From 0baf26b0fcd74bbfcef53c5d5e8bad2b99c8d0d2 Mon Sep 17 00:00:00 2001 From: Martin KaFai Lau Date: Wed, 8 Jan 2020 16:35:08 -0800 Subject: bpf: tcp: Support tcp_congestion_ops in bpf This patch makes "struct tcp_congestion_ops" to be the first user of BPF STRUCT_OPS. It allows implementing a tcp_congestion_ops in bpf. The BPF implemented tcp_congestion_ops can be used like regular kernel tcp-cc through sysctl and setsockopt. e.g. [root@arch-fb-vm1 bpf]# sysctl -a | egrep congestion net.ipv4.tcp_allowed_congestion_control = reno cubic bpf_cubic net.ipv4.tcp_available_congestion_control = reno bic cubic bpf_cubic net.ipv4.tcp_congestion_control = bpf_cubic There has been attempt to move the TCP CC to the user space (e.g. CCP in TCP). The common arguments are faster turn around, get away from long-tail kernel versions in production...etc, which are legit points. BPF has been the continuous effort to join both kernel and userspace upsides together (e.g. XDP to gain the performance advantage without bypassing the kernel). The recent BPF advancements (in particular BTF-aware verifier, BPF trampoline, BPF CO-RE...) made implementing kernel struct ops (e.g. tcp cc) possible in BPF. It allows a faster turnaround for testing algorithm in the production while leveraging the existing (and continue growing) BPF feature/framework instead of building one specifically for userspace TCP CC. This patch allows write access to a few fields in tcp-sock (in bpf_tcp_ca_btf_struct_access()). The optional "get_info" is unsupported now. It can be added later. One possible way is to output the info with a btf-id to describe the content. Signed-off-by: Martin KaFai Lau Signed-off-by: Alexei Starovoitov Acked-by: Andrii Nakryiko Acked-by: Yonghong Song Link: https://lore.kernel.org/bpf/20200109003508.3856115-1-kafai@fb.com --- 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 58c92a7d671c..377cfab422df 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -3368,8 +3368,8 @@ static void tcp_ca_dst_init(struct sock *sk, const struct dst_entry *dst) rcu_read_lock(); ca = tcp_ca_find_key(ca_key); - if (likely(ca && try_module_get(ca->owner))) { - module_put(icsk->icsk_ca_ops->owner); + if (likely(ca && bpf_try_module_get(ca, ca->owner))) { + bpf_module_put(icsk->icsk_ca_ops, icsk->icsk_ca_ops->owner); icsk->icsk_ca_dst_locked = tcp_ca_dst_locked(dst); icsk->icsk_ca_ops = ca; } -- cgit From 85712484110df308215077be6ee21c4e57d7dec2 Mon Sep 17 00:00:00 2001 From: Mat Martineau Date: Thu, 9 Jan 2020 07:59:20 -0800 Subject: tcp: coalesce/collapse must respect MPTCP extensions Coalesce and collapse of packets carrying MPTCP extensions is allowed when the newer packet has no extension or the extensions carried by both packets are equal. This allows merging of TSO packet trains and even cross-TSO packets, and does not require any additional action when moving data into existing SKBs. v3 -> v4: - allow collapsing, under mptcp_skb_can_collapse() constraint v5 -> v6: - clarify MPTCP skb extensions must always be cleared at allocation time Co-developed-by: Paolo Abeni Signed-off-by: Paolo Abeni Signed-off-by: Mat Martineau 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 58c92a7d671c..3ce7fe1c4076 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -2865,7 +2865,7 @@ static void tcp_retrans_try_collapse(struct sock *sk, struct sk_buff *to, if (!tcp_can_collapse(sk, skb)) break; - if (!tcp_skb_can_collapse_to(to)) + if (!tcp_skb_can_collapse(to, skb)) break; space -= skb->len; -- cgit From 9cfcca2389d7e07647ee69950f46ab5e6dfe03ac Mon Sep 17 00:00:00 2001 From: Mat Martineau Date: Thu, 9 Jan 2020 07:59:22 -0800 Subject: tcp: Check for filled TCP option space before SACK Update the SACK check to work with zero option space available, a case that's possible with MPTCP but not MD5+TS. Maintained