_t(unsigned long, limit, - sock_net(sk)->ipv4.sysctl_tcp_limit_output_bytes); + READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_limit_output_bytes)); limit <<= factor; if (static_branch_unlikely(&tcp_tx_delay_enabled) && -- cgit v1.2.3 From e0bb4ab9dfddd872622239f49fb2bd403b70853b Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Wed, 20 Jul 2022 09:50:22 -0700 Subject: tcp: Fix a data-race around sysctl_tcp_min_tso_segs. While reading sysctl_tcp_min_tso_segs, it can be changed concurrently. Thus, we need to add READ_ONCE() to its reader. Fixes: 95bd09eb2750 ("tcp: TSO packets automatic sizing") Signed-off-by: Kuniyuki Iwashima Signed-off-by: David S. Miller --- net/ipv4/tcp_output.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net/ipv4/tcp_output.c') diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 80c9bed73337..6e29cf391a64 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -1995,7 +1995,7 @@ static u32 tcp_tso_segs(struct sock *sk, unsigned int mss_now) min_tso = ca_ops->min_tso_segs ? ca_ops->min_tso_segs(sk) : - sock_net(sk)->ipv4.sysctl_tcp_min_tso_segs; + READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_min_tso_segs); tso_segs = tcp_tso_autosize(sk, mss_now, min_tso); return min_t(u32, tso_segs, sk->sk_gso_max_segs); -- cgit v1.2.3 From 2455e61b85e9c99af38cd889a7101f1d48b33cb4 Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Wed, 20 Jul 2022 09:50:23 -0700 Subject: tcp: Fix a data-race around sysctl_tcp_tso_rtt_log. While reading sysctl_tcp_tso_rtt_log, it can be changed concurrently. Thus, we need to add READ_ONCE() to its reader. Fixes: 65466904b015 ("tcp: adjust TSO packet sizes based on min_rtt") Signed-off-by: Kuniyuki Iwashima Signed-off-by: David S. Miller --- net/ipv4/tcp_output.c | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) (limited to 'net/ipv4/tcp_output.c') diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 6e29cf391a64..cf6713c9567e 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -1976,7 +1976,7 @@ static u32 tcp_tso_autosize(const struct sock *sk, unsigned int mss_now, bytes = sk->sk_pacing_rate >> READ_ONCE(sk->sk_pacing_shift); - r = tcp_min_rtt(tcp_sk(sk)) >> sock_net(sk)->ipv4.sysctl_tcp_tso_rtt_log; + r = tcp_min_rtt(tcp_sk(sk)) >> READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_tso_rtt_log); if (r < BITS_PER_TYPE(sk->sk_gso_max_size)) bytes += sk->sk_gso_max_size >> r; -- cgit v1.2.3 From 4d8f24eeedc58d5f87b650ddda73c16e8ba56559 Mon Sep 17 00:00:00 2001 From: Wei Wang Date: Thu, 21 Jul 2022 20:44:04 +0000 Subject: Revert "tcp: change pingpong threshold to 3" This reverts commit 4a41f453bedfd5e9cd040bad509d9da49feb3e2c. This to-be-reverted commit was meant to apply a stricter rule for the stack to enter pingpong mode. However, the condition used to check for interactive session "before(tp->lsndtime, icsk->icsk_ack.lrcvtime)" is jiffy based and might be too coarse, which delays the stack entering pingpong mode. We revert this patch so that we no longer use the above condition to determine interactive session, and also reduce pingpong threshold to 1. Fixes: 4a41f453bedf ("tcp: change pingpong threshold to 3") Reported-by: LemmyHuang Suggested-by: Neal Cardwell