[linux.git] / net / smc / smc.h

/* SPDX-License-Identifier: GPL-2.0 */
/*
 *  Shared Memory Communications over RDMA (SMC-R) and RoCE
 *
 *  Definitions for the SMC module (socket related)
 *
 *  Copyright IBM Corp. 2016
 *
 *  Author(s):  Ursula Braun <[email protected]>
 */
#ifndef __SMC_H
#define __SMC_H

#include <linux/socket.h>
#include <linux/types.h>
#include <linux/compiler.h> /* __aligned */
#include <net/sock.h>

#include "smc_ib.h"

#define SMCPROTO_SMC		0	/* SMC protocol, IPv4 */
#define SMCPROTO_SMC6		1	/* SMC protocol, IPv6 */

extern struct proto smc_proto;
extern struct proto smc_proto6;

#ifdef ATOMIC64_INIT
#define KERNEL_HAS_ATOMIC64
#endif

enum smc_state {		/* possible states of an SMC socket */
	SMC_ACTIVE	= 1,
	SMC_INIT	= 2,
	SMC_CLOSED	= 7,
	SMC_LISTEN	= 10,
	/* normal close */
	SMC_PEERCLOSEWAIT1	= 20,
	SMC_PEERCLOSEWAIT2	= 21,
	SMC_APPFINCLOSEWAIT	= 24,
	SMC_APPCLOSEWAIT1	= 22,
	SMC_APPCLOSEWAIT2	= 23,
	SMC_PEERFINCLOSEWAIT	= 25,
	/* abnormal close */
	SMC_PEERABORTWAIT	= 26,
	SMC_PROCESSABORT	= 27,
};

struct smc_link_group;

struct smc_wr_rx_hdr {	/* common prefix part of LLC and CDC to demultiplex */
	u8			type;
} __aligned(1);

struct smc_cdc_conn_state_flags {
#if defined(__BIG_ENDIAN_BITFIELD)
	u8	peer_done_writing : 1;	/* Sending done indicator */
	u8	peer_conn_closed : 1;	/* Peer connection closed indicator */
	u8	peer_conn_abort : 1;	/* Abnormal close indicator */
	u8	reserved : 5;
#elif defined(__LITTLE_ENDIAN_BITFIELD)
	u8	reserved : 5;
	u8	peer_conn_abort : 1;
	u8	peer_conn_closed : 1;
	u8	peer_done_writing : 1;
#endif
};

struct smc_cdc_producer_flags {
#if defined(__BIG_ENDIAN_BITFIELD)
	u8	write_blocked : 1;	/* Writing Blocked, no rx buf space */
	u8	urg_data_pending : 1;	/* Urgent Data Pending */
	u8	urg_data_present : 1;	/* Urgent Data Present */
	u8	cons_curs_upd_req : 1;	/* cursor update requested */
	u8	failover_validation : 1;/* message replay due to failover */
	u8	reserved : 3;
#elif defined(__LITTLE_ENDIAN_BITFIELD)
	u8	reserved : 3;
	u8	failover_validation : 1;
	u8	cons_curs_upd_req : 1;
	u8	urg_data_present : 1;
	u8	urg_data_pending : 1;
	u8	write_blocked : 1;
#endif
};

/* in host byte order */
union smc_host_cursor {	/* SMC cursor - an offset in an RMBE */
	struct {
		u16	reserved;
		u16	wrap;		/* window wrap sequence number */
		u32	count;		/* cursor (= offset) part */
	};
#ifdef KERNEL_HAS_ATOMIC64
	atomic64_t		acurs;	/* for atomic processing */
#else
	u64			acurs;	/* for atomic processing */
#endif
} __aligned(8);