only one conditional branch for insufficient SACK space. v1 -> v2: - Moves the check inside the SACK branch by taking recent SACK fix: 9424e2e7ad93 (tcp: md5: fix potential overestimation of TCP option space) in to account, but modifies it to work in MPTCP scenarios beyond the MD5+TS corner case. Co-developed-by: Paolo Abeni Signed-off-by: Paolo Abeni Reviewed-by: Eric Dumazet Signed-off-by: Mat Martineau Signed-off-by: David S. Miller --- net/ipv4/tcp_output.c | 10 +++++++--- 1 file changed, 7 insertions(+), 3 deletions(-) (limited to 'net/ipv4/tcp_output.c') diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 3ce7fe1c4076..05109d0c675b 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -754,13 +754,17 @@ static unsigned int tcp_established_options(struct sock *sk, struct sk_buff *skb eff_sacks = tp->rx_opt.num_sacks + tp->rx_opt.dsack; if (unlikely(eff_sacks)) { const unsigned int remaining = MAX_TCP_OPTION_SPACE - size; + if (unlikely(remaining < TCPOLEN_SACK_BASE_ALIGNED + + TCPOLEN_SACK_PERBLOCK)) + return size; + opts->num_sack_blocks = min_t(unsigned int, eff_sacks, (remaining - TCPOLEN_SACK_BASE_ALIGNED) / TCPOLEN_SACK_PERBLOCK); - if (likely(opts->num_sack_blocks)) - size += TCPOLEN_SACK_BASE_ALIGNED + - opts->num_sack_blocks * TCPOLEN_SACK_PERBLOCK; + + size += TCPOLEN_SACK_BASE_ALIGNED + + opts->num_sack_blocks * TCPOLEN_SACK_PERBLOCK; } return size; -- cgit From eda7acddf8080bb2d022a8d4b8b2345eb80c63ec Mon Sep 17 00:00:00 2001 From: Peter Krystad Date: Tue, 21 Jan 2020 16:56:16 -0800 Subject: mptcp: Handle MPTCP TCP options Add hooks to parse and format the MP_CAPABLE option. This option is handled according to MPTCP version 0 (RFC6824). MPTCP version 1 MP_CAPABLE (RFC6824bis/RFC8684) will be added later in coordination with related code changes. Co-developed-by: Matthieu Baerts Signed-off-by: Matthieu Baerts Co-developed-by: Florian Westphal Signed-off-by: Florian Westphal Co-developed-by: Davide Caratti Signed-off-by: Davide Caratti Signed-off-by: Peter Krystad Signed-off-by: Christoph Paasch Signed-off-by: David S. Miller --- net/ipv4/tcp_output.c | 13 +++++++++++++ 1 file changed, 13 insertions(+) (limited to 'net/ipv4/tcp_output.c') diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 786978cb2db7..0f0984f39f67 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -38,6 +38,7 @@ #define pr_fmt(fmt) "TCP: " fmt #include +#include #include #include @@ -414,6 +415,7 @@ static inline bool tcp_urg_mode(const struct tcp_sock *tp) #define OPTION_WSCALE (1 << 3) #define OPTION_FAST_OPEN_COOKIE (1 << 8) #define OPTION_SMC (1 << 9) +#define OPTION_MPTCP (1 << 10) static void smc_options_write(__be32 *ptr, u16 *options) { @@ -439,8 +441,17 @@ struct tcp_out_options { __u8 *hash_location; /* temporary pointer, overloaded */ __u32 tsval, tsecr; /* need to include OPTION_TS */ struct tcp_fastopen_cookie *fastopen_cookie; /* Fast open cookie */ + struct mptcp_out_options mptcp; }; +static void mptcp_options_write(__be32 *ptr, struct tcp_out_options *opts) +{ +#if IS_ENABLED(CONFIG_MPTCP) + if (unlikely(OPTION_MPTCP & opts->options)) + mptcp_write_options(ptr, &opts->mptcp); +#endif +} + /* Write previously computed TCP options to the packet. * * Beware: Something in the Internet is very sensitive to the ordering of @@ -549,6 +560,8 @@ static void tcp_options_write(__be32 *ptr, struct tcp_sock *tp, } smc_options_write(ptr, &options); + + mptcp_options_write(ptr, opts); } static void smc_set_option(const struct tcp_sock *tp, -- cgit From cec37a6e41aae7bf3df9a3da783380a4d9325fd8 Mon Sep 17 00:00:00 2001 From: Peter Krystad Date: Tue, 21 Jan 2020 16:56:18 -0800 Subject: mptcp: Handle MP_CAPABLE options for outgoing connections Add hooks to tcp_output.c to add MP_CAPABLE to an outgoing SYN request, to capture the MP_CAPABLE in the received SYN-ACK, to