[linux-block.git] / include / linux / skmsg.h

/* SPDX-License-Identifier: GPL-2.0 */
/* Copyright (c) 2017 - 2018 Covalent IO, Inc. http://covalent.io */

#ifndef _LINUX_SKMSG_H
#define _LINUX_SKMSG_H

#include <linux/bpf.h>
#include <linux/filter.h>
#include <linux/scatterlist.h>
#include <linux/skbuff.h>

#include <net/sock.h>
#include <net/tcp.h>
#include <net/strparser.h>

#define MAX_MSG_FRAGS			MAX_SKB_FRAGS
#define NR_MSG_FRAG_IDS			(MAX_MSG_FRAGS + 1)

enum __sk_action {
	__SK_DROP = 0,
	__SK_PASS,
	__SK_REDIRECT,
	__SK_NONE,
};

struct sk_msg_sg {
	u32				start;
	u32				curr;
	u32				end;
	u32				size;
	u32				copybreak;
	DECLARE_BITMAP(copy, MAX_MSG_FRAGS + 2);
	/* The extra two elements:
	 * 1) used for chaining the front and sections when the list becomes
	 *    partitioned (e.g. end < start). The crypto APIs require the
	 *    chaining;
	 * 2) to chain tailer SG entries after the message.
	 */
	struct scatterlist		data[MAX_MSG_FRAGS + 2];
};

/* UAPI in filter.c depends on struct sk_msg_sg being first element. */
struct sk_msg {
	struct sk_msg_sg		sg;
	void				*data;
	void				*data_end;
	u32				apply_bytes;
	u32				cork_bytes;
	u32				flags;
	struct sk_buff			*skb;
	struct sock			*sk_redir;
	struct sock			*sk;
	struct list_head		list;
};

struct sk_psock_progs {
	struct bpf_prog			*msg_parser;
	struct bpf_prog			*stream_parser;
	struct bpf_prog			*stream_verdict;
	struct bpf_prog			*skb_verdict;
};

enum sk_psock_state_bits {
	SK_PSOCK_TX_ENABLED,
};

struct sk_psock_link {
	struct list_head		list;
	struct bpf_map			*map;
	void				*link_raw;
};

struct sk_psock_work_state {
	struct sk_buff			*skb;
	u32				len;
	u32				off;
};

struct sk_psock {
	struct sock			*sk;
	struct sock			*sk_redir;
	u32				apply_bytes;
	u32				cork_bytes;
	u32				eval;
	struct sk_msg			*cork;
	struct sk_psock_progs		progs;
#if IS_ENABLED(CONFIG_BPF_STREAM_PARSER)
	struct strparser		strp;
#endif
	struct sk_buff_head		ingress_skb;
	struct list_head		ingress_msg;
	spinlock_t			ingress_lock;
	unsigned long			state;
	struct list_head		link;
	spinlock_t			link_lock;
	refcount_t			refcnt;
	void (*saved_unhash)(struct sock *sk);
	void (*saved_close)(struct sock *sk, long timeout);
	void (*saved_write_space)(struct sock *sk);
	void (*saved_data_ready)(struct sock *sk);
	int  (*psock_update_sk_prot)(struct sock *sk, struct sk_psock *psock,
				     bool restore);
	struct proto			*sk_proto;
	struct mutex			work_mutex;
	struct sk_psock_work_state	work_state;
	struct work_struct		work;
	struct rcu_work			rwork;
};

int sk_msg_alloc(struct sock *sk, struct sk_msg *msg, int len,
		 int elem_first_coalesce);
int sk_msg_clone(struct sock *sk, struct sk_msg *dst, struct sk_msg *src,
		 u32 off, u32 len);
void sk_msg_trim(struct sock *sk, struct sk_msg *msg, int len);
int sk_msg_free(struct sock *sk, struct sk_msg *msg);
int sk_msg_free_nocharge(struct sock *sk, struct sk_msg *msg);
void sk_msg_free_partial(struct sock *sk, struct sk_msg *msg, u32 bytes);
void sk_msg_free_partial_nocharge(struct sock *sk, struct sk_msg *msg,
				  u32 bytes);

void sk_msg_return(struct sock *sk, struct sk_msg *msg, int bytes);
void sk_msg_return_zero(struct sock *sk, struct sk_msg *msg, int bytes);

int sk_msg_zerocopy_from_iter(struct sock *sk, struct iov_iter *from,
			      struct sk_msg *msg, u32 bytes);
int sk_msg_memcopy_from_iter(struct sock *sk, struct iov_iter *from,
			     struct sk_msg *msg, u32 bytes);
int sk_msg_recvmsg(struct sock *sk, struct sk_psock *psock, struct msghdr *msg,
		   int len, int flags);
bool sk_msg_is_readable(struct sock *sk);

static inline void sk_msg_check_to_free(struct sk_msg *msg, u32 i, u32 bytes)
{
	WARN_ON(i == msg->sg.end && bytes);
}

static inline void sk_msg_apply_bytes(struct sk_psock *psock, u32 bytes)
{
	if (psock->apply_bytes) {
		if (psock->apply_bytes < bytes)
			psock->apply_bytes = 0;
		else
			psock->apply_bytes -= bytes;
	}
}

static inline u32 sk_msg_iter_dist(u32 start, u32 end)
{
	return end >= start ? end - start : end + (NR_MSG_FRAG_IDS - start);
}

