Subversion Repositories shark

Rev

Go to most recent revision | Blame | Compare with Previous | Last modification | View Log | RSS feed

/*
 * 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, <waltje@uWalt.NL.Mugnet.ORG>
 *
 *              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 _LINUX_TCP_H
#define _LINUX_TCP_H

#include <linux/types.h>
#include <asm/byteorder.h>

struct tcphdr {
        __u16   source;
        __u16   dest;
        __u32   seq;
        __u32   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 
        __u16   window;
        __u16   check;
        __u16   urg_ptr;
};


enum {
  TCP_ESTABLISHED = 1,
  TCP_SYN_SENT,
  TCP_SYN_RECV,
  TCP_FIN_WAIT1,
  TCP_FIN_WAIT2,
  TCP_TIME_WAIT,
  TCP_CLOSE,
  TCP_CLOSE_WAIT,
  TCP_LAST_ACK,
  TCP_LISTEN,
  TCP_CLOSING,   /* now a valid state */

  TCP_MAX_STATES /* Leave at the end! */
};

#define TCP_STATE_MASK  0xF
#define TCP_ACTION_FIN  (1 << 7)

enum {
  TCPF_ESTABLISHED = (1 << 1),
  TCPF_SYN_SENT  = (1 << 2),
  TCPF_SYN_RECV  = (1 << 3),
  TCPF_FIN_WAIT1 = (1 << 4),
  TCPF_FIN_WAIT2 = (1 << 5),
  TCPF_TIME_WAIT = (1 << 6),
  TCPF_CLOSE     = (1 << 7),
  TCPF_CLOSE_WAIT = (1 << 8),
  TCPF_LAST_ACK  = (1 << 9),
  TCPF_LISTEN    = (1 << 10),
  TCPF_CLOSING   = (1 << 11)
};

/*
 *      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;
        __u32             words[5];
};

#define tcp_flag_word(tp) ( ((union tcp_word_hdr *)(tp))->words [3])

enum {
        TCP_FLAG_CWR = __constant_htonl(0x00800000),
        TCP_FLAG_ECE = __constant_htonl(0x00400000),
        TCP_FLAG_URG = __constant_htonl(0x00200000),
        TCP_FLAG_ACK = __constant_htonl(0x00100000),
        TCP_FLAG_PSH = __constant_htonl(0x00080000),
        TCP_FLAG_RST = __constant_htonl(0x00040000),
        TCP_FLAG_SYN = __constant_htonl(0x00020000),
        TCP_FLAG_FIN = __constant_htonl(0x00010000),
        TCP_RESERVED_BITS = __constant_htonl(0x0F000000),
        TCP_DATA_OFFSET = __constant_htonl(0xF0000000)
};

/* 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 TCPI_OPT_TIMESTAMPS     1
#define TCPI_OPT_SACK           2
#define TCPI_OPT_WSCALE         4
#define TCPI_OPT_ECN            8

enum tcp_ca_state
{
        TCP_CA_Open = 0,
#define TCPF_CA_Open    (1<<TCP_CA_Open)
        TCP_CA_Disorder = 1,
#define TCPF_CA_Disorder (1<<TCP_CA_Disorder)
        TCP_CA_CWR = 2,
#define TCPF_CA_CWR     (1<<TCP_CA_CWR)
        TCP_CA_Recovery = 3,
#define TCPF_CA_Recovery (1<<TCP_CA_Recovery)
        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;

        __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;
};

#ifdef __KERNEL__

#include <linux/config.h>
#include <linux/skbuff.h>
#include <linux/ip.h>
#include <net/sock.h>

/* This defines a selective acknowledgement block. */
struct tcp_sack_block {
        __u32   start_seq;
        __u32   end_seq;
};

struct tcp_opt {
        int     tcp_header_len; /* Bytes of tcp header to send          */

/*
 *      Header prediction flags
 *      0x5?10 << 16 + snd_wnd in net byte order
 */

        __u32   pred_flags;

/*
 *      RFC793 variables by their proper names. This means you can
 *      read the code and the spec side by side (and laugh ...)
 *      See RFC793 and RFC1122. The RFC writes these in capitals.
 */

        __u32   rcv_nxt;        /* What we want to receive next         */
        __u32   snd_nxt;        /* Next sequence we send                */

