bpf: Remove extra lock_sock for TCP_ZEROCOPY_RECEIVE

Add custom implementation of getsockopt hook for TCP_ZEROCOPY_RECEIVE.
We skip generic hooks for TCP_ZEROCOPY_RECEIVE and have a custom
call in do_tcp_getsockopt using the on-stack data. This removes
3% overhead for locking/unlocking the socket.

Without this patch:
3.38% 0.07% tcp_mmap [kernel.kallsyms] [k] __cgroup_bpf_run_filter_getsockopt
|
--3.30%--__cgroup_bpf_run_filter_getsockopt
|
--0.81%--__kmalloc

With the patch applied:
0.52% 0.12% tcp_mmap [kernel.kallsyms] [k] __cgroup_bpf_run_filter_getsockopt_kern

Note, exporting uapi/tcp.h requires removing netinet/tcp.h
from test_progs.h because those headers have confliciting
definitions.

Signed-off-by: Stanislav Fomichev <[e-mail]>
Signed-off-by: Alexei Starovoitov <[e-mail]>
Acked-by: Martin KaFai Lau <[e-mail]>
Link: https://lore.kernel.org/bpf/[e-mail]

include/linux/bpf-cgroup.h

···
 				       int __user *optlen, int max_optlen,
 				       int retval);
 
+int __cgroup_bpf_run_filter_getsockopt_kern(struct sock *sk, int level,
+					    int optname, void *optval,
+					    int *optlen, int retval);
+
 static inline enum bpf_cgroup_storage_type cgroup_storage_type(
 	struct bpf_map *map)
 {
···
 ({									       \
 	int __ret = retval;						       \
 	if (cgroup_bpf_enabled)						       \
-		__ret = __cgroup_bpf_run_filter_getsockopt(sock, level,	       \
-							   optname, optval,    \
-							   optlen, max_optlen, \
-							   retval);	       \
+		if (!(sock)->sk_prot->bpf_bypass_getsockopt ||		       \
+		    !INDIRECT_CALL_INET_1((sock)->sk_prot->bpf_bypass_getsockopt, \
+					tcp_bpf_bypass_getsockopt,	       \
+					level, optname))		       \
+			__ret = __cgroup_bpf_run_filter_getsockopt(	       \
+				sock, level, optname, optval, optlen,	       \
+				max_optlen, retval);			       \
+	__ret;								       \
+})
+
+#define BPF_CGROUP_RUN_PROG_GETSOCKOPT_KERN(sock, level, optname, optval,      \
+					    optlen, retval)		       \
+({									       \
+	int __ret = retval;						       \
+	if (cgroup_bpf_enabled)						       \
+		__ret = __cgroup_bpf_run_filter_getsockopt_kern(	       \
+			sock, level, optname, optval, optlen, retval);	       \
 	__ret;								       \
 })
 
···
 #define BPF_CGROUP_GETSOCKOPT_MAX_OPTLEN(optlen) ({ 0; })
 #define BPF_CGROUP_RUN_PROG_GETSOCKOPT(sock, level, optname, optval, \
 				       optlen, max_optlen, retval) ({ retval; })
+#define BPF_CGROUP_RUN_PROG_GETSOCKOPT_KERN(sock, level, optname, optval, \
+					    optlen, retval) ({ retval; })
 #define BPF_CGROUP_RUN_PROG_SETSOCKOPT(sock, level, optname, optval, optlen, \
 				       kernel_optval) ({ 0; })
 

include/linux/indirect_call_wrapper.h

···
 #define INDIRECT_CALL_INET(f, f2, f1, ...) f(__VA_ARGS__)
 #endif
 
+#if IS_ENABLED(CONFIG_INET)
+#define INDIRECT_CALL_INET_1(f, f1, ...) INDIRECT_CALL_1(f, f1, __VA_ARGS__)
+#else
+#define INDIRECT_CALL_INET_1(f, f1, ...) f(__VA_ARGS__)
+#endif
+
 #endif

