[linux-block.git] / include / net / udp.h

/* SPDX-License-Identifier: GPL-2.0-or-later */
/*
 * 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 UDP module.
 *
 * Version:	@(#)udp.h	1.0.2	05/07/93
 *
 * Authors:	Ross Biro
 *		Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
 *
 * Fixes:
 *		Alan Cox	: Turned on udp checksums. I don't want to
 *				  chase 'memory corruption' bugs that aren't!
 */
#ifndef _UDP_H
#define _UDP_H

#include <linux/list.h>
#include <linux/bug.h>
#include <net/inet_sock.h>
#include <net/sock.h>
#include <net/snmp.h>
#include <net/ip.h>
#include <linux/ipv6.h>
#include <linux/seq_file.h>
#include <linux/poll.h>
#include <linux/indirect_call_wrapper.h>

/**
 *	struct udp_skb_cb  -  UDP(-Lite) private variables
 *
 *	@header:      private variables used by IPv4/IPv6
 *	@cscov:       checksum coverage length (UDP-Lite only)
 *	@partial_cov: if set indicates partial csum coverage
 */
struct udp_skb_cb {
	union {
		struct inet_skb_parm	h4;
#if IS_ENABLED(CONFIG_IPV6)
		struct inet6_skb_parm	h6;
#endif
	} header;
	__u16		cscov;
	__u8		partial_cov;
};
#define UDP_SKB_CB(__skb)	((struct udp_skb_cb *)((__skb)->cb))

/**
 *	struct udp_hslot - UDP hash slot
 *
 *	@head:	head of list of sockets
 *	@count:	number of sockets in 'head' list
 *	@lock:	spinlock protecting changes to head/count
 */
struct udp_hslot {
	struct hlist_head	head;
	int			count;
	spinlock_t		lock;
} __attribute__((aligned(2 * sizeof(long))));

/**
 *	struct udp_table - UDP table
 *
 *	@hash:	hash table, sockets are hashed on (local port)
 *	@hash2:	hash table, sockets are hashed on (local port, local address)
 *	@mask:	number of slots in hash tables, minus 1
 *	@log:	log2(number of slots in hash table)
 */
struct udp_table {
	struct udp_hslot	*hash;
	struct udp_hslot	*hash2;
	unsigned int		mask;
	unsigned int		log;
};
extern struct udp_table udp_table;
void udp_table_init(struct udp_table *, const char *);
static inline struct udp_hslot *udp_hashslot(struct udp_table *table,
					     struct net *net, unsigned int num)
{
	return &table->hash[udp_hashfn(net, num, table->mask)];
}
/*
 * For secondary hash, net_hash_mix() is performed before calling
 * udp_hashslot2(), this explains difference with udp_hashslot()
 */
static inline struct udp_hslot *udp_hashslot2(struct udp_table *table,
					      unsigned int hash)
{
	return &table->hash2[hash & table->mask];
}

extern struct proto udp_prot;

extern atomic_long_t udp_memory_allocated;
DECLARE_PER_CPU(int, udp_memory_per_cpu_fw_alloc);

/* sysctl variables for udp */
extern long sysctl_udp_mem[3];
extern int sysctl_udp_rmem_min;
extern int sysctl_udp_wmem_min;

struct sk_buff;

/*
 *	Generic checksumming routines for UDP(-Lite) v4 and v6
 */
static inline __sum16 __udp_lib_checksum_complete(struct sk_buff *skb)
{
	return (UDP_SKB_CB(skb)->cscov == skb->len ?
		__skb_checksum_complete(skb) :
		__skb_checksum_complete_head(skb, UDP_SKB_CB(skb)->cscov));
}

static inline int udp_lib_checksum_complete(struct sk_buff *skb)
{
	return !skb_csum_unnecessary(skb) &&
		__udp_lib_checksum_complete(skb);
}

/**
 * 	udp_csum_outgoing  -  compute UDPv4/v6 checksum over fragments
 * 	@sk: 	socket we are writing to
 * 	@skb: 	sk_buff containing the filled-in UDP header
 * 	        (checksum field must be zeroed out)
 */
static inline __wsum udp_csum_outgoing(struct sock *sk, struct sk_buff *skb)
{
	__wsum csum = csum_partial(skb_transport_header(skb),
				   sizeof(struct udphdr), 0);
	skb_queue_walk(&sk->sk_write_queue, skb) {
		csum = csum_add(csum, skb->csum);
	}
	return csum;
}

static inline __wsum udp_csum(struct sk_buff *skb)
{
	__wsum csum = csum_partial(skb_transport_header(skb),
				   sizeof(struct udphdr), skb->csum);

	for (skb = skb_shinfo(skb)->frag_list; skb; skb = skb->next) {
		csum = csum_add(csum, skb->csum);
	}
	return csum;
}

static inline __sum16 udp_v4_check(int len, __be32 saddr,
				   __be32 daddr, __wsum base)
{
	return csum_tcpudp_magic(saddr, daddr, len, IPPROTO_UDP, base);
}

void udp_set_csum(bool nocheck, struct sk_buff *skb,
		  __be32 saddr, __be32 daddr, int len);

static inline void udp_csum_pull_header(struct sk_buff *skb)
{
	if (!skb->csum_valid && skb->ip_summed == CHECKSUM_NONE)
		skb->csum = csum_partial(skb->data, sizeof(struct udphdr),
					 skb->csum);
	skb_pull_rcsum(skb, sizeof(struct udphdr));
	UDP_SKB_CB(skb)->cscov -= sizeof(struct udphdr);
}

typedef struct sock *(*udp_lookup_t)(const struct sk_buff *skb, __be16 sport,
				     __be16 dport);

void udp_v6_early_demux(struct sk_buff *skb);
INDIRECT_CALLABLE_DECLARE(int udpv6_rcv(struct sk_buff *));

struct sk_buff *__udp_gso_segment(struct sk_buff *gso_skb,
				  netdev_features_t features, bool is_ipv6);

static inline void udp_lib_init_sock(struct sock *sk)
{
	struct udp_sock *up = udp_sk(sk);

	skb_queue_head_init(&up->reader_queue);
	up->forward_threshold = sk->sk_rcvbuf >> 2;
	set_bit(SOCK_CUSTOM_SOCKOPT, &sk->sk_socket->flags);
}