/* in host byte order, except for flag bitfields in network byte order */
struct smc_host_cdc_msg {		/* Connection Data Control message */
	struct smc_wr_rx_hdr		common; /* .type = 0xFE */
	u8				len;	/* length = 44 */
	u16				seqno;	/* connection seq # */
	u32				token;	/* alert_token */
	union smc_host_cursor		prod;		/* producer cursor */
	union smc_host_cursor		cons;		/* consumer cursor,
							 * piggy backed "ack"
							 */
	struct smc_cdc_producer_flags	prod_flags;	/* conn. tx/rx status */
	struct smc_cdc_conn_state_flags	conn_state_flags; /* peer conn. status*/
	u8				reserved[18];
} __aligned(8);

enum smc_urg_state {
	SMC_URG_VALID	= 1,			/* data present */
	SMC_URG_NOTYET	= 2,			/* data pending */
	SMC_URG_READ	= 3,			/* data was already read */
};

struct smc_connection {
	struct rb_node		alert_node;
	struct smc_link_group	*lgr;		/* link group of connection */
	struct smc_link		*lnk;		/* assigned SMC-R link */
	u32			alert_token_local; /* unique conn. id */
	u8			peer_rmbe_idx;	/* from tcp handshake */
	int			peer_rmbe_size;	/* size of peer rx buffer */
	atomic_t		peer_rmbe_space;/* remaining free bytes in peer
						 * rmbe
						 */
	int			rtoken_idx;	/* idx to peer RMB rkey/addr */

	struct smc_buf_desc	*sndbuf_desc;	/* send buffer descriptor */
	struct smc_buf_desc	*rmb_desc;	/* RMBE descriptor */
	int			rmbe_size_short;/* compressed notation */
	int			rmbe_update_limit;
						/* lower limit for consumer
						 * cursor update
						 */

	struct smc_host_cdc_msg	local_tx_ctrl;	/* host byte order staging
						 * buffer for CDC msg send
						 * .prod cf. TCP snd_nxt
						 * .cons cf. TCP sends ack
						 */
	union smc_host_cursor	local_tx_ctrl_fin;
						/* prod crsr - confirmed by peer
						 */
	union smc_host_cursor	tx_curs_prep;	/* tx - prepared data
						 * snd_max..wmem_alloc
						 */
	union smc_host_cursor	tx_curs_sent;	/* tx - sent data
						 * snd_nxt ?
						 */
	union smc_host_cursor	tx_curs_fin;	/* tx - confirmed by peer
						 * snd-wnd-begin ?
						 */
	atomic_t		sndbuf_space;	/* remaining space in sndbuf */
	u16			tx_cdc_seq;	/* sequence # for CDC send */
	u16			tx_cdc_seq_fin;	/* sequence # - tx completed */
	spinlock_t		send_lock;	/* protect wr_sends */
	struct delayed_work	tx_work;	/* retry of smc_cdc_msg_send */
	u32			tx_off;		/* base offset in peer rmb */

	struct smc_host_cdc_msg	local_rx_ctrl;	/* filled during event_handl.
						 * .prod cf. TCP rcv_nxt
						 * .cons cf. TCP snd_una
						 */
	union smc_host_cursor	rx_curs_confirmed; /* confirmed to peer
						    * source of snd_una ?
						    */
	union smc_host_cursor	urg_curs;	/* points at urgent byte */
	enum smc_urg_state	urg_state;
	bool			urg_tx_pend;	/* urgent data staged */
	bool			urg_rx_skip_pend;
						/* indicate urgent oob data
						 * read, but previous regular
						 * data still pending
						 */
	char			urg_rx_byte;	/* urgent byte */
	atomic_t		bytes_to_rcv;	/* arrived data,
						 * not yet received
						 */
	atomic_t		splice_pending;	/* number of spliced bytes
						 * pending processing
						 */
#ifndef KERNEL_HAS_ATOMIC64
	spinlock_t		acurs_lock;	/* protect cursors */
#endif
	struct work_struct	close_work;	/* peer sent some closing */
	struct work_struct	abort_work;	/* abort the connection */
	struct tasklet_struct	rx_tsklet;	/* Receiver tasklet for SMC-D */
	u8			rx_off;		/* receive offset:
						 * 0 for SMC-R, 32 for SMC-D
						 */
	u64			peer_token;	/* SMC-D token of peer */
	u8			killed : 1;	/* abnormal termination */
	u8			out_of_sync : 1; /* out of sync with peer */
};