include/net/sock.h

···
 
 	int			(*backlog_rcv) (struct sock *sk,
 						struct sk_buff *skb);
+	bool			(*bpf_bypass_getsockopt)(int level,
+							 int optname);
 
 	void		(*release_cb)(struct sock *sk);
 

include/net/tcp.h

···
 		      struct poll_table_struct *wait);
 int tcp_getsockopt(struct sock *sk, int level, int optname,
 		   char __user *optval, int __user *optlen);
+bool tcp_bpf_bypass_getsockopt(int level, int optname);
 int tcp_setsockopt(struct sock *sk, int level, int optname, sockptr_t optval,
 		   unsigned int optlen);
 void tcp_set_keepalive(struct sock *sk, int val);

kernel/bpf/cgroup.c

···
 	sockopt_free_buf(&ctx);
 	return ret;
 }
+
+int __cgroup_bpf_run_filter_getsockopt_kern(struct sock *sk, int level,
+					    int optname, void *optval,
+					    int *optlen, int retval)
+{
+	struct cgroup *cgrp = sock_cgroup_ptr(&sk->sk_cgrp_data);
+	struct bpf_sockopt_kern ctx = {
+		.sk = sk,
+		.level = level,
+		.optname = optname,
+		.retval = retval,
+		.optlen = *optlen,
+		.optval = optval,
+		.optval_end = optval + *optlen,
+	};
+	int ret;
+
+	/* Note that __cgroup_bpf_run_filter_getsockopt doesn't copy
+	 * user data back into BPF buffer when reval != 0. This is
+	 * done as an optimization to avoid extra copy, assuming
+	 * kernel won't populate the data in case of an error.
+	 * Here we always pass the data and memset() should
+	 * be called if that data shouldn't be "exported".
+	 */
+
+	ret = BPF_PROG_RUN_ARRAY(cgrp->bpf.effective[BPF_CGROUP_GETSOCKOPT],
+				 &ctx, BPF_PROG_RUN);
+	if (!ret)
+		return -EPERM;
+
+	if (ctx.optlen > *optlen)
+		return -EFAULT;
+
+	/* BPF programs only allowed to set retval to 0, not some
+	 * arbitrary value.
+	 */
+	if (ctx.retval != 0 && ctx.retval != retval)
+		return -EFAULT;
+
+	/* BPF programs can shrink the buffer, export the modifications.
+	 */
+	if (ctx.optlen != 0)
+		*optlen = ctx.optlen;
+
+	return ctx.retval;
+}
 #endif
 
 static ssize_t sysctl_cpy_dir(const struct ctl_dir *dir, char **bufp,

net/ipv4/tcp.c

···
 			return -EFAULT;
 		lock_sock(sk);
 		err = tcp_zerocopy_receive(sk, &zc);
+		err = BPF_CGROUP_RUN_PROG_GETSOCKOPT_KERN(sk, level, optname,
+							  &zc, &len, err);
 		release_sock(sk);
 		if (len >= offsetofend(struct tcp_zerocopy_receive, err))
 			goto zerocopy_rcv_sk_err;
···
 		return -EFAULT;
 	return 0;
 }
+
+bool tcp_bpf_bypass_getsockopt(int level, int optname)
+{
+	/* TCP do_tcp_getsockopt has optimized getsockopt implementation
+	 * to avoid extra socket lock for TCP_ZEROCOPY_RECEIVE.
+	 */
+	if (level == SOL_TCP && optname == TCP_ZEROCOPY_RECEIVE)
+		return true;
+
+	return false;
+}
+EXPORT_SYMBOL(tcp_bpf_bypass_getsockopt);
 
 int tcp_getsockopt(struct sock *sk, int level, int optname, char __user *optval,
 		   int __user *optlen)