#define sk_msg_iter_var_prev(var)			\
	do {						\
		if (var == 0)				\
			var = NR_MSG_FRAG_IDS - 1;	\
		else					\
			var--;				\
	} while (0)

#define sk_msg_iter_var_next(var)			\
	do {						\
		var++;					\
		if (var == NR_MSG_FRAG_IDS)		\
			var = 0;			\
	} while (0)

#define sk_msg_iter_prev(msg, which)			\
	sk_msg_iter_var_prev(msg->sg.which)

#define sk_msg_iter_next(msg, which)			\
	sk_msg_iter_var_next(msg->sg.which)

static inline void sk_msg_init(struct sk_msg *msg)
{
	BUILD_BUG_ON(ARRAY_SIZE(msg->sg.data) - 1 != NR_MSG_FRAG_IDS);
	memset(msg, 0, sizeof(*msg));
	sg_init_marker(msg->sg.data, NR_MSG_FRAG_IDS);
}

static inline void sk_msg_xfer(struct sk_msg *dst, struct sk_msg *src,
			       int which, u32 size)
{
	dst->sg.data[which] = src->sg.data[which];
	dst->sg.data[which].length  = size;
	dst->sg.size		   += size;
	src->sg.size		   -= size;
	src->sg.data[which].length -= size;
	src->sg.data[which].offset += size;
}

static inline void sk_msg_xfer_full(struct sk_msg *dst, struct sk_msg *src)
{
	memcpy(dst, src, sizeof(*src));
	sk_msg_init(src);
}

static inline bool sk_msg_full(const struct sk_msg *msg)
{
	return sk_msg_iter_dist(msg->sg.start, msg->sg.end) == MAX_MSG_FRAGS;
}

static inline u32 sk_msg_elem_used(const struct sk_msg *msg)
{
	return sk_msg_iter_dist(msg->sg.start, msg->sg.end);
}

static inline struct scatterlist *sk_msg_elem(struct sk_msg *msg, int which)
{
	return &msg->sg.data[which];
}

static inline struct scatterlist sk_msg_elem_cpy(struct sk_msg *msg, int which)
{
	return msg->sg.data[which];
}

static inline struct page *sk_msg_page(struct sk_msg *msg, int which)
{
	return sg_page(sk_msg_elem(msg, which));
}

static inline bool sk_msg_to_ingress(const struct sk_msg *msg)
{
	return msg->flags & BPF_F_INGRESS;
}

static inline void sk_msg_compute_data_pointers(struct sk_msg *msg)
{
	struct scatterlist *sge = sk_msg_elem(msg, msg->sg.start);

	if (test_bit(msg->sg.start, msg->sg.copy)) {
		msg->data = NULL;
		msg->data_end = NULL;
	} else {
		msg->data = sg_virt(sge);
		msg->data_end = msg->data + sge->length;
	}
}

static inline void sk_msg_page_add(struct sk_msg *msg, struct page *page,
				   u32 len, u32 offset)
{
	struct scatterlist *sge;

	get_page(page);
	sge = sk_msg_elem(msg, msg->sg.end);
	sg_set_page(sge, page, len, offset);
	sg_unmark_end(sge);

	__set_bit(msg->sg.end, msg->sg.copy);
	msg->sg.size += len;
	sk_msg_iter_next(msg, end);
}

static inline void sk_msg_sg_copy(struct sk_msg *msg, u32 i, bool copy_state)
{
	do {
		if (copy_state)
			__set_bit(i, msg->sg.copy);
		else
			__clear_bit(i, msg->sg.copy);
		sk_msg_iter_var_next(i);
		if (i == msg->sg.end)
			break;
	} while (1);
}

static inline void sk_msg_sg_copy_set(struct sk_msg *msg, u32 start)
{
	sk_msg_sg_copy(msg, start, true);
}

static inline void sk_msg_sg_copy_clear(struct sk_msg *msg, u32 start)
{
	sk_msg_sg_copy(msg, start, false);
}

static inline struct sk_psock *sk_psock(const struct sock *sk)
{
	return rcu_dereference_sk_user_data(sk);
}

static inline void sk_psock_set_state(struct sk_psock *psock,
				      enum sk_psock_state_bits bit)
{
	set_bit(bit, &psock->state);
}

static inline void sk_psock_clear_state(struct sk_psock *psock,
					enum sk_psock_state_bits bit)
{
	clear_bit(bit, &psock->state);
}

static inline bool sk_psock_test_state(const struct sk_psock *psock,
				       enum sk_psock_state_bits bit)
{
	return test_bit(bit, &psock->state);
}

static inline void sock_drop(struct sock *sk, struct sk_buff *skb)
{
	sk_drops_add(sk, skb);
	kfree_skb(skb);
}

static inline void sk_psock_queue_msg(struct sk_psock *psock,
				      struct sk_msg *msg)
{
	spin_lock_bh(&psock->ingress_lock);
	if (sk_psock_test_state(psock, SK_PSOCK_TX_ENABLED))
		list_add_tail(&msg->list, &psock->ingress_msg);
	else {
		sk_msg_free(psock->sk, msg);
		kfree(msg);
	}
	spin_unlock_bh(&psock->ingress_lock);
}

