path: root/net/rds/rds.h

/* SPDX-License-Identifier: GPL-2.0 */
#ifndef _RDS_RDS_H
#define _RDS_RDS_H

#include <net/sock.h>
#include <linux/scatterlist.h>
#include <linux/highmem.h>
#include <rdma/rdma_cm.h>
#include <linux/mutex.h>
#include <linux/rds.h>
#include <linux/rhashtable.h>
#include <linux/refcount.h>
#include <linux/in6.h>

#include "info.h"

/*
 * RDS Network protocol version
 */
#define RDS_PROTOCOL_3_0	0x0300
#define RDS_PROTOCOL_3_1	0x0301
#define RDS_PROTOCOL_4_0	0x0400
#define RDS_PROTOCOL_4_1	0x0401
#define RDS_PROTOCOL_VERSION	RDS_PROTOCOL_3_1
#define RDS_PROTOCOL_MAJOR(v)	((v) >> 8)
#define RDS_PROTOCOL_MINOR(v)	((v) & 255)
#define RDS_PROTOCOL(maj, min)	(((maj) << 8) | min)
#define RDS_PROTOCOL_COMPAT_VERSION	RDS_PROTOCOL_3_1

/* The following ports, 16385, 18634, 18635, are registered with IANA as
 * the ports to be used for RDS over TCP and UDP.  Currently, only RDS over
 * TCP and RDS over IB/RDMA are implemented.  18634 is the historical value
 * used for the RDMA_CM listener port.  RDS/TCP uses port 16385.  After
 * IPv6 work, RDMA_CM also uses 16385 as the listener port.  18634 is kept
 * to ensure compatibility with older RDS modules.  Those ports are defined
 * in each transport's header file.
 */
#define RDS_PORT	18634

#ifdef ATOMIC64_INIT
#define KERNEL_HAS_ATOMIC64
#endif
#ifdef RDS_DEBUG
#define rdsdebug(fmt, args...) pr_debug("%s(): " fmt, __func__ , ##args)
#else
/* sigh, pr_debug() causes unused variable warnings */
static inline __printf(1, 2)
void rdsdebug(char *fmt, ...)
{
}
#endif

#define RDS_FRAG_SHIFT	12
#define RDS_FRAG_SIZE	((unsigned int)(1 << RDS_FRAG_SHIFT))

/* Used to limit both RDMA and non-RDMA RDS message to 1MB */
#define RDS_MAX_MSG_SIZE	((unsigned int)(1 << 20))

#define RDS_CONG_MAP_BYTES	(65536 / 8)
#define RDS_CONG_MAP_PAGES	(PAGE_ALIGN(RDS_CONG_MAP_BYTES) / PAGE_SIZE)
#define RDS_CONG_MAP_PAGE_BITS	(PAGE_SIZE * 8)

struct rds_cong_map {
	struct rb_node		m_rb_node;
	struct in6_addr		m_addr;
	wait_queue_head_t	m_waitq;
	struct list_head	m_conn_list;
	unsigned long		m_page_addrs[RDS_CONG_MAP_PAGES];
};


/*
 * This is how we will track the connection state:
 * A connection is always in one of the following
 * states. Updates to the state are atomic and imply
 * a memory barrier.
 */
enum {
	RDS_CONN_DOWN = 0,
	RDS_CONN_CONNECTING,
	RDS_CONN_DISCONNECTING,
	RDS_CONN_UP,
	RDS_CONN_RESETTING,
	RDS_CONN_ERROR,
};

/* Bits for c_flags */
#define RDS_LL_SEND_FULL	0
#define RDS_RECONNECT_PENDING	1
#define RDS_IN_XMIT		2
#define RDS_RECV_REFILL		3
#define	RDS_DESTROY_PENDING	4

/* Max number of multipaths per RDS connection. Must be a power of 2 */
#define	RDS_MPATH_WORKERS	8
#define	RDS_MPATH_HASH(rs, n) (jhash_1word((rs)->rs_bound_port, \
			       (rs)->rs_hash_initval) & ((n) - 1))

#define IS_CANONICAL(laddr, faddr) (htonl(laddr) < htonl(faddr))