add MP_CAPABLE to the final ACK of the three-way handshake. Use the .sk_rx_dst_set() handler in the subflow proto to capture when the responding SYN-ACK is received and notify the MPTCP connection layer. Co-developed-by: Paolo Abeni Signed-off-by: Paolo Abeni Co-developed-by: Florian Westphal Signed-off-by: Florian Westphal Signed-off-by: Peter Krystad Signed-off-by: Christoph Paasch Signed-off-by: David S. Miller --- net/ipv4/tcp_output.c | 44 ++++++++++++++++++++++++++++++++++++++++++++ 1 file changed, 44 insertions(+) (limited to 'net/ipv4/tcp_output.c') diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index 0f0984f39f67..5456076166da 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -597,6 +597,22 @@ static void smc_set_option_cond(const struct tcp_sock *tp, #endif } +static void mptcp_set_option_cond(const struct request_sock *req, + struct tcp_out_options *opts, + unsigned int *remaining) +{ + if (rsk_is_mptcp(req)) { + unsigned int size; + + if (mptcp_synack_options(req, &size, &opts->mptcp)) { + if (*remaining >= size) { + opts->options |= OPTION_MPTCP; + *remaining -= size; + } + } + } +} + /* Compute TCP options for SYN packets. This is not the final * network wire format yet. */ @@ -666,6 +682,15 @@ static unsigned int tcp_syn_options(struct sock *sk, struct sk_buff *skb, smc_set_option(tp, opts, &remaining); + if (sk_is_mptcp(sk)) { + unsigned int size; + + if (mptcp_syn_options(sk, &size, &opts->mptcp)) { + opts->options |= OPTION_MPTCP; + remaining -= size; + } + } + return MAX_TCP_OPTION_SPACE - remaining; } @@ -727,6 +752,8 @@ static unsigned int tcp_synack_options(const struct sock *sk, } } + mptcp_set_option_cond(req, opts, &remaining); + smc_set_option_cond(tcp_sk(sk), ireq, opts, &remaining); return MAX_TCP_OPTION_SPACE - remaining; @@ -764,6 +791,23 @@ static unsigned int tcp_established_options(struct sock *sk, struct sk_buff *skb size += TCPOLEN_TSTAMP_ALIGNED; } + /* MPTCP options have precedence over SACK for the limited TCP + * option space because a MPTCP connection would be forced to + * fall back to regular TCP if a required multipath option is + * missing. SACK still gets a chance to use whatever space is + * left. + */ + if (sk_is_mptcp(sk)) { + unsigned int remaining = MAX_TCP_OPTION_SPACE - size; + unsigned int opt_size = 0; + + if (mptcp_established_options(sk, skb, &opt_size, remaining, + &opts->mptcp)) { + opts->options |= OPTION_MPTCP; + size += opt_size; + } + } + eff_sacks = tp->rx_opt.num_sacks + tp->rx_opt.dsack; if (unlikely(eff_sacks)) { const unsigned int remaining = MAX_TCP_OPTION_SPACE - size; -- cgit From cc7972ea1932335e0a0ee00ac8a24b3e8304630d Mon Sep 17 00:00:00 2001 From: Christoph Paasch Date: Tue, 21 Jan 2020 16:56:31 -0800 Subject: mptcp: parse and emit MP_CAPABLE option according to v1 spec This implements MP_CAPABLE options parsing and writing according to RFC 6824 bis / RFC 8684: MPTCP v1. Local key is sent on syn/ack, and both keys are sent on 3rd ack. MP_CAPABLE messages len are updated accordingly. We need the skbuff to correctly emit the above, so we push the skbuff struct as an argument all the way from tcp code to the relevant mptcp callbacks. When processing incoming MP_CAPABLE + data, build a full blown DSS-like map info, to simplify later processing. On child socket creation, we need to record the remote key, if available. Signed-off-by: Christoph Paasch 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 5456076166da..fec4b3a4b22d 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -685,7 +685,7 @@ static unsigned int tcp_syn_options(struct sock *sk, struct sk_buff *skb, if (sk_is_mptcp(sk)) { unsigned int size; - if (mptcp_syn_options(sk, &size, &opts->mptcp)) { + if (mptcp_syn_options(sk, skb, &size, &opts->mptcp)) { opts->options |= OPTION_MPTCP; remaining -= size; } -- cgit