/* hash routines shared between UDPv4/6 and UDP-Litev4/6 */
static inline int udp_lib_hash(struct sock *sk)
{
	BUG();
	return 0;
}

void udp_lib_unhash(struct sock *sk);
void udp_lib_rehash(struct sock *sk, u16 new_hash);

static inline void udp_lib_close(struct sock *sk, long timeout)
{
	sk_common_release(sk);
}

int udp_lib_get_port(struct sock *sk, unsigned short snum,
		     unsigned int hash2_nulladdr);

u32 udp_flow_hashrnd(void);

static inline __be16 udp_flow_src_port(struct net *net, struct sk_buff *skb,
				       int min, int max, bool use_eth)
{
	u32 hash;

	if (min >= max) {
		/* Use default range */
		inet_get_local_port_range(net, &min, &max);
	}

	hash = skb_get_hash(skb);
	if (unlikely(!hash)) {
		if (use_eth) {
			/* Can't find a normal hash, caller has indicated an
			 * Ethernet packet so use that to compute a hash.
			 */
			hash = jhash(skb->data, 2 * ETH_ALEN,
				     (__force u32) skb->protocol);
		} else {
			/* Can't derive any sort of hash for the packet, set
			 * to some consistent random value.
			 */
			hash = udp_flow_hashrnd();
		}
	}

	/* Since this is being sent on the wire obfuscate hash a bit
	 * to minimize possbility that any useful information to an
	 * attacker is leaked. Only upper 16 bits are relevant in the
	 * computation for 16 bit port value.
	 */
	hash ^= hash << 16;

	return htons((((u64) hash * (max - min)) >> 32) + min);
}

static inline int udp_rqueue_get(struct sock *sk)
{
	return sk_rmem_alloc_get(sk) - READ_ONCE(udp_sk(sk)->forward_deficit);
}

static inline bool udp_sk_bound_dev_eq(struct net *net, int bound_dev_if,
				       int dif, int sdif)
{
#if IS_ENABLED(CONFIG_NET_L3_MASTER_DEV)
	return inet_bound_dev_eq(!!READ_ONCE(net->ipv4.sysctl_udp_l3mdev_accept),
				 bound_dev_if, dif, sdif);
#else
	return inet_bound_dev_eq(true, bound_dev_if, dif, sdif);
#endif
}

/* net/ipv4/udp.c */
void udp_destruct_common(struct sock *sk);
void skb_consume_udp(struct sock *sk, struct sk_buff *skb, int len);
int __udp_enqueue_schedule_skb(struct sock *sk, struct sk_buff *skb);
void udp_skb_destructor(struct sock *sk, struct sk_buff *skb);
struct sk_buff *__skb_recv_udp(struct sock *sk, unsigned int flags, int *off,
			       int *err);
static inline struct sk_buff *skb_recv_udp(struct sock *sk, unsigned int flags,
					   int *err)
{
	int off = 0;

	return __skb_recv_udp(sk, flags, &off, err);
}

int udp_v4_early_demux(struct sk_buff *skb);
bool udp_sk_rx_dst_set(struct sock *sk, struct dst_entry *dst);
int udp_get_port(struct sock *sk, unsigned short snum,
		 int (*saddr_cmp)(const struct sock *,
				  const struct sock *));
int udp_err(struct sk_buff *, u32);
int udp_abort(struct sock *sk, int err);
int udp_sendmsg(struct sock *sk, struct msghdr *msg, size_t len);
int udp_push_pending_frames(struct sock *sk);
void udp_flush_pending_frames(struct sock *sk);
int udp_cmsg_send(struct sock *sk, struct msghdr *msg, u16 *gso_size);
void udp4_hwcsum(struct sk_buff *skb, __be32 src, __be32 dst);
int udp_rcv(struct sk_buff *skb);
int udp_ioctl(struct sock *sk, int cmd, unsigned long arg);
int udp_init_sock(struct sock *sk);
int udp_pre_connect(struct sock *sk, struct sockaddr *uaddr, int addr_len);
int __udp_disconnect(struct sock *sk, int flags);
int udp_disconnect(struct sock *sk, int flags);
__poll_t udp_poll(struct file *file, struct socket *sock, poll_table *wait);
struct sk_buff *skb_udp_tunnel_segment(struct sk_buff *skb,
				       netdev_features_t features,
				       bool is_ipv6);
int udp_lib_getsockopt(struct sock *sk, int level, int optname,
		       char __user *optval, int __user *optlen);
int udp_lib_setsockopt(struct sock *sk, int level, int optname,
		       sockptr_t optval, unsigned int optlen,
		       int (*push_pending_frames)(struct sock *));
struct sock *udp4_lib_lookup(struct net *net, __be32 saddr, __be16 sport,
			     __be32 daddr, __be16 dport, int dif);
struct sock *__udp4_lib_lookup(struct net *net, __be32 saddr, __be16 sport,
			       __be32 daddr, __be16 dport, int dif, int sdif,
			       struct udp_table *tbl, struct sk_buff *skb);
struct sock *udp4_lib_lookup_skb(const struct sk_buff *skb,
				 __be16 sport, __be16 dport);
struct sock *udp6_lib_lookup(struct net *net,
			     const struct in6_addr *saddr, __be16 sport,
			     const struct in6_addr *daddr, __be16 dport,
			     int dif);
struct sock *__udp6_lib_lookup(struct net *net,
			       const struct in6_addr *saddr, __be16 sport,
			       const struct in6_addr *daddr, __be16 dport,
			       int dif, int sdif, struct udp_table *tbl,
			       struct sk_buff *skb);
struct sock *udp6_lib_lookup_skb(const struct sk_buff *skb,
				 __be16 sport, __be16 dport);
int udp_read_skb(struct sock *sk, skb_read_actor_t recv_actor);