net/ipv4/tcp_ipv4.c

···
 	.shutdown		= tcp_shutdown,
 	.setsockopt		= tcp_setsockopt,
 	.getsockopt		= tcp_getsockopt,
+	.bpf_bypass_getsockopt	= tcp_bpf_bypass_getsockopt,
 	.keepalive		= tcp_set_keepalive,
 	.recvmsg		= tcp_recvmsg,
 	.sendmsg		= tcp_sendmsg,

net/ipv6/tcp_ipv6.c

···
 	.shutdown		= tcp_shutdown,
 	.setsockopt		= tcp_setsockopt,
 	.getsockopt		= tcp_getsockopt,
+	.bpf_bypass_getsockopt	= tcp_bpf_bypass_getsockopt,
 	.keepalive		= tcp_set_keepalive,
 	.recvmsg		= tcp_recvmsg,
 	.sendmsg		= tcp_sendmsg,

net/socket.c

···
 	return __sys_setsockopt(fd, level, optname, optval, optlen);
 }
 
+INDIRECT_CALLABLE_DECLARE(bool tcp_bpf_bypass_getsockopt(int level,
+							 int optname));
+
 /*
  *	Get a socket option. Because we don't know the option lengths we have
  *	to pass a user mode parameter for the protocols to sort out.