struct smc_sock {				/* smc sock container */
	struct sock		sk;
	struct socket		*clcsock;	/* internal tcp socket */
	struct smc_connection	conn;		/* smc connection */
	struct smc_sock		*listen_smc;	/* listen parent */
	struct work_struct	connect_work;	/* handle non-blocking connect*/
	struct work_struct	tcp_listen_work;/* handle tcp socket accepts */
	struct work_struct	smc_listen_work;/* prepare new accept socket */
	struct list_head	accept_q;	/* sockets to be accepted */
	spinlock_t		accept_q_lock;	/* protects accept_q */
	bool			use_fallback;	/* fallback to tcp */
	int			fallback_rsn;	/* reason for fallback */
	u32			peer_diagnosis; /* decline reason from peer */
	int			sockopt_defer_accept;
						/* sockopt TCP_DEFER_ACCEPT
						 * value
						 */
	u8			wait_close_tx_prepared : 1;
						/* shutdown wr or close
						 * started, waiting for unsent
						 * data to be sent
						 */
	u8			connect_nonblock : 1;
						/* non-blocking connect in
						 * flight
						 */
	struct mutex            clcsock_release_lock;
						/* protects clcsock of a listen
						 * socket
						 * */
};

static inline struct smc_sock *smc_sk(const struct sock *sk)
{
	return (struct smc_sock *)sk;
}

#define SMC_SYSTEMID_LEN		8

extern u8	local_systemid[SMC_SYSTEMID_LEN]; /* unique system identifier */

/* convert an u32 value into network byte order, store it into a 3 byte field */
static inline void hton24(u8 *net, u32 host)
{
	__be32 t;

	t = cpu_to_be32(host);
	memcpy(net, ((u8 *)&t) + 1, 3);
}

/* convert a received 3 byte field into host byte order*/
static inline u32 ntoh24(u8 *net)
{
	__be32 t = 0;

	memcpy(((u8 *)&t) + 1, net, 3);
	return be32_to_cpu(t);
}

#ifdef CONFIG_XFRM
static inline bool using_ipsec(struct smc_sock *smc)
{
	return (smc->clcsock->sk->sk_policy[0] ||
		smc->clcsock->sk->sk_policy[1]) ? true : false;
}
#else
static inline bool using_ipsec(struct smc_sock *smc)
{
	return false;
}
#endif

struct sock *smc_accept_dequeue(struct sock *parent, struct socket *new_sock);
void smc_close_non_accepted(struct sock *sk);