/* UDP uses skb->dev_scratch to cache as much information as possible and avoid
 * possibly multiple cache miss on dequeue()
 */
struct udp_dev_scratch {
	/* skb->truesize and the stateless bit are embedded in a single field;
	 * do not use a bitfield since the compiler emits better/smaller code
	 * this way
	 */
	u32 _tsize_state;

#if BITS_PER_LONG == 64
	/* len and the bit needed to compute skb_csum_unnecessary
	 * will be on cold cache lines at recvmsg time.
	 * skb->len can be stored on 16 bits since the udp header has been
	 * already validated and pulled.
	 */
	u16 len;
	bool is_linear;
	bool csum_unnecessary;
#endif
};

static inline struct udp_dev_scratch *udp_skb_scratch(struct sk_buff *skb)
{
	return (struct udp_dev_scratch *)&skb->dev_scratch;
}

#if BITS_PER_LONG == 64
static inline unsigned int udp_skb_len(struct sk_buff *skb)
{
	return udp_skb_scratch(skb)->len;
}

static inline bool udp_skb_csum_unnecessary(struct sk_buff *skb)
{
	return udp_skb_scratch(skb)->csum_unnecessary;
}

static inline bool udp_skb_is_linear(struct sk_buff *skb)
{
	return udp_skb_scratch(skb)->is_linear;
}

#else
static inline unsigned int udp_skb_len(struct sk_buff *skb)
{
	return skb->len;
}

static inline bool udp_skb_csum_unnecessary(struct sk_buff *skb)
{
	return skb_csum_unnecessary(skb);
}

static inline bool udp_skb_is_linear(struct sk_buff *skb)
{
	return !skb_is_nonlinear(skb);
}
#endif

static inline int copy_linear_skb(struct sk_buff *skb, int len, int off,
				  struct iov_iter *to)
{
	int n;

	n = copy_to_iter(skb->data + off, len, to);
	if (n == len)
		return 0;

	iov_iter_revert(to, n);
	return -EFAULT;
}

/*
 * 	SNMP statistics for UDP and UDP-Lite
 */
#define UDP_INC_STATS(net, field, is_udplite)		      do { \
	if (is_udplite) SNMP_INC_STATS((net)->mib.udplite_statistics, field);       \
	else		SNMP_INC_STATS((net)->mib.udp_statistics, field);  }  while(0)
#define __UDP_INC_STATS(net, field, is_udplite) 	      do { \
	if (is_udplite) __SNMP_INC_STATS((net)->mib.udplite_statistics, field);         \
	else		__SNMP_INC_STATS((net)->mib.udp_statistics, field);    }  while(0)

#define __UDP6_INC_STATS(net, field, is_udplite)	    do { \
	if (is_udplite) __SNMP_INC_STATS((net)->mib.udplite_stats_in6, field);\
	else		__SNMP_INC_STATS((net)->mib.udp_stats_in6, field);  \
} while(0)
#define UDP6_INC_STATS(net, field, __lite)		    do { \
	if (__lite) SNMP_INC_STATS((net)->mib.udplite_stats_in6, field);  \
	else	    SNMP_INC_STATS((net)->mib.udp_stats_in6, field);      \
} while(0)

#if IS_ENABLED(CONFIG_IPV6)
#define __UDPX_MIB(sk, ipv4)						\
({									\
	ipv4 ? (IS_UDPLITE(sk) ? sock_net(sk)->mib.udplite_statistics :	\
				 sock_net(sk)->mib.udp_statistics) :	\
		(IS_UDPLITE(sk) ? sock_net(sk)->mib.udplite_stats_in6 :	\
				 sock_net(sk)->mib.udp_stats_in6);	\
})
#else
#define __UDPX_MIB(sk, ipv4)						\
({									\
	IS_UDPLITE(sk) ? sock_net(sk)->mib.udplite_statistics :		\
			 sock_net(sk)->mib.udp_statistics;		\
})
#endif

#define __UDPX_INC_STATS(sk, field) \
	__SNMP_INC_STATS(__UDPX_MIB(sk, (sk)->sk_family == AF_INET), field)

#ifdef CONFIG_PROC_FS
struct udp_seq_afinfo {
	sa_family_t			family;
	struct udp_table		*udp_table;
};

struct udp_iter_state {
	struct seq_net_private  p;
	int			bucket;
	struct udp_seq_afinfo	*bpf_seq_afinfo;
};

void *udp_seq_start(struct seq_file *seq, loff_t *pos);
void *udp_seq_next(struct seq_file *seq, void *v, loff_t *pos);
void udp_seq_stop(struct seq_file *seq, void *v);

extern const struct seq_operations udp_seq_ops;
extern const struct seq_operations udp6_seq_ops;

int udp4_proc_init(void);
void udp4_proc_exit(void);
#endif /* CONFIG_PROC_FS */

int udpv4_offload_init(void);

void udp_init(void);

DECLARE_STATIC_KEY_FALSE(udp_encap_needed_key);
void udp_encap_enable(void);
void udp_encap_disable(void);
#if IS_ENABLED(CONFIG_IPV6)
DECLARE_STATIC_KEY_FALSE(udpv6_encap_needed_key);
void udpv6_encap_enable(void);
#endif

static inline struct sk_buff *udp_rcv_segment(struct sock *sk,
					      struct sk_buff *skb, bool ipv4)
{
	netdev_features_t features = NETIF_F_SG;
	struct sk_buff *segs;

	/* Avoid csum recalculation by skb_segment unless userspace explicitly
	 * asks for the final checksum values
	 */
	if (!inet_get_convert_csum(sk))
		features |= NETIF_F_IP_CSUM | NETIF_F_IPV6_CSUM;