tools/include/uapi/linux/tcp.h

···
+/* SPDX-License-Identifier: GPL-2.0+ WITH Linux-syscall-note */
+/*
+ * INET		An implementation of the TCP/IP protocol suite for the LINUX
+ *		operating system.  INET is implemented using the  BSD Socket
+ *		interface as the means of communication with the user level.
+ *
+ *		Definitions for the TCP protocol.
+ *
+ * Version:	@(#)tcp.h	1.0.2	04/28/93
+ *
+ * Author:	Fred N. van Kempen, <[e-mail]>
+ *
+ *		This program is free software; you can redistribute it and/or
+ *		modify it under the terms of the GNU General Public License
+ *		as published by the Free Software Foundation; either version
+ *		2 of the License, or (at your option) any later version.
+ */
+#ifndef _UAPI_LINUX_TCP_H
+#define _UAPI_LINUX_TCP_H
+
+#include <linux/types.h>
+#include <asm/byteorder.h>
+#include <linux/socket.h>
+
+struct tcphdr {
+	__be16	source;
+	__be16	dest;
+	__be32	seq;
+	__be32	ack_seq;
+#if defined(__LITTLE_ENDIAN_BITFIELD)
+	__u16	res1:4,
+		doff:4,
+		fin:1,
+		syn:1,
+		rst:1,
+		psh:1,
+		ack:1,
+		urg:1,
+		ece:1,
+		cwr:1;
+#elif defined(__BIG_ENDIAN_BITFIELD)
+	__u16	doff:4,
+		res1:4,
+		cwr:1,
+		ece:1,
+		urg:1,
+		ack:1,
+		psh:1,
+		rst:1,
+		syn:1,
+		fin:1;
+#else
+#error	"Adjust your <asm/byteorder.h> defines"
+#endif	
+	__be16	window;
+	__sum16	check;
+	__be16	urg_ptr;
+};
+
+/*
+ *	The union cast uses a gcc extension to avoid aliasing problems
+ *  (union is compatible to any of its members)
+ *  This means this part of the code is -fstrict-aliasing safe now.
+ */
+union tcp_word_hdr { 
+	struct tcphdr hdr;
+	__be32 		  words[5];
+}; 
+
+#define tcp_flag_word(tp) ( ((union tcp_word_hdr *)(tp))->words [3]) 
+
+enum { 
+	TCP_FLAG_CWR = __constant_cpu_to_be32(0x00800000),
+	TCP_FLAG_ECE = __constant_cpu_to_be32(0x00400000),
+	TCP_FLAG_URG = __constant_cpu_to_be32(0x00200000),
+	TCP_FLAG_ACK = __constant_cpu_to_be32(0x00100000),
+	TCP_FLAG_PSH = __constant_cpu_to_be32(0x00080000),
+	TCP_FLAG_RST = __constant_cpu_to_be32(0x00040000),
+	TCP_FLAG_SYN = __constant_cpu_to_be32(0x00020000),
+	TCP_FLAG_FIN = __constant_cpu_to_be32(0x00010000),
+	TCP_RESERVED_BITS = __constant_cpu_to_be32(0x0F000000),
+	TCP_DATA_OFFSET = __constant_cpu_to_be32(0xF0000000)
+}; 
+
+/*
+ * TCP general constants
+ */
+#define TCP_MSS_DEFAULT		 536U	/* IPv4 (RFC1122, RFC2581) */
+#define TCP_MSS_DESIRED		1220U	/* IPv6 (tunneled), EDNS0 (RFC3226) */
+
+/* TCP socket options */
+#define TCP_NODELAY		1	/* Turn off Nagle's algorithm. */
+#define TCP_MAXSEG		2	/* Limit MSS */
+#define TCP_CORK		3	/* Never send partially complete segments */
+#define TCP_KEEPIDLE		4	/* Start keeplives after this period */
+#define TCP_KEEPINTVL		5	/* Interval between keepalives */
+#define TCP_KEEPCNT		6	/* Number of keepalives before death */
+#define TCP_SYNCNT		7	/* Number of SYN retransmits */
+#define TCP_LINGER2		8	/* Life time of orphaned FIN-WAIT-2 state */
+#define TCP_DEFER_ACCEPT	9	/* Wake up listener only when data arrive */
+#define TCP_WINDOW_CLAMP	10	/* Bound advertised window */