#endif	/* __SMC_H */
Commit	Line	Data
b2441318	1	/* SPDX-License-Identifier: GPL-2.0 */
ac713874 UB	2	/*
	3	* Shared Memory Communications over RDMA (SMC-R) and RoCE
	4	*
	5	* Definitions for the SMC module (socket related)
	6	*
	7	* Copyright IBM Corp. 2016
	8	*
	9	* Author(s): Ursula Braun <[email protected]>
	10	*/
	11	#ifndef __SMC_H
	12	#define __SMC_H
	13
	14	#include <linux/socket.h>
	15	#include <linux/types.h>
f38ba179	16	#include <linux/compiler.h> /* __aligned */
ac713874 UB	17	#include <net/sock.h>
ac713874 UB	18
0cfdd8f9 UB	19	#include "smc_ib.h"
0cfdd8f9 UB	20
aaa4d33f KG	21	#define SMCPROTO_SMC 0 /* SMC protocol, IPv4 */
aaa4d33f KG	22	#define SMCPROTO_SMC6 1 /* SMC protocol, IPv6 */
ac713874	23
f16a7dd5	24	extern struct proto smc_proto;
aaa4d33f	25	extern struct proto smc_proto6;
f16a7dd5	26
5f08318f UB	27	#ifdef ATOMIC64_INIT
	28	#define KERNEL_HAS_ATOMIC64
	29	#endif
	30
ac713874 UB	31	enum smc_state { /* possible states of an SMC socket */
	32	SMC_ACTIVE = 1,
	33	SMC_INIT = 2,
	34	SMC_CLOSED = 7,
	35	SMC_LISTEN = 10,
b38d7324 UB	36	/* normal close */
	37	SMC_PEERCLOSEWAIT1 = 20,
	38	SMC_PEERCLOSEWAIT2 = 21,
	39	SMC_APPFINCLOSEWAIT = 24,
	40	SMC_APPCLOSEWAIT1 = 22,
	41	SMC_APPCLOSEWAIT2 = 23,
	42	SMC_PEERFINCLOSEWAIT = 25,
	43	/* abnormal close */
	44	SMC_PEERABORTWAIT = 26,
	45	SMC_PROCESSABORT = 27,
ac713874 UB	46	};
ac713874 UB	47
0cfdd8f9 UB	48	struct smc_link_group;
0cfdd8f9 UB	49
f38ba179 UB	50	struct smc_wr_rx_hdr { /* common prefix part of LLC and CDC to demultiplex */
	51	u8 type;
	52	} __aligned(1);
	53
5f08318f UB	54	struct smc_cdc_conn_state_flags {
	55	#if defined(__BIG_ENDIAN_BITFIELD)
	56	u8 peer_done_writing : 1; /* Sending done indicator */
	57	u8 peer_conn_closed : 1; /* Peer connection closed indicator */
	58	u8 peer_conn_abort : 1; /* Abnormal close indicator */
	59	u8 reserved : 5;
	60	#elif defined(__LITTLE_ENDIAN_BITFIELD)
	61	u8 reserved : 5;
	62	u8 peer_conn_abort : 1;
	63	u8 peer_conn_closed : 1;
	64	u8 peer_done_writing : 1;
	65	#endif
	66	};
	67
	68	struct smc_cdc_producer_flags {
	69	#if defined(__BIG_ENDIAN_BITFIELD)
	70	u8 write_blocked : 1; /* Writing Blocked, no rx buf space */
	71	u8 urg_data_pending : 1; /* Urgent Data Pending */
	72	u8 urg_data_present : 1; /* Urgent Data Present */
	73	u8 cons_curs_upd_req : 1; /* cursor update requested */
	74	u8 failover_validation : 1;/* message replay due to failover */
	75	u8 reserved : 3;
	76	#elif defined(__LITTLE_ENDIAN_BITFIELD)
	77	u8 reserved : 3;
	78	u8 failover_validation : 1;
	79	u8 cons_curs_upd_req : 1;
	80	u8 urg_data_present : 1;
	81	u8 urg_data_pending : 1;
	82	u8 write_blocked : 1;
	83	#endif
	84	};
	85
	86	/* in host byte order */
	87	union smc_host_cursor { /* SMC cursor - an offset in an RMBE */
	88	struct {
	89	u16 reserved;
	90	u16 wrap; /* window wrap sequence number */
	91	u32 count; /* cursor (= offset) part */
	92	};
	93	#ifdef KERNEL_HAS_ATOMIC64
	94	atomic64_t acurs; /* for atomic processing */
	95	#else
	96	u64 acurs; /* for atomic processing */
	97	#endif
	98	} __aligned(8);
	99
	100	/* in host byte order, except for flag bitfields in network byte order */
	101	struct smc_host_cdc_msg { /* Connection Data Control message */