	/* UDP segmentation expects packets of type CHECKSUM_PARTIAL or
	 * CHECKSUM_NONE in __udp_gso_segment. UDP GRO indeed builds partial
	 * packets in udp_gro_complete_segment. As does UDP GSO, verified by
	 * udp_send_skb. But when those packets are looped in dev_loopback_xmit
	 * their ip_summed CHECKSUM_NONE is changed to CHECKSUM_UNNECESSARY.
	 * Reset in this specific case, where PARTIAL is both correct and
	 * required.
	 */
	if (skb->pkt_type == PACKET_LOOPBACK)
		skb->ip_summed = CHECKSUM_PARTIAL;

	/* the GSO CB lays after the UDP one, no need to save and restore any
	 * CB fragment
	 */
	segs = __skb_gso_segment(skb, features, false);
	if (IS_ERR_OR_NULL(segs)) {
		int segs_nr = skb_shinfo(skb)->gso_segs;

		atomic_add(segs_nr, &sk->sk_drops);
		SNMP_ADD_STATS(__UDPX_MIB(sk, ipv4), UDP_MIB_INERRORS, segs_nr);
		kfree_skb(skb);
		return NULL;
	}

	consume_skb(skb);
	return segs;
}

static inline void udp_post_segment_fix_csum(struct sk_buff *skb)
{
	/* UDP-lite can't land here - no GRO */
	WARN_ON_ONCE(UDP_SKB_CB(skb)->partial_cov);

	/* UDP packets generated with UDP_SEGMENT and traversing:
	 *
	 * UDP tunnel(xmit) -> veth (segmentation) -> veth (gro) -> UDP tunnel (rx)
	 *
	 * can reach an UDP socket with CHECKSUM_NONE, because
	 * __iptunnel_pull_header() converts CHECKSUM_PARTIAL into NONE.
	 * SKB_GSO_UDP_L4 or SKB_GSO_FRAGLIST packets with no UDP tunnel will
	 * have a valid checksum, as the GRO engine validates the UDP csum
	 * before the aggregation and nobody strips such info in between.
	 * Instead of adding another check in the tunnel fastpath, we can force
	 * a valid csum after the segmentation.
	 * Additionally fixup the UDP CB.
	 */
	UDP_SKB_CB(skb)->cscov = skb->len;
	if (skb->ip_summed == CHECKSUM_NONE && !skb->csum_valid)
		skb->csum_valid = 1;
}

#ifdef CONFIG_BPF_SYSCALL
struct sk_psock;
struct proto *udp_bpf_get_proto(struct sock *sk, struct sk_psock *psock);
int udp_bpf_update_proto(struct sock *sk, struct sk_psock *psock, bool restore);
#endif