+#define TCP_INFO		11	/* Information about this connection. */
+#define TCP_QUICKACK		12	/* Block/reenable quick acks */
+#define TCP_CONGESTION		13	/* Congestion control algorithm */
+#define TCP_MD5SIG		14	/* TCP MD5 Signature (RFC2385) */
+#define TCP_THIN_LINEAR_TIMEOUTS 16      /* Use linear timeouts for thin streams*/
+#define TCP_THIN_DUPACK         17      /* Fast retrans. after 1 dupack */
+#define TCP_USER_TIMEOUT	18	/* How long for loss retry before timeout */
+#define TCP_REPAIR		19	/* TCP sock is under repair right now */
+#define TCP_REPAIR_QUEUE	20
+#define TCP_QUEUE_SEQ		21
+#define TCP_REPAIR_OPTIONS	22
+#define TCP_FASTOPEN		23	/* Enable FastOpen on listeners */
+#define TCP_TIMESTAMP		24
+#define TCP_NOTSENT_LOWAT	25	/* limit number of unsent bytes in write queue */
+#define TCP_CC_INFO		26	/* Get Congestion Control (optional) info */
+#define TCP_SAVE_SYN		27	/* Record SYN headers for new connections */
+#define TCP_SAVED_SYN		28	/* Get SYN headers recorded for connection */
+#define TCP_REPAIR_WINDOW	29	/* Get/set window parameters */
+#define TCP_FASTOPEN_CONNECT	30	/* Attempt FastOpen with connect */
+#define TCP_ULP			31	/* Attach a ULP to a TCP connection */
+#define TCP_MD5SIG_EXT		32	/* TCP MD5 Signature with extensions */
+#define TCP_FASTOPEN_KEY	33	/* Set the key for Fast Open (cookie) */
+#define TCP_FASTOPEN_NO_COOKIE	34	/* Enable TFO without a TFO cookie */
+#define TCP_ZEROCOPY_RECEIVE	35
+#define TCP_INQ			36	/* Notify bytes available to read as a cmsg on read */
+
+#define TCP_CM_INQ		TCP_INQ
+
+#define TCP_TX_DELAY		37	/* delay outgoing packets by XX usec */
+
+
+#define TCP_REPAIR_ON		1
+#define TCP_REPAIR_OFF		0
+#define TCP_REPAIR_OFF_NO_WP	-1	/* Turn off without window probes */
+
+struct tcp_repair_opt {
+	__u32	opt_code;
+	__u32	opt_val;
+};
+
+struct tcp_repair_window {
+	__u32	snd_wl1;
+	__u32	snd_wnd;
+	__u32	max_window;
+
+	__u32	rcv_wnd;
+	__u32	rcv_wup;
+};
+
+enum {
+	TCP_NO_QUEUE,
+	TCP_RECV_QUEUE,
+	TCP_SEND_QUEUE,
+	TCP_QUEUES_NR,
+};
+
+/* why fastopen failed from client perspective */
+enum tcp_fastopen_client_fail {
+	TFO_STATUS_UNSPEC, /* catch-all */
+	TFO_COOKIE_UNAVAILABLE, /* if not in TFO_CLIENT_NO_COOKIE mode */
+	TFO_DATA_NOT_ACKED, /* SYN-ACK did not ack SYN data */
+	TFO_SYN_RETRANSMITTED, /* SYN-ACK did not ack SYN data after timeout */
+};
+
+/* for TCP_INFO socket option */
+#define TCPI_OPT_TIMESTAMPS	1
+#define TCPI_OPT_SACK		2
+#define TCPI_OPT_WSCALE		4
+#define TCPI_OPT_ECN		8 /* ECN was negociated at TCP session init */
+#define TCPI_OPT_ECN_SEEN	16 /* we received at least one packet with ECT */
+#define TCPI_OPT_SYN_DATA	32 /* SYN-ACK acked data in SYN sent or rcvd */
+
+/*
+ * Sender's congestion state indicating normal or abnormal situations
+ * in the last round of packets sent. The state is driven by the ACK
+ * information and timer events.
+ */
+enum tcp_ca_state {
+	/*
+	 * Nothing bad has been observed recently.
+	 * No apparent reordering, packet loss, or ECN marks.
+	 */
+	TCP_CA_Open = 0,