	102	struct smc_wr_rx_hdr common; /* .type = 0xFE */
	103	u8 len; /* length = 44 */
	104	u16 seqno; /* connection seq # */
	105	u32 token; /* alert_token */
	106	union smc_host_cursor prod; /* producer cursor */
	107	union smc_host_cursor cons; /* consumer cursor,
	108	* piggy backed "ack"
	109	*/
	110	struct smc_cdc_producer_flags prod_flags; /* conn. tx/rx status */
	111	struct smc_cdc_conn_state_flags conn_state_flags; /* peer conn. status*/
	112	u8 reserved[18];
	113	} __aligned(8);
	114
de8474eb	115	enum smc_urg_state {
d7cf4a3b UB	116	SMC_URG_VALID = 1, /* data present */
	117	SMC_URG_NOTYET = 2, /* data pending */
	118	SMC_URG_READ = 3, /* data was already read */
de8474eb SR	119	};
de8474eb SR	120
0cfdd8f9 UB	121	struct smc_connection {
	122	struct rb_node alert_node;
	123	struct smc_link_group lgr; / link group of connection */
387707fd	124	struct smc_link lnk; / assigned SMC-R link */
0cfdd8f9	125	u32 alert_token_local; /* unique conn. id */
92a138e3	126	u8 peer_rmbe_idx; /* from tcp handshake */
cd6851f3 UB	127	int peer_rmbe_size; /* size of peer rx buffer */
	128	atomic_t peer_rmbe_space;/* remaining free bytes in peer
	129	* rmbe
	130	*/
bd4ad577	131	int rtoken_idx; /* idx to peer RMB rkey/addr */
cd6851f3 UB	132
cd6851f3 UB	133	struct smc_buf_desc sndbuf_desc; / send buffer descriptor */
cd6851f3	134	struct smc_buf_desc rmb_desc; / RMBE descriptor */
cd6851f3	135	int rmbe_size_short;/* compressed notation */
952310cc UB	136	int rmbe_update_limit;
	137	/* lower limit for consumer
	138	* cursor update
	139	*/
5f08318f UB	140
	141	struct smc_host_cdc_msg local_tx_ctrl; /* host byte order staging
	142	* buffer for CDC msg send
	143	* .prod cf. TCP snd_nxt
	144	* .cons cf. TCP sends ack
	145	*/
f0ec4f1d KG	146	union smc_host_cursor local_tx_ctrl_fin;
	147	/* prod crsr - confirmed by peer
	148	*/
5f08318f UB	149	union smc_host_cursor tx_curs_prep; /* tx - prepared data
	150	* snd_max..wmem_alloc
	151	*/
	152	union smc_host_cursor tx_curs_sent; /* tx - sent data
	153	* snd_nxt ?
	154	*/
	155	union smc_host_cursor tx_curs_fin; /* tx - confirmed by peer
	156	* snd-wnd-begin ?
	157	*/
	158	atomic_t sndbuf_space; /* remaining space in sndbuf */
	159	u16 tx_cdc_seq; /* sequence # for CDC send */
f0ec4f1d	160	u16 tx_cdc_seq_fin; /* sequence # - tx completed */
5f08318f	161	spinlock_t send_lock; /* protect wr_sends */
18e537cd	162	struct delayed_work tx_work; /* retry of smc_cdc_msg_send */
95d8d263	163	u32 tx_off; /* base offset in peer rmb */
5f08318f UB	164
	165	struct smc_host_cdc_msg local_rx_ctrl; /* filled during event_handl.
	166	* .prod cf. TCP rcv_nxt
	167	* .cons cf. TCP snd_una
	168	*/
	169	union smc_host_cursor rx_curs_confirmed; /* confirmed to peer
	170	* source of snd_una ?
	171	*/
de8474eb SR	172	union smc_host_cursor urg_curs; /* points at urgent byte */
	173	enum smc_urg_state urg_state;
	174	bool urg_tx_pend; /* urgent data staged */
	175	bool urg_rx_skip_pend;
	176	/* indicate urgent oob data
	177	* read, but previous regular
	178	* data still pending
	179	*/
	180	char urg_rx_byte; /* urgent byte */
5f08318f UB	181	atomic_t bytes_to_rcv; /* arrived data,
	182	* not yet received
	183	*/
9014db20 SR	184	atomic_t splice_pending; /* number of spliced bytes