static inline struct sk_msg *sk_psock_dequeue_msg(struct sk_psock *psock)
{
	struct sk_msg *msg;

	spin_lock_bh(&psock->ingress_lock);
	msg = list_first_entry_or_null(&psock->ingress_msg, struct sk_msg, list);
	if (msg)
		list_del(&msg->list);
	spin_unlock_bh(&psock->ingress_lock);
	return msg;
}

static inline struct sk_msg *sk_psock_peek_msg(struct sk_psock *psock)
{
	struct sk_msg *msg;

	spin_lock_bh(&psock->ingress_lock);
	msg = list_first_entry_or_null(&psock->ingress_msg, struct sk_msg, list);
	spin_unlock_bh(&psock->ingress_lock);
	return msg;
}

static inline struct sk_msg *sk_psock_next_msg(struct sk_psock *psock,
					       struct sk_msg *msg)
{
	struct sk_msg *ret;

	spin_lock_bh(&psock->ingress_lock);
	if (list_is_last(&msg->list, &psock->ingress_msg))
		ret = NULL;
	else
		ret = list_next_entry(msg, list);
	spin_unlock_bh(&psock->ingress_lock);
	return ret;
}

static inline bool sk_psock_queue_empty(const struct sk_psock *psock)
{
	return psock ? list_empty(&psock->ingress_msg) : true;
}

static inline void kfree_sk_msg(struct sk_msg *msg)
{
	if (msg->skb)
		consume_skb(msg->skb);
	kfree(msg);
}

static inline void sk_psock_report_error(struct sk_psock *psock, int err)
{
	struct sock *sk = psock->sk;

	sk->sk_err = err;
	sk_error_report(sk);
}

struct sk_psock *sk_psock_init(struct sock *sk, int node);
void sk_psock_stop(struct sk_psock *psock, bool wait);

#if IS_ENABLED(CONFIG_BPF_STREAM_PARSER)
int sk_psock_init_strp(struct sock *sk, struct sk_psock *psock);
void sk_psock_start_strp(struct sock *sk, struct sk_psock *psock);
void sk_psock_stop_strp(struct sock *sk, struct sk_psock *psock);
#else
static inline int sk_psock_init_strp(struct sock *sk, struct sk_psock *psock)
{
	return -EOPNOTSUPP;
}

static inline void sk_psock_start_strp(struct sock *sk, struct sk_psock *psock)
{
}

static inline void sk_psock_stop_strp(struct sock *sk, struct sk_psock *psock)
{
}
#endif

void sk_psock_start_verdict(struct sock *sk, struct sk_psock *psock);
void sk_psock_stop_verdict(struct sock *sk, struct sk_psock *psock);

int sk_psock_msg_verdict(struct sock *sk, struct sk_psock *psock,
			 struct sk_msg *msg);

static inline struct sk_psock_link *sk_psock_init_link(void)
{
	return kzalloc(sizeof(struct sk_psock_link),
		       GFP_ATOMIC | __GFP_NOWARN);
}

static inline void sk_psock_free_link(struct sk_psock_link *link)
{
	kfree(link);
}

struct sk_psock_link *sk_psock_link_pop(struct sk_psock *psock);

static inline void sk_psock_cork_free(struct sk_psock *psock)
{
	if (psock->cork) {
		sk_msg_free(psock->sk, psock->cork);
		kfree(psock->cork);
		psock->cork = NULL;
	}
}

static inline void sk_psock_restore_proto(struct sock *sk,
					  struct sk_psock *psock)
{
	if (psock->psock_update_sk_prot)
		psock->psock_update_sk_prot(sk, psock, true);
}

static inline struct sk_psock *sk_psock_get(struct sock *sk)
{
	struct sk_psock *psock;

	rcu_read_lock();
	psock = sk_psock(sk);
	if (psock && !refcount_inc_not_zero(&psock->refcnt))
		psock = NULL;
	rcu_read_unlock();
	return psock;
}

void sk_psock_drop(struct sock *sk, struct sk_psock *psock);

static inline void sk_psock_put(struct sock *sk, struct sk_psock *psock)
{
	if (refcount_dec_and_test(&psock->refcnt))
		sk_psock_drop(sk, psock);
}

static inline void sk_psock_data_ready(struct sock *sk, struct sk_psock *psock)
{
	if (psock->saved_data_ready)
		psock->saved_data_ready(sk);
	else
		sk->sk_data_ready(sk);
}

static inline void psock_set_prog(struct bpf_prog **pprog,
				  struct bpf_prog *prog)
{
	prog = xchg(pprog, prog);
	if (prog)
		bpf_prog_put(prog);
}

static inline int psock_replace_prog(struct bpf_prog **pprog,
				     struct bpf_prog *prog,
				     struct bpf_prog *old)
{
	if (cmpxchg(pprog, old, prog) != old)
		return -ENOENT;

	if (old)
		bpf_prog_put(old);

	return 0;
}

static inline void psock_progs_drop(struct sk_psock_progs *progs)
{
	psock_set_prog(&progs->msg_parser, NULL);
	psock_set_prog(&progs->stream_parser, NULL);
	psock_set_prog(&progs->stream_verdict, NULL);
	psock_set_prog(&progs->skb_verdict, NULL);
}

int sk_psock_tls_strp_read(struct sk_psock *psock, struct sk_buff *skb);

static inline bool sk_psock_strp_enabled(struct sk_psock *psock)
{
	if (!psock)
		return false;
	return !!psock->saved_data_ready;
}

static inline bool sk_is_udp(const struct sock *sk)
{
	return sk->sk_type == SOCK_DGRAM &&
	       sk->sk_protocol == IPPROTO_UDP;
}