#endif	/* _UDP_H */
Commit	Line	Data
2874c5fd	1	/* SPDX-License-Identifier: GPL-2.0-or-later */
1da177e4 LT	2	/*
	3	* INET An implementation of the TCP/IP protocol suite for the LINUX
	4	* operating system. INET is implemented using the BSD Socket
	5	* interface as the means of communication with the user level.
	6	*
	7	* Definitions for the UDP module.
	8	*
	9	* Version: @(#)udp.h 1.0.2 05/07/93
	10	*
02c30a84	11	* Authors: Ross Biro
1da177e4 LT	12	* Fred N. van Kempen, <waltje@uWalt.NL.Mugnet.ORG>
	13	*
	14	* Fixes:
	15	* Alan Cox : Turned on udp checksums. I don't want to
	16	* chase 'memory corruption' bugs that aren't!
1da177e4 LT	17	*/
	18	#ifndef _UDP_H
	19	#define _UDP_H
	20
1da177e4	21	#include <linux/list.h>
187f1882	22	#include <linux/bug.h>
14c85021	23	#include <net/inet_sock.h>
1da177e4 LT	24	#include <net/sock.h>
1da177e4 LT	25	#include <net/snmp.h>
ba4e58ec GR	26	#include <net/ip.h>
ba4e58ec GR	27	#include <linux/ipv6.h>
1da177e4	28	#include <linux/seq_file.h>
bd01f843	29	#include <linux/poll.h>
6db69328	30	#include <linux/indirect_call_wrapper.h>
1da177e4	31
ba4e58ec GR	32	/**
	33	* struct udp_skb_cb - UDP(-Lite) private variables
	34	*
	35	* @header: private variables used by IPv4/IPv6
	36	* @cscov: checksum coverage length (UDP-Lite only)
	37	* @partial_cov: if set indicates partial csum coverage
	38	*/
	39	struct udp_skb_cb {
	40	union {
	41	struct inet_skb_parm h4;
dfd56b8b	42	#if IS_ENABLED(CONFIG_IPV6)
ba4e58ec GR	43	struct inet6_skb_parm h6;
	44	#endif
	45	} header;
	46	__u16 cscov;
	47	__u8 partial_cov;
	48	};
	49	#define UDP_SKB_CB(__skb) ((struct udp_skb_cb *)((__skb)->cb))
1da177e4	50
fdcc8aa9 ED	51	/**
	52	* struct udp_hslot - UDP hash slot
	53	*
	54	* @head: head of list of sockets
	55	* @count: number of sockets in 'head' list
	56	* @lock: spinlock protecting changes to head/count
	57	*/
645ca708	58	struct udp_hslot {
ca065d0c	59	struct hlist_head head;
fdcc8aa9	60	int count;
645ca708 ED	61	spinlock_t lock;
645ca708 ED	62	} __attribute__((aligned(2 * sizeof(long))));
f86dcc5a	63
512615b6 ED	64	/**
	65	* struct udp_table - UDP table
	66	*
	67	* @hash: hash table, sockets are hashed on (local port)
	68	* @hash2: hash table, sockets are hashed on (local port, local address)
	69	* @mask: number of slots in hash tables, minus 1
	70	* @log: log2(number of slots in hash table)
	71	*/
645ca708	72	struct udp_table {
f86dcc5a	73	struct udp_hslot *hash;
512615b6 ED	74	struct udp_hslot *hash2;
	75	unsigned int mask;
	76	unsigned int log;
645ca708 ED	77	};
645ca708 ED	78	extern struct udp_table udp_table;
1a50bd06	79	void udp_table_init(struct udp_table , const char );
f86dcc5a	80	static inline struct udp_hslot udp_hashslot(struct udp_table table,
95c96174	81	struct net *net, unsigned int num)
f86dcc5a ED	82	{
	83	return &table->hash[udp_hashfn(net, num, table->mask)];
	84	}
512615b6 ED	85	/*
	86	* For secondary hash, net_hash_mix() is performed before calling
	87	* udp_hashslot2(), this explains difference with udp_hashslot()
	88	*/
	89	static inline struct udp_hslot udp_hashslot2(struct udp_table table,
	90	unsigned int hash)
	91	{
	92	return &table->hash2[hash & table->mask];
	93	}
1da177e4	94
1da177e4 LT	95	extern struct proto udp_prot;
1da177e4 LT	96
8d987e5c	97	extern atomic_long_t udp_memory_allocated;
0defbb0a	98	DECLARE_PER_CPU(int, udp_memory_per_cpu_fw_alloc);
95766fff HA	99
95766fff HA	100	/* sysctl variables for udp */
8d987e5c	101	extern long sysctl_udp_mem[3];
95766fff HA	102	extern int sysctl_udp_rmem_min;
	103	extern int sysctl_udp_wmem_min;
	104
14c85021	105	struct sk_buff;
1da177e4	106
ba4e58ec GR	107	/*
	108	* Generic checksumming routines for UDP(-Lite) v4 and v6
	109	*/
b51655b9	110	static inline __sum16 __udp_lib_checksum_complete(struct sk_buff *skb)
ba4e58ec	111	{
bbdff225 TH	112	return (UDP_SKB_CB(skb)->cscov == skb->len ?
	113	__skb_checksum_complete(skb) :
	114	__skb_checksum_complete_head(skb, UDP_SKB_CB(skb)->cscov));
ba4e58ec GR	115	}
ba4e58ec GR	116
8e5200f5	117	static inline int udp_lib_checksum_complete(struct sk_buff *skb)
ba4e58ec	118	{
60476372	119	return !skb_csum_unnecessary(skb) &&
ba4e58ec GR	120	__udp_lib_checksum_complete(skb);
	121	}
	122
	123	/**
	124	* udp_csum_outgoing - compute UDPv4/v6 checksum over fragments
	125	* @sk: socket we are writing to
	126	* @skb: sk_buff containing the filled-in UDP header
	127	* (checksum field must be zeroed out)
	128	*/
868c86bc	129	static inline __wsum udp_csum_outgoing(struct sock sk, struct sk_buff skb)
ba4e58ec	130	{
9c70220b ACM	131	__wsum csum = csum_partial(skb_transport_header(skb),
9c70220b ACM	132	sizeof(struct udphdr), 0);
ba4e58ec GR	133	skb_queue_walk(&sk->sk_write_queue, skb) {
	134	csum = csum_add(csum, skb->csum);
	135	}
	136	return csum;
	137	}
	138
f6b9664f HX	139	static inline __wsum udp_csum(struct sk_buff *skb)
	140	{
	141	__wsum csum = csum_partial(skb_transport_header(skb),
	142	sizeof(struct udphdr), skb->csum);
	143
	144	for (skb = skb_shinfo(skb)->frag_list; skb; skb = skb->next) {
	145	csum = csum_add(csum, skb->csum);
	146	}