	185	* pending processing
	186	*/
5f08318f UB	187	#ifndef KERNEL_HAS_ATOMIC64
	188	spinlock_t acurs_lock; /* protect cursors */
	189	#endif
46c28dbd	190	struct work_struct close_work; /* peer sent some closing */
b286a065	191	struct work_struct abort_work; /* abort the connection */
be244f28 HW	192	struct tasklet_struct rx_tsklet; /* Receiver tasklet for SMC-D */
	193	u8 rx_off; /* receive offset:
	194	* 0 for SMC-R, 32 for SMC-D
	195	*/
	196	u64 peer_token; /* SMC-D token of peer */
b2900980	197	u8 killed : 1; /* abnormal termination */
b286a065	198	u8 out_of_sync : 1; /* out of sync with peer */
0cfdd8f9 UB	199	};
0cfdd8f9 UB	200
ac713874 UB	201	struct smc_sock { /* smc sock container */
	202	struct sock sk;
	203	struct socket clcsock; / internal tcp socket */
0cfdd8f9	204	struct smc_connection conn; /* smc connection */
a046d57d	205	struct smc_sock listen_smc; / listen parent */
24ac3a08	206	struct work_struct connect_work; /* handle non-blocking connect*/
a046d57d UB	207	struct work_struct tcp_listen_work;/* handle tcp socket accepts */
	208	struct work_struct smc_listen_work;/* prepare new accept socket */
	209	struct list_head accept_q; /* sockets to be accepted */
	210	spinlock_t accept_q_lock; /* protects accept_q */
ac713874	211	bool use_fallback; /* fallback to tcp */
603cc149 KG	212	int fallback_rsn; /* reason for fallback */
603cc149 KG	213	u32 peer_diagnosis; /* decline reason from peer */
abb190f1 UB	214	int sockopt_defer_accept;
	215	/* sockopt TCP_DEFER_ACCEPT
	216	* value
	217	*/
b38d7324 UB	218	u8 wait_close_tx_prepared : 1;
	219	/* shutdown wr or close
	220	* started, waiting for unsent
	221	* data to be sent
	222	*/
50717a37 UB	223	u8 connect_nonblock : 1;
	224	/* non-blocking connect in
	225	* flight
	226	*/
78abe3d0 MJ	227	struct mutex clcsock_release_lock;
	228	/* protects clcsock of a listen
	229	* socket
	230	* */
ac713874 UB	231	};
	232
	233	static inline struct smc_sock smc_sk(const struct sock sk)
	234	{
	235	return (struct smc_sock *)sk;
	236	}
	237
a4cf0443 UB	238	#define SMC_SYSTEMID_LEN 8
	239
	240	extern u8 local_systemid[SMC_SYSTEMID_LEN]; /* unique system identifier */
	241
0cfdd8f9 UB	242	/* convert an u32 value into network byte order, store it into a 3 byte field */
	243	static inline void hton24(u8 *net, u32 host)
	244	{
	245	__be32 t;
	246
	247	t = cpu_to_be32(host);
	248	memcpy(net, ((u8 *)&t) + 1, 3);
	249	}
	250
	251	/* convert a received 3 byte field into host byte order*/
	252	static inline u32 ntoh24(u8 *net)
	253	{
	254	__be32 t = 0;
	255
	256	memcpy(((u8 *)&t) + 1, net, 3);
	257	return be32_to_cpu(t);
	258	}
	259
a046d57d UB	260	#ifdef CONFIG_XFRM
	261	static inline bool using_ipsec(struct smc_sock *smc)
	262	{
	263	return (smc->clcsock->sk->sk_policy[0] \|\|
a8fbf8e7	264	smc->clcsock->sk->sk_policy[1]) ? true : false;
a046d57d UB	265	}
	266	#else
	267	static inline bool using_ipsec(struct smc_sock *smc)
	268	{
a8fbf8e7	269	return false;
a046d57d UB	270	}
	271	#endif
	272
b38d7324 UB	273	struct sock smc_accept_dequeue(struct sock parent, struct socket *new_sock);
b38d7324 UB	274	void smc_close_non_accepted(struct sock *sk);
a046d57d	275
ac713874	276	#endif /* __SMC_H */