#if IS_ENABLED(CONFIG_NET_SOCK_MSG)

#define BPF_F_STRPARSER	(1UL << 1)

/* We only have two bits so far. */
#define BPF_F_PTR_MASK ~(BPF_F_INGRESS | BPF_F_STRPARSER)

static inline bool skb_bpf_strparser(const struct sk_buff *skb)
{
	unsigned long sk_redir = skb->_sk_redir;

	return sk_redir & BPF_F_STRPARSER;
}

static inline void skb_bpf_set_strparser(struct sk_buff *skb)
{
	skb->_sk_redir |= BPF_F_STRPARSER;
}

static inline bool skb_bpf_ingress(const struct sk_buff *skb)
{
	unsigned long sk_redir = skb->_sk_redir;

	return sk_redir & BPF_F_INGRESS;
}

static inline void skb_bpf_set_ingress(struct sk_buff *skb)
{
	skb->_sk_redir |= BPF_F_INGRESS;
}

static inline void skb_bpf_set_redir(struct sk_buff *skb, struct sock *sk_redir,
				     bool ingress)
{
	skb->_sk_redir = (unsigned long)sk_redir;
	if (ingress)
		skb->_sk_redir |= BPF_F_INGRESS;
}

static inline struct sock *skb_bpf_redirect_fetch(const struct sk_buff *skb)
{
	unsigned long sk_redir = skb->_sk_redir;