/* Per mpath connection state */
struct rds_conn_path {
	struct rds_connection	*cp_conn;
	struct rds_message	*cp_xmit_rm;
	unsigned long		cp_xmit_sg;
	unsigned int		cp_xmit_hdr_off;
	unsigned int		cp_xmit_data_off;
	unsigned int		cp_xmit_atomic_sent;
	unsigned int		cp_xmit_rdma_sent;
	unsigned int		cp_xmit_data_sent;

	spinlock_t		cp_lock;		/* protect msg queues */
	u64			cp_next_tx_seq;
	struct list_head	cp_send_queue;
	struct list_head	cp_retrans;

	u64			cp_next_rx_seq;

	void			*cp_transport_data;

	atomic_t		cp_state;
	unsigned long		cp_send_gen;
	unsigned long		cp_flags;
	unsigned long		cp_reconnect_jiffies;
	struct delayed_work	cp_send_w;
	struct delayed_work	cp_recv_w;
	struct delayed_work	cp_conn_w;
	struct work_struct	cp_down_w;
	struct mutex		cp_cm_lock;	/* protect cp_state & cm */
	wait_queue_head_t	cp_waitq;

	unsigned int		cp_unacked_packets;
	unsigned int		cp_unacked_bytes;
	unsigned int		cp_index;
};

/* One rds_connection per RDS address pair */
struct rds_connection {
	struct hlist_node	c_hash_node;
	struct in6_addr		c_laddr;
	struct in6_addr		c_faddr;
	int			c_dev_if; /* ifindex used for this conn */
	int			c_bound_if; /* ifindex of c_laddr */
	unsigned int		c_loopback:1,
				c_isv6:1,
				c_ping_triggered:1,
				c_pad_to_32:29;
	int			c_npaths;
	struct rds_connection	*c_passive;
	struct rds_transport	*c_trans;

	struct rds_cong_map	*c_lcong;
	struct rds_cong_map	*c_fcong;

	/* Protocol version */
	unsigned int		c_proposed_version;
	unsigned int		c_version;
	possible_net_t		c_net;

	/* TOS */
	u8			c_tos;

	struct list_head	c_map_item;
	unsigned long		c_map_queued;

	struct rds_conn_path	*c_path;
	wait_queue_head_t	c_hs_waitq; /* handshake waitq */

	u32			c_my_gen_num;
	u32			c_peer_gen_num;
};

static inline
struct net *rds_conn_net(struct rds_connection *conn)
{
	return read_pnet(&conn->c_net);
}

static inline
void rds_conn_net_set(struct rds_connection *conn, struct net *net)
{
	write_pnet(&conn->c_net, net);
}

#define RDS_FLAG_CONG_BITMAP	0x01
#define RDS_FLAG_ACK_REQUIRED	0x02
#define RDS_FLAG_RETRANSMITTED	0x04
#define RDS_MAX_ADV_CREDIT	255

/* RDS_FLAG_PROBE_PORT is the reserved sport used for sending a ping
 * probe to exchange control information before establishing a connection.
 * Currently the control information that is exchanged is the number of
 * supported paths. If the peer is a legacy (older kernel revision) peer,
 * it would return a pong message without additional control information
 * that would then alert the sender that the peer was an older rev.
 */
#define RDS_FLAG_PROBE_PORT	1
#define	RDS_HS_PROBE(sport, dport) \
		((sport == RDS_FLAG_PROBE_PORT && dport == 0) || \
		 (sport == 0 && dport == RDS_FLAG_PROBE_PORT))
/*
 * Maximum space available for extension headers.
 */
#define RDS_HEADER_EXT_SPACE	16

struct rds_header {
	__be64	h_sequence;
	__be64	h_ack;
	__be32	h_len;
	__be16	h_sport;
	__be16	h_dport;
	u8	h_flags;
	u8	h_credit;
	u8	h_padding[4];
	__sum16	h_csum;

	u8	h_exthdr[RDS_HEADER_EXT_SPACE];
};

/*
 * Reserved - indicates end of extensions
 */
#define RDS_EXTHDR_NONE		0

/*
 * This extension header is included in the very
 * first message that is sent on a new connection,
 * and identifies the protocol level. This will help
 * rolling updates if a future change requires breaking
 * the protocol.
 * NB: This is no longer true for IB, where we do a version