+#define TCPF_CA_Open	(1<<TCP_CA_Open)
+	/*
+	 * The sender enters disordered state when it has received DUPACKs or
+	 * SACKs in the last round of packets sent. This could be due to packet
+	 * loss or reordering but needs further information to confirm packets
+	 * have been lost.
+	 */
+	TCP_CA_Disorder = 1,
+#define TCPF_CA_Disorder (1<<TCP_CA_Disorder)
+	/*
+	 * The sender enters Congestion Window Reduction (CWR) state when it
+	 * has received ACKs with ECN-ECE marks, or has experienced congestion
+	 * or packet discard on the sender host (e.g. qdisc).
+	 */
+	TCP_CA_CWR = 2,
+#define TCPF_CA_CWR	(1<<TCP_CA_CWR)
+	/*
+	 * The sender is in fast recovery and retransmitting lost packets,
+	 * typically triggered by ACK events.
+	 */
+	TCP_CA_Recovery = 3,
+#define TCPF_CA_Recovery (1<<TCP_CA_Recovery)
+	/*
+	 * The sender is in loss recovery triggered by retransmission timeout.
+	 */
+	TCP_CA_Loss = 4
+#define TCPF_CA_Loss	(1<<TCP_CA_Loss)
+};
+
+struct tcp_info {
+	__u8	tcpi_state;
+	__u8	tcpi_ca_state;
+	__u8	tcpi_retransmits;
+	__u8	tcpi_probes;
+	__u8	tcpi_backoff;
+	__u8	tcpi_options;
+	__u8	tcpi_snd_wscale : 4, tcpi_rcv_wscale : 4;
+	__u8	tcpi_delivery_rate_app_limited:1, tcpi_fastopen_client_fail:2;
+
+	__u32	tcpi_rto;
+	__u32	tcpi_ato;
+	__u32	tcpi_snd_mss;
+	__u32	tcpi_rcv_mss;
+
+	__u32	tcpi_unacked;
+	__u32	tcpi_sacked;
+	__u32	tcpi_lost;
+	__u32	tcpi_retrans;
+	__u32	tcpi_fackets;
+
+	/* Times. */
+	__u32	tcpi_last_data_sent;
+	__u32	tcpi_last_ack_sent;     /* Not remembered, sorry. */
+	__u32	tcpi_last_data_recv;
+	__u32	tcpi_last_ack_recv;
+
+	/* Metrics. */
+	__u32	tcpi_pmtu;
+	__u32	tcpi_rcv_ssthresh;
+	__u32	tcpi_rtt;
+	__u32	tcpi_rttvar;
+	__u32	tcpi_snd_ssthresh;
+	__u32	tcpi_snd_cwnd;
+	__u32	tcpi_advmss;
+	__u32	tcpi_reordering;
+
+	__u32	tcpi_rcv_rtt;
+	__u32	tcpi_rcv_space;
+
+	__u32	tcpi_total_retrans;
+
+	__u64	tcpi_pacing_rate;
+	__u64	tcpi_max_pacing_rate;
+	__u64	tcpi_bytes_acked;    /* RFC4898 tcpEStatsAppHCThruOctetsAcked */
+	__u64	tcpi_bytes_received; /* RFC4898 tcpEStatsAppHCThruOctetsReceived */
+	__u32	tcpi_segs_out;	     /* RFC4898 tcpEStatsPerfSegsOut */
+	__u32	tcpi_segs_in;	     /* RFC4898 tcpEStatsPerfSegsIn */
+
+	__u32	tcpi_notsent_bytes;
+	__u32	tcpi_min_rtt;
+	__u32	tcpi_data_segs_in;	/* RFC4898 tcpEStatsDataSegsIn */
+	__u32	tcpi_data_segs_out;	/* RFC4898 tcpEStatsDataSegsOut */
+
+	__u64   tcpi_delivery_rate;
+
+	__u64	tcpi_busy_time;      /* Time (usec) busy sending data */
+	__u64	tcpi_rwnd_limited;   /* Time (usec) limited by receive window */
+	__u64	tcpi_sndbuf_limited; /* Time (usec) limited by send buffer */
+
+	__u32	tcpi_delivered;
+	__u32	tcpi_delivered_ce;
+
+	__u64	tcpi_bytes_sent;     /* RFC4898 tcpEStatsPerfHCDataOctetsOut */
+	__u64	tcpi_bytes_retrans;  /* RFC4898 tcpEStatsPerfOctetsRetrans */
+	__u32	tcpi_dsack_dups;     /* RFC4898 tcpEStatsStackDSACKDups */
+	__u32	tcpi_reord_seen;     /* reordering events seen */
+
+	__u32	tcpi_rcv_ooopack;    /* Out-of-order packets received */
+
+	__u32	tcpi_snd_wnd;	     /* peer's advertised receive window after