	return (struct sock *)(sk_redir & BPF_F_PTR_MASK);
}

static inline void skb_bpf_redirect_clear(struct sk_buff *skb)
{
	skb->_sk_redir = 0;
}
#endif /* CONFIG_NET_SOCK_MSG */
#endif /* _LINUX_SKMSG_H */
Commit	Line	Data
604326b4 DB	1	/* SPDX-License-Identifier: GPL-2.0 */
	2	/* Copyright (c) 2017 - 2018 Covalent IO, Inc. http://covalent.io */
	3
	4	#ifndef _LINUX_SKMSG_H
	5	#define _LINUX_SKMSG_H
	6
	7	#include <linux/bpf.h>
	8	#include <linux/filter.h>
	9	#include <linux/scatterlist.h>
	10	#include <linux/skbuff.h>
	11
	12	#include <net/sock.h>
	13	#include <net/tcp.h>
	14	#include <net/strparser.h>
	15
	16	#define MAX_MSG_FRAGS MAX_SKB_FRAGS
031097d9	17	#define NR_MSG_FRAG_IDS (MAX_MSG_FRAGS + 1)
604326b4 DB	18
	19	enum __sk_action {
	20	__SK_DROP = 0,
	21	__SK_PASS,
	22	__SK_REDIRECT,
	23	__SK_NONE,
	24	};
	25
	26	struct sk_msg_sg {
	27	u32 start;
	28	u32 curr;
	29	u32 end;
	30	u32 size;
	31	u32 copybreak;
5a8fb33e	32	DECLARE_BITMAP(copy, MAX_MSG_FRAGS + 2);
031097d9 JK	33	/* The extra two elements:
	34	* 1) used for chaining the front and sections when the list becomes
	35	* partitioned (e.g. end < start). The crypto APIs require the
	36	* chaining;
	37	* 2) to chain tailer SG entries after the message.
d3b18ad3	38	*/
031097d9	39	struct scatterlist data[MAX_MSG_FRAGS + 2];
604326b4 DB	40	};
604326b4 DB	41
7a69c0f2	42	/* UAPI in filter.c depends on struct sk_msg_sg being first element. */
604326b4 DB	43	struct sk_msg {
	44	struct sk_msg_sg sg;
	45	void *data;
	46	void *data_end;
	47	u32 apply_bytes;
	48	u32 cork_bytes;
	49	u32 flags;
	50	struct sk_buff *skb;
	51	struct sock *sk_redir;
	52	struct sock *sk;
	53	struct list_head list;
	54	};
	55
	56	struct sk_psock_progs {
	57	struct bpf_prog *msg_parser;
ae8b8332 CW	58	struct bpf_prog *stream_parser;
ae8b8332 CW	59	struct bpf_prog *stream_verdict;
a7ba4558	60	struct bpf_prog *skb_verdict;
604326b4 DB	61	};
	62
	63	enum sk_psock_state_bits {
	64	SK_PSOCK_TX_ENABLED,
	65	};
	66
	67	struct sk_psock_link {
	68	struct list_head list;
	69	struct bpf_map *map;
	70	void *link_raw;
	71	};
	72
604326b4 DB	73	struct sk_psock_work_state {
	74	struct sk_buff *skb;
	75	u32 len;
	76	u32 off;
	77	};
	78
	79	struct sk_psock {
	80	struct sock *sk;
	81	struct sock *sk_redir;
	82	u32 apply_bytes;
	83	u32 cork_bytes;
	84	u32 eval;
	85	struct sk_msg *cork;
	86	struct sk_psock_progs progs;
5a685cd9 CW	87	#if IS_ENABLED(CONFIG_BPF_STREAM_PARSER)
	88	struct strparser strp;
	89	#endif
604326b4 DB	90	struct sk_buff_head ingress_skb;
604326b4 DB	91	struct list_head ingress_msg;
b01fd6e8	92	spinlock_t ingress_lock;
604326b4 DB	93	unsigned long state;
	94	struct list_head link;
	95	spinlock_t link_lock;
	96	refcount_t refcnt;
	97	void (saved_unhash)(struct sock sk);
	98	void (saved_close)(struct sock sk, long timeout);
	99	void (saved_write_space)(struct sock sk);
5a685cd9	100	void (saved_data_ready)(struct sock sk);
51e0158a CW	101	int (psock_update_sk_prot)(struct sock sk, struct sk_psock *psock,
51e0158a CW	102	bool restore);
604326b4	103	struct proto *sk_proto;
799aa7f9	104	struct mutex work_mutex;
604326b4 DB	105	struct sk_psock_work_state work_state;
604326b4 DB	106	struct work_struct work;
7786dfc4	107	struct rcu_work rwork;
604326b4 DB	108	};
	109
	110	int sk_msg_alloc(struct sock sk, struct sk_msg msg, int len,
	111	int elem_first_coalesce);
d829e9c4 DB	112	int sk_msg_clone(struct sock sk, struct sk_msg dst, struct sk_msg *src,
d829e9c4 DB	113	u32 off, u32 len);
604326b4 DB	114	void sk_msg_trim(struct sock sk, struct sk_msg msg, int len);
	115	int sk_msg_free(struct sock sk, struct sk_msg msg);
	116	int sk_msg_free_nocharge(struct sock sk, struct sk_msg msg);
	117	void sk_msg_free_partial(struct sock sk, struct sk_msg msg, u32 bytes);
	118	void sk_msg_free_partial_nocharge(struct sock sk, struct sk_msg msg,
	119	u32 bytes);
	120
	121	void sk_msg_return(struct sock sk, struct sk_msg msg, int bytes);
d3b18ad3	122	void sk_msg_return_zero(struct sock sk, struct sk_msg msg, int bytes);
604326b4 DB	123
	124	int sk_msg_zerocopy_from_iter(struct sock sk, struct iov_iter from,
	125	struct sk_msg *msg, u32 bytes);
	126	int sk_msg_memcopy_from_iter(struct sock sk, struct iov_iter from,
	127	struct sk_msg *msg, u32 bytes);