Signed-off-by: Wei Wang Acked-by: Neal Cardwell Reviewed-by: Eric Dumazet Link: https://lore.kernel.org/r/20220721204404.388396-1-weiwan@google.com Signed-off-by: Jakub Kicinski --- include/net/inet_connection_sock.h | 10 +--------- net/ipv4/tcp_output.c | 15 ++++++--------- 2 files changed, 7 insertions(+), 18 deletions(-) (limited to 'net/ipv4/tcp_output.c') diff --git a/include/net/inet_connection_sock.h b/include/net/inet_connection_sock.h index 85cd695e7fd1..ee88f0f1350f 100644 --- a/include/net/inet_connection_sock.h +++ b/include/net/inet_connection_sock.h @@ -321,7 +321,7 @@ void inet_csk_update_fastreuse(struct inet_bind_bucket *tb, struct dst_entry *inet_csk_update_pmtu(struct sock *sk, u32 mtu); -#define TCP_PINGPONG_THRESH 3 +#define TCP_PINGPONG_THRESH 1 static inline void inet_csk_enter_pingpong_mode(struct sock *sk) { @@ -338,14 +338,6 @@ static inline bool inet_csk_in_pingpong_mode(struct sock *sk) return inet_csk(sk)->icsk_ack.pingpong >= TCP_PINGPONG_THRESH; } -static inline void inet_csk_inc_pingpong_cnt(struct sock *sk) -{ - struct inet_connection_sock *icsk = inet_csk(sk); - - if (icsk->icsk_ack.pingpong < U8_MAX) - icsk->icsk_ack.pingpong++; -} - static inline bool inet_csk_has_ulp(struct sock *sk) { return inet_sk(sk)->is_icsk && !!inet_csk(sk)->icsk_ulp_ops; diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index cf6713c9567e..2efe41c84ee8 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -167,16 +167,13 @@ static void tcp_event_data_sent(struct tcp_sock *tp, if (tcp_packets_in_flight(tp) == 0) tcp_ca_event(sk, CA_EVENT_TX_START); - /* If this is the first data packet sent in response to the - * previous received data, - * and it is a reply for ato after last received packet, - * increase pingpong count. - */ - if (before(tp->lsndtime, icsk->icsk_ack.lrcvtime) && - (u32)(now - icsk->icsk_ack.lrcvtime) < icsk->icsk_ack.ato) - inet_csk_inc_pingpong_cnt(sk); - tp->lsndtime = now; + + /* If it is a reply for ato after last received + * packet, enter pingpong mode. + */ + if ((u32)(now - icsk->icsk_ack.lrcvtime) < icsk->icsk_ack.ato) + inet_csk_enter_pingpong_mode(sk); } /* Account for an ACK we sent. */ -- cgit v1.2.3 From 02739545951ad4c1215160db7fbf9b7a918d3c0b Mon Sep 17 00:00:00 2001 From: Kuniyuki Iwashima Date: Fri, 22 Jul 2022 11:22:00 -0700 Subject: net: Fix data-races around sysctl_[rw]mem(_offset)?. While reading these sysctl variables, they can be changed concurrently. Thus, we need to add READ_ONCE() to their readers. - .sysctl_rmem - .sysctl_rwmem - .sysctl_rmem_offset - .sysctl_wmem_offset - sysctl_tcp_rmem[1, 2] - sysctl_tcp_wmem[1, 2] - sysctl_decnet_rmem[1] - sysctl_decnet_wmem[1] - sysctl_tipc_rmem[1] Fixes: 1da177e4c3f4 ("Linux-2.6.12-rc2") Signed-off-by: Kuniyuki Iwashima Signed-off-by: David S. Miller --- include/net/sock.h | 8 ++++---- net/decnet/af_decnet.c | 4 ++-- net/ipv4/tcp.c | 6 +++--- net/ipv4/tcp_input.c | 13 +++++++------ net/ipv4/tcp_output.c | 2 +- net/mptcp/protocol.c | 6 +++--- net/tipc/socket.c | 2 +- 7 files changed, 21 insertions(+), 20 deletions(-) (limited to 'net/ipv4/tcp_output.c') diff --git a/include/net/sock.h b/include/net/sock.h index 9fa54762e077..7a48991cdb19 100644 --- a/include/net/sock.h +++ b/include/net/sock.h @@ -2843,18 +2843,18 @@ static inline int sk_get_wmem0(const struct sock *sk, const struct proto *proto) { /* Does this proto have per netns sysctl_wmem ? */ if (proto->sysctl_wmem_offset) - return *(int *)((void *)sock_net(sk) + proto->sysctl_wmem_offset); + return READ_ONCE(*(int *)((void *)sock_net(sk) + proto->sysctl_wmem_offset)); - return *proto->sysctl_wmem; + return READ_ONCE(*proto->sysctl_wmem); } static inline int sk_get_rmem0(const struct sock *sk, const struct proto *proto) { /* Does this proto have per netns sysctl_rmem ? */ if (proto->sysctl_rmem_offset) - return *(int *)((void *)sock_net(sk) + proto->sysctl_rmem_offset); + return READ_ONCE(*(int *)((void *)sock_net(sk) + proto->sysctl_rmem_offset)); - return *proto->sysctl_rmem; + return READ_ONCE(*proto->sysctl_rmem); } /* Default TCP Small queue budget is ~1 ms of data (1sec >> 10) diff --git a/net/decnet/af_decnet.c b/net/decnet/af_decnet.c index dc92a67baea3..7d542eb46172 100644 --- a/net/decnet/af_decnet.c +++ b/net/decnet/af_decnet.c @@ -480,8 +480,8 @@ static struct sock *dn_alloc_sock(struct net *net, struct socket *sock, gfp_t gf sk->sk_family = PF_DECnet; sk->sk_protocol = 0; sk->sk_allocation = gfp; - sk->sk_sndbuf = sysctl_decnet_wmem[1]; - sk->sk_rcvbuf = sysctl_decnet_rmem[1]; + sk->sk_sndbuf = READ_ONCE(sysctl_decnet_wmem[1]); + sk->sk_rcvbuf = READ_ONCE(sysctl_decnet_rmem[1]); /* Initialization of DECnet Session Control Port */ scp = DN_SK(sk); diff --git a/net/ipv4/tcp.c b/net/ipv4/tcp.c index a11e5de3a4c3..002a4a04efbe 100644 --- a/net/ipv4/tcp.c +++ b/net/ipv4/tcp.c @@ -452,8 +452,8 @@ void tcp_init_sock(struct sock *sk) icsk->icsk_sync_mss = tcp_sync_mss; - WRITE_ONCE(sk->sk_sndbuf, sock_net(sk)->ipv4.sysctl_tcp_wmem[1]); - WRITE_ONCE(sk->sk_rcvbuf, sock_net(sk)->ipv4.sysctl_tcp_rmem[1]); + WRITE_ONCE(sk->sk_sndbuf, READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_wmem[1])); + WRITE_ONCE(sk->sk_rcvbuf, READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_rmem[1])); sk_sockets_allocated_inc(sk); } @@ -1724,7 +1724,7 @@ int tcp_set_rcvlowat(struct sock *sk, int val) if (sk->sk_userlocks & SOCK_RCVBUF_LOCK) cap = sk->sk_rcvbuf >> 1; else - cap = sock_net(sk)->ipv4.sysctl_tcp_rmem[2] >> 1; + cap = READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_rmem[2]) >> 1; val = min(val, cap); WRITE_ONCE(sk->sk_rcvlowat, val ? : 1); diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index dd05238f79f6..ff2e0d87aee4 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -426,7 +426,7 @@ static void tcp_sndbuf_expand(struct sock *sk) if (sk->sk_sndbuf < sndmem) WRITE_ONCE(sk->sk_sndbuf, - min(sndmem, sock_net(sk)->ipv4.sysctl_tcp_wmem[2])); + min(sndmem, READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_wmem[2]))); } /* 2. Tuning advertised window (window_clamp, rcv_ssthresh) @@ -461,7 +461,7 @@ static int __tcp_grow_window(const struct sock *sk, const struct sk_buff *skb, struct tcp_sock *tp = tcp_sk(sk); /* Optimize this! */ int truesize = tcp_win_from_space(sk, skbtruesize) >> 1; - int window = tcp_win_from_space(sk, sock_net(sk)->ipv4.sysctl_tcp_rmem[2]) >> 1; + int window = tcp_win_from_space(sk, READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_rmem[2])) >> 1; while (tp->rcv_ssthresh <= window) { if (truesize <= skb->len) @@ -574,16 +574,17 @@ static void tcp_clamp_window(struct sock *sk) struct tcp_sock *tp = tcp_sk(sk); struct inet_connection_sock *icsk = inet_csk(sk); struct net *net = sock_net(sk); + int rmem2; icsk->icsk_ack.quick = 0; + rmem2 = READ_ONCE(net->ipv4.sysctl_tcp_rmem[2]); - if (sk->sk_rcvbuf < net->ipv4.sysctl_tcp_rmem[2] && + if (sk->sk_rcvbuf < rmem2 && !(sk->sk_userlocks & SOCK_RCVBUF_LOCK) && !tcp_under_memory_pressure(sk) && sk_memory_allocated(sk) < sk_prot_mem_limits(sk, 0)) { WRITE_ONCE(sk->sk_rcvbuf, - min(atomic_read(&sk->sk_rmem_alloc), - net->ipv4.sysctl_tcp_rmem[2])); + min(atomic_read(&sk->sk_rmem_alloc), rmem2)); } if (atomic_read(&sk->sk_rmem_alloc) > sk->sk_rcvbuf) tp->rcv_ssthresh = min(tp->window_clamp, 2U * tp->advmss); @@ -745,7 +746,7 @@ void tcp_rcv_space_adjust(struct sock *sk) do_div(rcvwin, tp->advmss); rcvbuf = min_t(u64, rcvwin * rcvmem, - sock_net(sk)->ipv4.sysctl_tcp_rmem[2]); + READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_rmem[2])); if (rcvbuf > sk->sk_rcvbuf) { WRITE_ONCE(sk->sk_rcvbuf, rcvbuf); diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 2efe41c84ee8..4c376b6d8764 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -238,7 +238,7 @@ void tcp_select_initial_window(const struct sock *sk, int __space, __u32 mss, *rcv_wscale = 0; if (wscale_ok) { /* Set window scaling on max possible window */ - space = max_t(u32, space, sock_net(sk)->ipv4.sysctl_tcp_rmem[2]); + space = max_t(u32, space, READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_rmem[2])); space = max_t(u32, space, sysctl_rmem_max); space = min_t(u32, space, *window_clamp); *rcv_wscale = clamp_t(int, ilog2(space) - 15, diff --git a/net/mptcp/protocol.c b/net/mptcp/protocol.c index 9bbd8cbe0acb..7e1518bb6115 100644 --- a/net/mptcp/protocol.c +++ b/net/mptcp/protocol.c @@ -1926,7 +1926,7 @@ static void mptcp_rcv_space_adjust(struct mptcp_sock *msk, int copied) do_div(rcvwin, advmss); rcvbuf = min_t(u64, rcvwin * rcvmem, - sock_net(sk)->ipv4.sysctl_tcp_rmem[2]); + READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_rmem[2])); if (rcvbuf > sk->sk_rcvbuf) { u32 window_clamp; @@ -2669,8 +2669,8 @@ static int mptcp_init_sock(struct sock *sk) mptcp_ca_reset(sk); sk_sockets_allocated_inc(sk); - sk->sk_rcvbuf = sock_net(sk)->ipv4.sysctl_tcp_rmem[1]; - sk->sk_sndbuf = sock_net(sk)->ipv4.sysctl_tcp_wmem[1]; + sk->sk_rcvbuf = READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_rmem[1]); + sk->sk_sndbuf = READ_ONCE(sock_net(sk)->ipv4.sysctl_tcp_wmem[1]); return 0; } diff --git a/net/tipc/socket.c b/net/tipc/socket.c index 43509c7e90fc..f1c3b8eb4b3d 100644 --- a/net/tipc/socket.c +++ b/net/tipc/socket.c @@ -517,7 +517,7 @@ static int tipc_sk_create(struct net *net, struct socket *sock, timer_setup(&sk->sk_timer, tipc_sk_timeout, 0); sk->sk_shutdown = 0; sk->sk_backlog_rcv = tipc_sk_backlog_rcv; - sk->sk_rcvbuf = sysctl_tipc_rmem[1]; + sk->sk_rcvbuf = READ_ONCE(sysctl_tipc_rmem[1]); sk->sk_data_ready = tipc_data_ready; sk->sk_write_space = tipc_write_space; sk->sk_destruct = tipc_sock_destruct; -- cgit v1.2.3