+				      * scaling (bytes)
+				      */
+};
+
+/* netlink attributes types for SCM_TIMESTAMPING_OPT_STATS */
+enum {
+	TCP_NLA_PAD,
+	TCP_NLA_BUSY,		/* Time (usec) busy sending data */
+	TCP_NLA_RWND_LIMITED,	/* Time (usec) limited by receive window */
+	TCP_NLA_SNDBUF_LIMITED,	/* Time (usec) limited by send buffer */
+	TCP_NLA_DATA_SEGS_OUT,	/* Data pkts sent including retransmission */
+	TCP_NLA_TOTAL_RETRANS,	/* Data pkts retransmitted */
+	TCP_NLA_PACING_RATE,    /* Pacing rate in bytes per second */
+	TCP_NLA_DELIVERY_RATE,  /* Delivery rate in bytes per second */
+	TCP_NLA_SND_CWND,       /* Sending congestion window */
+	TCP_NLA_REORDERING,     /* Reordering metric */
+	TCP_NLA_MIN_RTT,        /* minimum RTT */
+	TCP_NLA_RECUR_RETRANS,  /* Recurring retransmits for the current pkt */
+	TCP_NLA_DELIVERY_RATE_APP_LMT, /* delivery rate application limited ? */
+	TCP_NLA_SNDQ_SIZE,	/* Data (bytes) pending in send queue */
+	TCP_NLA_CA_STATE,	/* ca_state of socket */
+	TCP_NLA_SND_SSTHRESH,	/* Slow start size threshold */
+	TCP_NLA_DELIVERED,	/* Data pkts delivered incl. out-of-order */
+	TCP_NLA_DELIVERED_CE,	/* Like above but only ones w/ CE marks */
+	TCP_NLA_BYTES_SENT,	/* Data bytes sent including retransmission */
+	TCP_NLA_BYTES_RETRANS,	/* Data bytes retransmitted */
+	TCP_NLA_DSACK_DUPS,	/* DSACK blocks received */
+	TCP_NLA_REORD_SEEN,	/* reordering events seen */
+	TCP_NLA_SRTT,		/* smoothed RTT in usecs */
+	TCP_NLA_TIMEOUT_REHASH, /* Timeout-triggered rehash attempts */
+	TCP_NLA_BYTES_NOTSENT,	/* Bytes in write queue not yet sent */
+	TCP_NLA_EDT,		/* Earliest departure time (CLOCK_MONOTONIC) */
+};
+
+/* for TCP_MD5SIG socket option */
+#define TCP_MD5SIG_MAXKEYLEN	80
+
+/* tcp_md5sig extension flags for TCP_MD5SIG_EXT */
+#define TCP_MD5SIG_FLAG_PREFIX		0x1	/* address prefix length */
+#define TCP_MD5SIG_FLAG_IFINDEX		0x2	/* ifindex set */
+
+struct tcp_md5sig {
+	struct __kernel_sockaddr_storage tcpm_addr;	/* address associated */
+	__u8	tcpm_flags;				/* extension flags */
+	__u8	tcpm_prefixlen;				/* address prefix */
+	__u16	tcpm_keylen;				/* key length */
+	int	tcpm_ifindex;				/* device index for scope */
+	__u8	tcpm_key[TCP_MD5SIG_MAXKEYLEN];		/* key (binary) */
+};
+
+/* INET_DIAG_MD5SIG */
+struct tcp_diag_md5sig {
+	__u8	tcpm_family;
+	__u8	tcpm_prefixlen;
+	__u16	tcpm_keylen;
+	__be32	tcpm_addr[4];
+	__u8	tcpm_key[TCP_MD5SIG_MAXKEYLEN];
+};
+
+/* setsockopt(fd, IPPROTO_TCP, TCP_ZEROCOPY_RECEIVE, ...) */
+
+#define TCP_RECEIVE_ZEROCOPY_FLAG_TLB_CLEAN_HINT 0x1
+struct tcp_zerocopy_receive {
+	__u64 address;		/* in: address of mapping */
+	__u32 length;		/* in/out: number of bytes to map/mapped */
+	__u32 recv_skip_hint;	/* out: amount of bytes to skip */
+	__u32 inq; /* out: amount of bytes in read queue */
+	__s32 err; /* out: socket error */
+	__u64 copybuf_address;	/* in: copybuf address (small reads) */
+	__s32 copybuf_len; /* in/out: copybuf bytes avail/used or error */
+	__u32 flags; /* in: flags */
+};
+#endif /* _UAPI_LINUX_TCP_H */