	147	return csum;
	148	}
	149
af5fcba7 TH	150	static inline __sum16 udp_v4_check(int len, __be32 saddr,
	151	__be32 daddr, __wsum base)
	152	{
	153	return csum_tcpudp_magic(saddr, daddr, len, IPPROTO_UDP, base);
	154	}
	155
	156	void udp_set_csum(bool nocheck, struct sk_buff *skb,
	157	__be32 saddr, __be32 daddr, int len);
	158
e6afc8ac	159	static inline void udp_csum_pull_header(struct sk_buff *skb)
e6afc8ac	160	{
595d0b29 ED	161	if (!skb->csum_valid && skb->ip_summed == CHECKSUM_NONE)
595d0b29 ED	162	skb->csum = csum_partial(skb->data, sizeof(struct udphdr),
e6afc8ac	163	skb->csum);
	164	skb_pull_rcsum(skb, sizeof(struct udphdr));
	165	UDP_SKB_CB(skb)->cscov -= sizeof(struct udphdr);
	166	}
	167
7b58e63e	168	typedef struct sock (udp_lookup_t)(const struct sk_buff *skb, __be16 sport,
a6024562 TH	169	__be16 dport);
a6024562 TH	170
11052589	171	void udp_v6_early_demux(struct sk_buff *skb);
f9a4719c LR	172	INDIRECT_CALLABLE_DECLARE(int udpv6_rcv(struct sk_buff *));
f9a4719c LR	173
ee80d1eb	174	struct sk_buff __udp_gso_segment(struct sk_buff gso_skb,
c3df39ac	175	netdev_features_t features, bool is_ipv6);
ee80d1eb	176
8a3854c7 PA	177	static inline void udp_lib_init_sock(struct sock *sk)
	178	{
	179	struct udp_sock *up = udp_sk(sk);
	180
	181	skb_queue_head_init(&up->reader_queue);
	182	up->forward_threshold = sk->sk_rcvbuf >> 2;
	183	set_bit(SOCK_CUSTOM_SOCKOPT, &sk->sk_socket->flags);
	184	}
	185
ba4e58ec	186	/* hash routines shared between UDPv4/6 and UDP-Litev4/6 */
086c653f	187	static inline int udp_lib_hash(struct sock *sk)
ba4e58ec GR	188	{
ba4e58ec GR	189	BUG();
086c653f	190	return 0;
ba4e58ec GR	191	}
ba4e58ec GR	192
1a50bd06 JP	193	void udp_lib_unhash(struct sock *sk);
1a50bd06 JP	194	void udp_lib_rehash(struct sock *sk, u16 new_hash);
ba4e58ec GR	195
	196	static inline void udp_lib_close(struct sock *sk, long timeout)
	197	{
	198	sk_common_release(sk);
	199	}
	200
1a50bd06	201	int udp_lib_get_port(struct sock *sk, unsigned short snum,
1a50bd06	202	unsigned int hash2_nulladdr);
ba4e58ec	203
723b8e46 TH	204	u32 udp_flow_hashrnd(void);
723b8e46 TH	205
b8f1a556 TH	206	static inline __be16 udp_flow_src_port(struct net net, struct sk_buff skb,
	207	int min, int max, bool use_eth)
	208	{
	209	u32 hash;
	210
	211	if (min >= max) {
	212	/* Use default range */
	213	inet_get_local_port_range(net, &min, &max);
	214	}
	215
	216	hash = skb_get_hash(skb);
723b8e46 TH	217	if (unlikely(!hash)) {
	218	if (use_eth) {
	219	/* Can't find a normal hash, caller has indicated an
	220	* Ethernet packet so use that to compute a hash.
	221	*/
	222	hash = jhash(skb->data, 2 * ETH_ALEN,
	223	(__force u32) skb->protocol);
	224	} else {
	225	/* Can't derive any sort of hash for the packet, set
	226	* to some consistent random value.
	227	*/
	228	hash = udp_flow_hashrnd();
	229	}
b8f1a556 TH	230	}
	231
	232	/* Since this is being sent on the wire obfuscate hash a bit
	233	* to minimize possbility that any useful information to an
	234	* attacker is leaked. Only upper 16 bits are relevant in the
	235	* computation for 16 bit port value.
	236	*/
	237	hash ^= hash << 16;
	238
	239	return htons((((u64) hash * (max - min)) >> 32) + min);
	240	}
	241
6c206b20 PA	242	static inline int udp_rqueue_get(struct sock *sk)
	243	{
	244	return sk_rmem_alloc_get(sk) - READ_ONCE(udp_sk(sk)->forward_deficit);
	245	}
	246
6da5b0f0 MM	247	static inline bool udp_sk_bound_dev_eq(struct net *net, int bound_dev_if,
	248	int dif, int sdif)
	249	{
	250	#if IS_ENABLED(CONFIG_NET_L3_MASTER_DEV)
3d72bb41	251	return inet_bound_dev_eq(!!READ_ONCE(net->ipv4.sysctl_udp_l3mdev_accept),
6da5b0f0 MM	252	bound_dev_if, dif, sdif);
	253	#else
	254	return inet_bound_dev_eq(true, bound_dev_if, dif, sdif);
	255	#endif
	256	}
	257
ba4e58ec	258	/* net/ipv4/udp.c */
d38afeec	259	void udp_destruct_common(struct sock *sk);
f970bd9e PA	260	void skb_consume_udp(struct sock sk, struct sk_buff skb, int len);
f970bd9e PA	261	int __udp_enqueue_schedule_skb(struct sock sk, struct sk_buff skb);
7c13f97f	262	void udp_skb_destructor(struct sock sk, struct sk_buff skb);
ec095263 OH	263	struct sk_buff __skb_recv_udp(struct sock sk, unsigned int flags, int *off,
ec095263 OH	264	int *err);
7c13f97f	265	static inline struct sk_buff skb_recv_udp(struct sock sk, unsigned int flags,
ec095263	266	int *err)
7c13f97f	267	{
fd69c399	268	int off = 0;
7c13f97f	269
ec095263	270	return __skb_recv_udp(sk, flags, &off, err);
7c13f97f	271	}
f970bd9e	272
7487449c	273	int udp_v4_early_demux(struct sk_buff *skb);
64f0f5d1	274	bool udp_sk_rx_dst_set(struct sock sk, struct dst_entry dst);
1a50bd06 JP	275	int udp_get_port(struct sock *sk, unsigned short snum,
	276	int (saddr_cmp)(const struct sock ,
	277	const struct sock *));
32bbd879	278	int udp_err(struct sk_buff *, u32);
5d77dca8	279	int udp_abort(struct sock *sk, int err);
1b784140	280	int udp_sendmsg(struct sock sk, struct msghdr msg, size_t len);
1a50bd06 JP	281	int udp_push_pending_frames(struct sock *sk);
1a50bd06 JP	282	void udp_flush_pending_frames(struct sock *sk);