2bc793e3 CW	128	int sk_msg_recvmsg(struct sock sk, struct sk_psock psock, struct msghdr *msg,
2bc793e3 CW	129	int len, int flags);
fb4e0a5e	130	bool sk_msg_is_readable(struct sock *sk);
604326b4 DB	131
	132	static inline void sk_msg_check_to_free(struct sk_msg *msg, u32 i, u32 bytes)
	133	{
	134	WARN_ON(i == msg->sg.end && bytes);
	135	}
	136
	137	static inline void sk_msg_apply_bytes(struct sk_psock *psock, u32 bytes)
	138	{
	139	if (psock->apply_bytes) {
	140	if (psock->apply_bytes < bytes)
	141	psock->apply_bytes = 0;
	142	else
	143	psock->apply_bytes -= bytes;
	144	}
	145	}
	146
683916f6 JK	147	static inline u32 sk_msg_iter_dist(u32 start, u32 end)
683916f6 JK	148	{
031097d9	149	return end >= start ? end - start : end + (NR_MSG_FRAG_IDS - start);
683916f6 JK	150	}
683916f6 JK	151
604326b4 DB	152	#define sk_msg_iter_var_prev(var) \
	153	do { \
	154	if (var == 0) \
031097d9	155	var = NR_MSG_FRAG_IDS - 1; \
604326b4 DB	156	else \
	157	var--; \
	158	} while (0)
	159
	160	#define sk_msg_iter_var_next(var) \
	161	do { \
	162	var++; \
031097d9	163	if (var == NR_MSG_FRAG_IDS) \
604326b4 DB	164	var = 0; \
	165	} while (0)
	166
	167	#define sk_msg_iter_prev(msg, which) \
	168	sk_msg_iter_var_prev(msg->sg.which)
	169
	170	#define sk_msg_iter_next(msg, which) \
	171	sk_msg_iter_var_next(msg->sg.which)
	172
604326b4 DB	173	static inline void sk_msg_init(struct sk_msg *msg)
604326b4 DB	174	{
031097d9	175	BUILD_BUG_ON(ARRAY_SIZE(msg->sg.data) - 1 != NR_MSG_FRAG_IDS);
604326b4	176	memset(msg, 0, sizeof(*msg));
031097d9	177	sg_init_marker(msg->sg.data, NR_MSG_FRAG_IDS);
604326b4 DB	178	}
	179
	180	static inline void sk_msg_xfer(struct sk_msg dst, struct sk_msg src,
	181	int which, u32 size)
	182	{
	183	dst->sg.data[which] = src->sg.data[which];
	184	dst->sg.data[which].length = size;
3f4c3127	185	dst->sg.size += size;
81aabbb9	186	src->sg.size -= size;
604326b4 DB	187	src->sg.data[which].length -= size;
	188	src->sg.data[which].offset += size;
	189	}
	190
d3b18ad3 JF	191	static inline void sk_msg_xfer_full(struct sk_msg dst, struct sk_msg src)
	192	{
	193	memcpy(dst, src, sizeof(*src));
	194	sk_msg_init(src);
	195	}
	196
8734a162 JF	197	static inline bool sk_msg_full(const struct sk_msg *msg)
8734a162 JF	198	{
031097d9	199	return sk_msg_iter_dist(msg->sg.start, msg->sg.end) == MAX_MSG_FRAGS;
8734a162 JF	200	}
8734a162 JF	201
604326b4 DB	202	static inline u32 sk_msg_elem_used(const struct sk_msg *msg)
604326b4 DB	203	{
683916f6	204	return sk_msg_iter_dist(msg->sg.start, msg->sg.end);
604326b4 DB	205	}
604326b4 DB	206
604326b4 DB	207	static inline struct scatterlist sk_msg_elem(struct sk_msg msg, int which)
	208	{
	209	return &msg->sg.data[which];
	210	}
	211
6fff607e JF	212	static inline struct scatterlist sk_msg_elem_cpy(struct sk_msg *msg, int which)
	213	{
	214	return msg->sg.data[which];
	215	}
	216
604326b4 DB	217	static inline struct page sk_msg_page(struct sk_msg msg, int which)
	218	{
	219	return sg_page(sk_msg_elem(msg, which));
	220	}
	221
	222	static inline bool sk_msg_to_ingress(const struct sk_msg *msg)
	223	{
	224	return msg->flags & BPF_F_INGRESS;
	225	}
	226
	227	static inline void sk_msg_compute_data_pointers(struct sk_msg *msg)
	228	{
	229	struct scatterlist *sge = sk_msg_elem(msg, msg->sg.start);
	230
5a8fb33e	231	if (test_bit(msg->sg.start, msg->sg.copy)) {
604326b4 DB	232	msg->data = NULL;
	233	msg->data_end = NULL;
	234	} else {
	235	msg->data = sg_virt(sge);
	236	msg->data_end = msg->data + sge->length;
	237	}
	238	}
	239
	240	static inline void sk_msg_page_add(struct sk_msg msg, struct page page,
	241	u32 len, u32 offset)
	242	{
	243	struct scatterlist *sge;
	244
	245	get_page(page);
	246	sge = sk_msg_elem(msg, msg->sg.end);
	247	sg_set_page(sge, page, len, offset);
	248	sg_unmark_end(sge);
	249
5a8fb33e	250	__set_bit(msg->sg.end, msg->sg.copy);
604326b4 DB	251	msg->sg.size += len;
	252	sk_msg_iter_next(msg, end);
	253	}
	254
d3b18ad3 JF	255	static inline void sk_msg_sg_copy(struct sk_msg *msg, u32 i, bool copy_state)
	256	{
	257	do {
163ab96b	258	if (copy_state)
5a8fb33e	259	__set_bit(i, msg->sg.copy);
163ab96b	260	else
5a8fb33e	261	__clear_bit(i, msg->sg.copy);
d3b18ad3 JF	262	sk_msg_iter_var_next(i);
	263	if (i == msg->sg.end)
	264	break;
	265	} while (1);
	266	}
	267