tools/testing/selftests/bpf/prog_tests/bpf_tcp_ca.c

···
 /* Copyright (c) 2019 Facebook */
 
 #include <linux/err.h>
+#include <netinet/tcp.h>
 #include <test_progs.h>
 #include "bpf_dctcp.skel.h"
 #include "bpf_cubic.skel.h"

tools/testing/selftests/bpf/prog_tests/cls_redirect.c

···
 #include <string.h>
 
 #include <linux/pkt_cls.h>
+#include <netinet/tcp.h>
 
 #include <test_progs.h>
 

tools/testing/selftests/bpf/prog_tests/sockmap_basic.c

···
 // SPDX-License-Identifier: GPL-2.0
 // Copyright (c) 2020 Cloudflare
 #include <error.h>
+#include <netinet/tcp.h>
 
 #include "test_progs.h"
 #include "test_skmsg_load_helpers.skel.h"

tools/testing/selftests/bpf/prog_tests/sockopt_sk.c

···
 #include <test_progs.h>
 #include "cgroup_helpers.h"
 
+#include <linux/tcp.h>
+
+#ifndef SOL_TCP
+#define SOL_TCP IPPROTO_TCP
+#endif
+
 #define SOL_CUSTOM			0xdeadbeef
 
 static int getsetsockopt(void)
···
 		char u8[4];
 		__u32 u32;
 		char cc[16]; /* TCP_CA_NAME_MAX */
+		struct tcp_zerocopy_receive zc;
 	} buf = {};
 	socklen_t optlen;
 	char *big_buf = NULL;
···
 	if (strcmp(buf.cc, "cubic") != 0) {
 		log_err("Unexpected getsockopt(TCP_CONGESTION) %s != %s",
 			buf.cc, "cubic");
+		goto err;
+	}
+
+	/* TCP_ZEROCOPY_RECEIVE triggers */
+	memset(&buf, 0, sizeof(buf));
+	optlen = sizeof(buf.zc);
+	err = getsockopt(fd, SOL_TCP, TCP_ZEROCOPY_RECEIVE, &buf, &optlen);
+	if (err) {
+		log_err("Unexpected getsockopt(TCP_ZEROCOPY_RECEIVE) err=%d errno=%d",
+			err, errno);
+		goto err;
+	}
+
+	memset(&buf, 0, sizeof(buf));
+	buf.zc.address = 12345; /* rejected by BPF */
+	optlen = sizeof(buf.zc);
+	errno = 0;
+	err = getsockopt(fd, SOL_TCP, TCP_ZEROCOPY_RECEIVE, &buf, &optlen);
+	if (errno != EPERM) {
+		log_err("Unexpected getsockopt(TCP_ZEROCOPY_RECEIVE) err=%d errno=%d",
+			err, errno);
 		goto err;
 	}
 

tools/testing/selftests/bpf/progs/sockopt_sk.c

···
 // SPDX-License-Identifier: GPL-2.0
 #include <string.h>
-#include <netinet/in.h>
-#include <netinet/tcp.h>
+#include <linux/tcp.h>
 #include <linux/bpf.h>
+#include <netinet/in.h>
 #include <bpf/bpf_helpers.h>
 
 char _license[] SEC("license") = "GPL";
···
 
 #ifndef PAGE_SIZE
 #define PAGE_SIZE 4096
+#endif
+
+#ifndef SOL_TCP
+#define SOL_TCP IPPROTO_TCP
 #endif
 
 #define SOL_CUSTOM			0xdeadbeef
···
 		 * let next BPF program in the cgroup chain or kernel
 		 * handle it.
 		 */
+		return 1;
+	}
+
+	if (ctx->level == SOL_TCP && ctx->optname == TCP_ZEROCOPY_RECEIVE) {
+		/* Verify that TCP_ZEROCOPY_RECEIVE triggers.
+		 * It has a custom implementation for performance
+		 * reasons.
+		 */
+
+		if (optval + sizeof(struct tcp_zerocopy_receive) > optval_end)
+			return 0; /* EPERM, bounds check */
+
+		if (((struct tcp_zerocopy_receive *)optval)->address != 0)
+			return 0; /* EPERM, unexpected data */
+
 		return 1;
 	}
 

tools/testing/selftests/bpf/test_progs.h

···
 #include <linux/if_packet.h>
 #include <linux/ip.h>
 #include <linux/ipv6.h>
-#include <netinet/tcp.h>
 #include <linux/filter.h>
 #include <linux/perf_event.h>
 #include <linux/socket.h>