2e8de857	283	int udp_cmsg_send(struct sock sk, struct msghdr msg, u16 *gso_size);
1a50bd06 JP	284	void udp4_hwcsum(struct sk_buff *skb, __be32 src, __be32 dst);
	285	int udp_rcv(struct sk_buff *skb);
	286	int udp_ioctl(struct sock *sk, int cmd, unsigned long arg);
f970bd9e	287	int udp_init_sock(struct sock *sk);
d74bad4e	288	int udp_pre_connect(struct sock sk, struct sockaddr uaddr, int addr_len);
286c72de	289	int __udp_disconnect(struct sock *sk, int flags);
1a50bd06	290	int udp_disconnect(struct sock *sk, int flags);
a11e1d43	291	__poll_t udp_poll(struct file file, struct socket sock, poll_table *wait);
1a50bd06	292	struct sk_buff skb_udp_tunnel_segment(struct sk_buff skb,
8bce6d7d TH	293	netdev_features_t features,
8bce6d7d TH	294	bool is_ipv6);
1a50bd06 JP	295	int udp_lib_getsockopt(struct sock *sk, int level, int optname,
	296	char __user optval, int __user optlen);
	297	int udp_lib_setsockopt(struct sock *sk, int level, int optname,
91ac1cca	298	sockptr_t optval, unsigned int optlen,
1a50bd06 JP	299	int (push_pending_frames)(struct sock ));
	300	struct sock udp4_lib_lookup(struct net net, __be32 saddr, __be16 sport,
	301	__be32 daddr, __be16 dport, int dif);
	302	struct sock __udp4_lib_lookup(struct net net, __be32 saddr, __be16 sport,
fb74c277	303	__be32 daddr, __be16 dport, int dif, int sdif,
538950a1	304	struct udp_table tbl, struct sk_buff skb);
7b58e63e	305	struct sock udp4_lib_lookup_skb(const struct sk_buff skb,
63058308	306	__be16 sport, __be16 dport);
1a50bd06 JP	307	struct sock udp6_lib_lookup(struct net net,
	308	const struct in6_addr *saddr, __be16 sport,
	309	const struct in6_addr *daddr, __be16 dport,
	310	int dif);
	311	struct sock __udp6_lib_lookup(struct net net,
	312	const struct in6_addr *saddr, __be16 sport,
	313	const struct in6_addr *daddr, __be16 dport,
1801b570	314	int dif, int sdif, struct udp_table *tbl,
538950a1	315	struct sk_buff *skb);
7b58e63e	316	struct sock udp6_lib_lookup_skb(const struct sk_buff skb,
63058308	317	__be16 sport, __be16 dport);
965b57b4	318	int udp_read_skb(struct sock *sk, skb_read_actor_t recv_actor);
bcd41303	319
b26bbdae PA	320	/* UDP uses skb->dev_scratch to cache as much information as possible and avoid
	321	* possibly multiple cache miss on dequeue()
	322	*/
b26bbdae	323	struct udp_dev_scratch {
dce4551c PA	324	/* skb->truesize and the stateless bit are embedded in a single field;
	325	* do not use a bitfield since the compiler emits better/smaller code
	326	* this way
	327	*/
	328	u32 _tsize_state;
	329
	330	#if BITS_PER_LONG == 64
	331	/* len and the bit needed to compute skb_csum_unnecessary
	332	* will be on cold cache lines at recvmsg time.
	333	* skb->len can be stored on 16 bits since the udp header has been
	334	* already validated and pulled.
	335	*/
b26bbdae PA	336	u16 len;
	337	bool is_linear;
	338	bool csum_unnecessary;
dce4551c	339	#endif
b26bbdae PA	340	};
b26bbdae PA	341
dce4551c PA	342	static inline struct udp_dev_scratch udp_skb_scratch(struct sk_buff skb)
	343	{
	344	return (struct udp_dev_scratch *)&skb->dev_scratch;
	345	}
	346
	347	#if BITS_PER_LONG == 64
b26bbdae PA	348	static inline unsigned int udp_skb_len(struct sk_buff *skb)
b26bbdae PA	349	{
dce4551c	350	return udp_skb_scratch(skb)->len;
b26bbdae PA	351	}
	352
	353	static inline bool udp_skb_csum_unnecessary(struct sk_buff *skb)
	354	{
dce4551c	355	return udp_skb_scratch(skb)->csum_unnecessary;
b26bbdae PA	356	}
	357
	358	static inline bool udp_skb_is_linear(struct sk_buff *skb)
	359	{
dce4551c	360	return udp_skb_scratch(skb)->is_linear;
b26bbdae PA	361	}
	362
	363	#else
	364	static inline unsigned int udp_skb_len(struct sk_buff *skb)
	365	{
	366	return skb->len;
	367	}
	368
	369	static inline bool udp_skb_csum_unnecessary(struct sk_buff *skb)
	370	{
	371	return skb_csum_unnecessary(skb);
	372	}
	373
	374	static inline bool udp_skb_is_linear(struct sk_buff *skb)
	375	{
	376	return !skb_is_nonlinear(skb);
	377	}
	378	#endif
	379
	380	static inline int copy_linear_skb(struct sk_buff *skb, int len, int off,
	381	struct iov_iter *to)
	382	{
42b73059	383	int n;
b26bbdae	384
42b73059 AV	385	n = copy_to_iter(skb->data + off, len, to);
42b73059 AV	386	if (n == len)
b26bbdae PA	387	return 0;
b26bbdae PA	388
42b73059	389	iov_iter_revert(to, n);
b26bbdae PA	390	return -EFAULT;
	391	}
	392
ba4e58ec GR	393	/*
	394	* SNMP statistics for UDP and UDP-Lite
	395	*/
6aef70a8 ED	396	#define UDP_INC_STATS(net, field, is_udplite) do { \
	397	if (is_udplite) SNMP_INC_STATS((net)->mib.udplite_statistics, field); \
	398	else SNMP_INC_STATS((net)->mib.udp_statistics, field); } while(0)
02c22347	399	#define __UDP_INC_STATS(net, field, is_udplite) do { \
13415e46 ED	400	if (is_udplite) __SNMP_INC_STATS((net)->mib.udplite_statistics, field); \
13415e46 ED	401	else __SNMP_INC_STATS((net)->mib.udp_statistics, field); } while(0)
1da177e4	402
02c22347	403	#define __UDP6_INC_STATS(net, field, is_udplite) do { \
13415e46 ED	404	if (is_udplite) __SNMP_INC_STATS((net)->mib.udplite_stats_in6, field);\
13415e46 ED	405	else __SNMP_INC_STATS((net)->mib.udp_stats_in6, field); \
0c7ed677	406	} while(0)