	268	static inline void sk_msg_sg_copy_set(struct sk_msg *msg, u32 start)
	269	{
	270	sk_msg_sg_copy(msg, start, true);
	271	}
	272
	273	static inline void sk_msg_sg_copy_clear(struct sk_msg *msg, u32 start)
	274	{
	275	sk_msg_sg_copy(msg, start, false);
	276	}
	277
604326b4 DB	278	static inline struct sk_psock sk_psock(const struct sock sk)
	279	{
	280	return rcu_dereference_sk_user_data(sk);
	281	}
	282
9635720b JF	283	static inline void sk_psock_set_state(struct sk_psock *psock,
	284	enum sk_psock_state_bits bit)
	285	{
	286	set_bit(bit, &psock->state);
	287	}
	288
	289	static inline void sk_psock_clear_state(struct sk_psock *psock,
	290	enum sk_psock_state_bits bit)
	291	{
	292	clear_bit(bit, &psock->state);
	293	}
	294
	295	static inline bool sk_psock_test_state(const struct sk_psock *psock,
	296	enum sk_psock_state_bits bit)
	297	{
	298	return test_bit(bit, &psock->state);
	299	}
	300
	301	static inline void sock_drop(struct sock sk, struct sk_buff skb)
	302	{
	303	sk_drops_add(sk, skb);
	304	kfree_skb(skb);
	305	}
	306
604326b4 DB	307	static inline void sk_psock_queue_msg(struct sk_psock *psock,
	308	struct sk_msg *msg)
	309	{
b01fd6e8	310	spin_lock_bh(&psock->ingress_lock);
9635720b JF	311	if (sk_psock_test_state(psock, SK_PSOCK_TX_ENABLED))
9635720b JF	312	list_add_tail(&msg->list, &psock->ingress_msg);
938d3480 WY	313	else {
	314	sk_msg_free(psock->sk, msg);
	315	kfree(msg);
	316	}
b01fd6e8 CW	317	spin_unlock_bh(&psock->ingress_lock);
	318	}
	319
	320	static inline struct sk_msg sk_psock_dequeue_msg(struct sk_psock psock)
	321	{
	322	struct sk_msg *msg;
	323
	324	spin_lock_bh(&psock->ingress_lock);
	325	msg = list_first_entry_or_null(&psock->ingress_msg, struct sk_msg, list);
	326	if (msg)
	327	list_del(&msg->list);
	328	spin_unlock_bh(&psock->ingress_lock);
	329	return msg;
	330	}
	331
	332	static inline struct sk_msg sk_psock_peek_msg(struct sk_psock psock)
	333	{
	334	struct sk_msg *msg;
	335
	336	spin_lock_bh(&psock->ingress_lock);
	337	msg = list_first_entry_or_null(&psock->ingress_msg, struct sk_msg, list);
	338	spin_unlock_bh(&psock->ingress_lock);
	339	return msg;
	340	}
	341
	342	static inline struct sk_msg sk_psock_next_msg(struct sk_psock psock,
	343	struct sk_msg *msg)
	344	{
	345	struct sk_msg *ret;
	346
	347	spin_lock_bh(&psock->ingress_lock);
	348	if (list_is_last(&msg->list, &psock->ingress_msg))
	349	ret = NULL;
	350	else
	351	ret = list_next_entry(msg, list);
	352	spin_unlock_bh(&psock->ingress_lock);
	353	return ret;
604326b4 DB	354	}
604326b4 DB	355
d3b18ad3 JF	356	static inline bool sk_psock_queue_empty(const struct sk_psock *psock)
	357	{
	358	return psock ? list_empty(&psock->ingress_msg) : true;
	359	}
	360
b01fd6e8 CW	361	static inline void kfree_sk_msg(struct sk_msg *msg)
	362	{
	363	if (msg->skb)
	364	consume_skb(msg->skb);
	365	kfree(msg);
	366	}
	367
604326b4 DB	368	static inline void sk_psock_report_error(struct sk_psock *psock, int err)
	369	{
	370	struct sock *sk = psock->sk;
	371
	372	sk->sk_err = err;
e3ae2365	373	sk_error_report(sk);
604326b4 DB	374	}
	375
	376	struct sk_psock sk_psock_init(struct sock sk, int node);
799aa7f9	377	void sk_psock_stop(struct sk_psock *psock, bool wait);
604326b4	378
88759609	379	#if IS_ENABLED(CONFIG_BPF_STREAM_PARSER)
604326b4 DB	380	int sk_psock_init_strp(struct sock sk, struct sk_psock psock);
	381	void sk_psock_start_strp(struct sock sk, struct sk_psock psock);
	382	void sk_psock_stop_strp(struct sock sk, struct sk_psock psock);
88759609 CW	383	#else
	384	static inline int sk_psock_init_strp(struct sock sk, struct sk_psock psock)
	385	{
	386	return -EOPNOTSUPP;
	387	}
	388
	389	static inline void sk_psock_start_strp(struct sock sk, struct sk_psock psock)
	390	{
	391	}
	392
	393	static inline void sk_psock_stop_strp(struct sock sk, struct sk_psock psock)
	394	{
	395	}
	396	#endif
	397
ef565928 JF	398	void sk_psock_start_verdict(struct sock sk, struct sk_psock psock);
ef565928 JF	399	void sk_psock_stop_verdict(struct sock sk, struct sk_psock psock);
604326b4 DB	400
	401	int sk_psock_msg_verdict(struct sock sk, struct sk_psock psock,
	402	struct sk_msg *msg);
	403
	404	static inline struct sk_psock_link *sk_psock_init_link(void)
	405	{
	406	return kzalloc(sizeof(struct sk_psock_link),
	407	GFP_ATOMIC \| __GFP_NOWARN);
	408	}
	409
	410	static inline void sk_psock_free_link(struct sk_psock_link *link)
	411	{
	412	kfree(link);
	413	}
	414
	415	struct sk_psock_link sk_psock_link_pop(struct sk_psock psock);
604326b4	416
604326b4 DB	417	static inline void sk_psock_cork_free(struct sk_psock *psock)
	418	{
	419	if (psock->cork) {
	420	sk_msg_free(psock->sk, psock->cork);
	421	kfree(psock->cork);
	422	psock->cork = NULL;
	423	}
	424	}
	425
604326b4 DB	426	static inline void sk_psock_restore_proto(struct sock *sk,
	427	struct sk_psock *psock)
	428	{
8a59f9d1	429	if (psock->psock_update_sk_prot)
51e0158a	430	psock->psock_update_sk_prot(sk, psock, true);
604326b4 DB	431	}
604326b4 DB	432
604326b4 DB	433	static inline struct sk_psock sk_psock_get(struct sock sk)
	434	{
	435	struct sk_psock *psock;
	436
	437	rcu_read_lock();
	438	psock = sk_psock(sk);
	439	if (psock && !refcount_inc_not_zero(&psock->refcnt))
	440	psock = NULL;
	441	rcu_read_unlock();
	442	return psock;
	443	}
	444
604326b4 DB	445	void sk_psock_drop(struct sock sk, struct sk_psock psock);
	446
	447	static inline void sk_psock_put(struct sock sk, struct sk_psock psock)
	448	{
	449	if (refcount_dec_and_test(&psock->refcnt))
	450	sk_psock_drop(sk, psock);
	451	}
	452
552de910 JF	453	static inline void sk_psock_data_ready(struct sock sk, struct sk_psock psock)
552de910 JF	454	{
5a685cd9 CW	455	if (psock->saved_data_ready)
5a685cd9 CW	456	psock->saved_data_ready(sk);
552de910 JF	457	else
	458	sk->sk_data_ready(sk);
	459	}
	460
604326b4 DB	461	static inline void psock_set_prog(struct bpf_prog **pprog,
	462	struct bpf_prog *prog)
	463	{
	464	prog = xchg(pprog, prog);
	465	if (prog)
	466	bpf_prog_put(prog);
	467	}
	468
bb0de313 LB	469	static inline int psock_replace_prog(struct bpf_prog **pprog,
	470	struct bpf_prog *prog,
	471	struct bpf_prog *old)
	472	{
	473	if (cmpxchg(pprog, old, prog) != old)
	474	return -ENOENT;
	475
	476	if (old)
	477	bpf_prog_put(old);
	478
	479	return 0;
	480	}
	481
604326b4 DB	482	static inline void psock_progs_drop(struct sk_psock_progs *progs)
	483	{
	484	psock_set_prog(&progs->msg_parser, NULL);
ae8b8332 CW	485	psock_set_prog(&progs->stream_parser, NULL);
ae8b8332 CW	486	psock_set_prog(&progs->stream_verdict, NULL);
a7ba4558	487	psock_set_prog(&progs->skb_verdict, NULL);
604326b4 DB	488	}
604326b4 DB	489
e91de6af JF	490	int sk_psock_tls_strp_read(struct sk_psock psock, struct sk_buff skb);
	491
	492	static inline bool sk_psock_strp_enabled(struct sk_psock *psock)
	493	{
	494	if (!psock)
	495	return false;
5a685cd9	496	return !!psock->saved_data_ready;
e91de6af	497	}
e3526bb9	498
40a34121 JF	499	static inline bool sk_is_udp(const struct sock *sk)
	500	{
	501	return sk->sk_type == SOCK_DGRAM &&
	502	sk->sk_protocol == IPPROTO_UDP;
	503	}
	504
e3526bb9 CW	505	#if IS_ENABLED(CONFIG_NET_SOCK_MSG)
e3526bb9 CW	506
7303524e LJ	507	#define BPF_F_STRPARSER (1UL << 1)
	508
	509	/* We only have two bits so far. */
	510	#define BPF_F_PTR_MASK ~(BPF_F_INGRESS \| BPF_F_STRPARSER)
	511
	512	static inline bool skb_bpf_strparser(const struct sk_buff *skb)
	513	{
	514	unsigned long sk_redir = skb->_sk_redir;
	515
	516	return sk_redir & BPF_F_STRPARSER;
	517	}
	518
	519	static inline void skb_bpf_set_strparser(struct sk_buff *skb)
	520	{
	521	skb->_sk_redir \|= BPF_F_STRPARSER;
	522	}
e3526bb9 CW	523
	524	static inline bool skb_bpf_ingress(const struct sk_buff *skb)
	525	{
	526	unsigned long sk_redir = skb->_sk_redir;
	527
	528	return sk_redir & BPF_F_INGRESS;
	529	}
	530
	531	static inline void skb_bpf_set_ingress(struct sk_buff *skb)
	532	{
	533	skb->_sk_redir \|= BPF_F_INGRESS;
	534	}
	535
	536	static inline void skb_bpf_set_redir(struct sk_buff skb, struct sock sk_redir,
	537	bool ingress)
	538	{
	539	skb->_sk_redir = (unsigned long)sk_redir;
	540	if (ingress)
	541	skb->_sk_redir \|= BPF_F_INGRESS;
	542	}
	543
	544	static inline struct sock skb_bpf_redirect_fetch(const struct sk_buff skb)
	545	{
	546	unsigned long sk_redir = skb->_sk_redir;
	547
	548	return (struct sock *)(sk_redir & BPF_F_PTR_MASK);
	549	}
	550
	551	static inline void skb_bpf_redirect_clear(struct sk_buff *skb)
	552	{
	553	skb->_sk_redir = 0;
	554	}
	555	#endif /* CONFIG_NET_SOCK_MSG */
604326b4	556	#endif /* _LINUX_SKMSG_H */