6aef70a8 ED	407	#define UDP6_INC_STATS(net, field, __lite) do { \
	408	if (__lite) SNMP_INC_STATS((net)->mib.udplite_stats_in6, field); \
	409	else SNMP_INC_STATS((net)->mib.udp_stats_in6, field); \
0c7ed677	410	} while(0)
1781f7f5	411
dfd56b8b	412	#if IS_ENABLED(CONFIG_IPV6)
cf329aa4 PA	413	#define __UDPX_MIB(sk, ipv4) \
	414	({ \
	415	ipv4 ? (IS_UDPLITE(sk) ? sock_net(sk)->mib.udplite_statistics : \
	416	sock_net(sk)->mib.udp_statistics) : \
	417	(IS_UDPLITE(sk) ? sock_net(sk)->mib.udplite_stats_in6 : \
	418	sock_net(sk)->mib.udp_stats_in6); \
	419	})
1781f7f5	420	#else
cf329aa4 PA	421	#define __UDPX_MIB(sk, ipv4) \
	422	({ \
	423	IS_UDPLITE(sk) ? sock_net(sk)->mib.udplite_statistics : \
	424	sock_net(sk)->mib.udp_statistics; \
	425	})
1781f7f5 HX	426	#endif
1781f7f5 HX	427
cf329aa4 PA	428	#define __UDPX_INC_STATS(sk, field) \
	429	__SNMP_INC_STATS(__UDPX_MIB(sk, (sk)->sk_family == AF_INET), field)
	430
a3d2599b	431	#ifdef CONFIG_PROC_FS
1da177e4	432	struct udp_seq_afinfo {
73cb88ec AV	433	sa_family_t family;
73cb88ec AV	434	struct udp_table *udp_table;
1da177e4 LT	435	};
	436
	437	struct udp_iter_state {
6f191efe	438	struct seq_net_private p;
1da177e4	439	int bucket;
9e8ca27a	440	struct udp_seq_afinfo *bpf_seq_afinfo;
1da177e4 LT	441	};
1da177e4 LT	442
a3d2599b CH	443	void udp_seq_start(struct seq_file seq, loff_t *pos);
	444	void udp_seq_next(struct seq_file seq, void v, loff_t pos);
	445	void udp_seq_stop(struct seq_file seq, void v);
	446
c3506372 CH	447	extern const struct seq_operations udp_seq_ops;
c3506372 CH	448	extern const struct seq_operations udp6_seq_ops;
20380731	449
1a50bd06 JP	450	int udp4_proc_init(void);
1a50bd06 JP	451	void udp4_proc_exit(void);
a3d2599b	452	#endif /* CONFIG_PROC_FS */
95766fff	453
1a50bd06	454	int udpv4_offload_init(void);
da5bab07	455
1a50bd06	456	void udp_init(void);
d7ca4cc0	457
f2e9de21	458	DECLARE_STATIC_KEY_FALSE(udp_encap_needed_key);
1a50bd06	459	void udp_encap_enable(void);
a4a600dd	460	void udp_encap_disable(void);
d7f3f621	461	#if IS_ENABLED(CONFIG_IPV6)
f2e9de21	462	DECLARE_STATIC_KEY_FALSE(udpv6_encap_needed_key);
1a50bd06	463	void udpv6_encap_enable(void);
d7f3f621	464	#endif
dddb64bc	465
cf329aa4 PA	466	static inline struct sk_buff udp_rcv_segment(struct sock sk,
	467	struct sk_buff *skb, bool ipv4)
	468	{
f2696099	469	netdev_features_t features = NETIF_F_SG;
cf329aa4 PA	470	struct sk_buff *segs;
cf329aa4 PA	471
f2696099 ST	472	/* Avoid csum recalculation by skb_segment unless userspace explicitly
	473	* asks for the final checksum values
	474	*/
	475	if (!inet_get_convert_csum(sk))
	476	features \|= NETIF_F_IP_CSUM \| NETIF_F_IPV6_CSUM;
	477
d0208bf4 WB	478	/* UDP segmentation expects packets of type CHECKSUM_PARTIAL or
	479	* CHECKSUM_NONE in __udp_gso_segment. UDP GRO indeed builds partial
	480	* packets in udp_gro_complete_segment. As does UDP GSO, verified by
	481	* udp_send_skb. But when those packets are looped in dev_loopback_xmit
9122a70a CS	482	* their ip_summed CHECKSUM_NONE is changed to CHECKSUM_UNNECESSARY.
	483	* Reset in this specific case, where PARTIAL is both correct and
	484	* required.
d0208bf4	485	*/
6cd021a5 WB	486	if (skb->pkt_type == PACKET_LOOPBACK)
	487	skb->ip_summed = CHECKSUM_PARTIAL;
	488
cf329aa4 PA	489	/* the GSO CB lays after the UDP one, no need to save and restore any
	490	* CB fragment
	491	*/
f2696099	492	segs = __skb_gso_segment(skb, features, false);
974ceb21	493	if (IS_ERR_OR_NULL(segs)) {
cf329aa4 PA	494	int segs_nr = skb_shinfo(skb)->gso_segs;
	495
	496	atomic_add(segs_nr, &sk->sk_drops);
	497	SNMP_ADD_STATS(__UDPX_MIB(sk, ipv4), UDP_MIB_INERRORS, segs_nr);
	498	kfree_skb(skb);
	499	return NULL;
	500	}
	501
	502	consume_skb(skb);
	503	return segs;
	504	}
	505
000ac44d PA	506	static inline void udp_post_segment_fix_csum(struct sk_buff *skb)
	507	{
	508	/* UDP-lite can't land here - no GRO */
	509	WARN_ON_ONCE(UDP_SKB_CB(skb)->partial_cov);
	510
	511	/* UDP packets generated with UDP_SEGMENT and traversing:
	512	*
	513	* UDP tunnel(xmit) -> veth (segmentation) -> veth (gro) -> UDP tunnel (rx)
	514	*
	515	* can reach an UDP socket with CHECKSUM_NONE, because
	516	* __iptunnel_pull_header() converts CHECKSUM_PARTIAL into NONE.
	517	* SKB_GSO_UDP_L4 or SKB_GSO_FRAGLIST packets with no UDP tunnel will
	518	* have a valid checksum, as the GRO engine validates the UDP csum
	519	* before the aggregation and nobody strips such info in between.
	520	* Instead of adding another check in the tunnel fastpath, we can force
	521	* a valid csum after the segmentation.
	522	* Additionally fixup the UDP CB.
	523	*/
	524	UDP_SKB_CB(skb)->cscov = skb->len;
	525	if (skb->ip_summed == CHECKSUM_NONE && !skb->csum_valid)
	526	skb->csum_valid = 1;
	527	}
	528
88759609	529	#ifdef CONFIG_BPF_SYSCALL
edc6741c LB	530	struct sk_psock;
edc6741c LB	531	struct proto udp_bpf_get_proto(struct sock sk, struct sk_psock *psock);
51e0158a	532	int udp_bpf_update_proto(struct sock sk, struct sk_psock psock, bool restore);
88759609	533	#endif
edc6741c	534
1da177e4	535	#endif /* _UDP_H */