        __u32   snd_una;        /* First byte we want an ack for        */
        __u32   snd_sml;        /* Last byte of the most recently transmitted small packet */
        __u32   rcv_tstamp;     /* timestamp of last received ACK (for keepalives) */
        __u32   lsndtime;       /* timestamp of last sent data packet (for restart window) */
        struct tcp_bind_bucket *bind_hash;
        /* Delayed ACK control data */
        struct {
                __u8    pending;        /* ACK is pending */
                __u8    quick;          /* Scheduled number of quick acks       */
                __u8    pingpong;       /* The session is interactive           */
                __u8    blocked;        /* Delayed ACK was blocked by socket lock*/
                __u32   ato;            /* Predicted tick of soft clock         */
                unsigned long timeout;  /* Currently scheduled timeout          */
                __u32   lrcvtime;       /* timestamp of last received data packet*/
                __u16   last_seg_size;  /* Size of last incoming segment        */
                __u16   rcv_mss;        /* MSS used for delayed ACK decisions   */
        } ack;

        /* Data for direct copy to user */
        struct {
                struct sk_buff_head     prequeue;
                struct task_struct      *task;
                struct iovec            *iov;
                int                     memory;
                int                     len;
        } ucopy;

        __u32   snd_wl1;        /* Sequence for window update           */
        __u32   snd_wnd;        /* The window we expect to receive      */
        __u32   max_window;     /* Maximal window ever seen from peer   */
        __u32   pmtu_cookie;    /* Last pmtu seen by socket             */
        __u32   mss_cache;      /* Cached effective mss, not including SACKS */
        __u16   mss_cache_std;  /* Like mss_cache, but without TSO */
        __u16   mss_clamp;      /* Maximal mss, negotiated at connection setup */
        __u16   ext_header_len; /* Network protocol overhead (IP/IPv6 options) */
        __u16   ext2_header_len;/* Options depending on route */
        __u8    ca_state;       /* State of fast-retransmit machine     */
        __u8    retransmits;    /* Number of unrecovered RTO timeouts.  */

        __u8    reordering;     /* Packet reordering metric.            */
        __u8    queue_shrunk;   /* Write queue has been shrunk recently.*/
        __u8    defer_accept;   /* User waits for some data after accept() */

/* RTT measurement */
        __u8    backoff;        /* backoff                              */
        __u32   srtt;           /* smothed round trip time << 3         */
        __u32   mdev;           /* medium deviation                     */
        __u32   mdev_max;       /* maximal mdev for the last rtt period */
        __u32   rttvar;         /* smoothed mdev_max                    */
        __u32   rtt_seq;        /* sequence number to update rttvar     */
        __u32   rto;            /* retransmit timeout                   */

        __u32   packets_out;    /* Packets which are "in flight"        */
        __u32   left_out;       /* Packets which leaved network         */
        __u32   retrans_out;    /* Retransmitted packets out            */


/*
 *      Slow start and congestion control (see also Nagle, and Karn & Partridge)
 */

        __u32   snd_ssthresh;   /* Slow start size threshold            */
        __u32   snd_cwnd;       /* Sending congestion window            */
        __u16   snd_cwnd_cnt;   /* Linear increase counter              */
        __u16   snd_cwnd_clamp; /* Do not allow snd_cwnd to grow above this */
        __u32   snd_cwnd_used;
        __u32   snd_cwnd_stamp;

        /* Two commonly used timers in both sender and receiver paths. */
        unsigned long           timeout;
        struct timer_list       retransmit_timer;       /* Resend (no ack)      */
        struct timer_list       delack_timer;           /* Ack delay            */

        struct sk_buff_head     out_of_order_queue; /* Out of order segments go here */

        struct tcp_func         *af_specific;   /* Operations which are AF_INET{4,6} specific   */
        struct sk_buff          *send_head;     /* Front of stuff to transmit                   */

        __u32   rcv_wnd;        /* Current receiver window              */
        __u32   rcv_wup;        /* rcv_nxt on last window update sent   */
        __u32   write_seq;      /* Tail(+1) of data held in tcp send buffer */
        __u32   pushed_seq;     /* Last pushed seq, required to talk to windows */
        __u32   copied_seq;     /* Head of yet unread data              */
/*
 *      Options received (usually on last packet, some only on SYN packets).
 */

        char    tstamp_ok,      /* TIMESTAMP seen on SYN packet         */
                wscale_ok,      /* Wscale seen on SYN packet            */
                sack_ok;        /* SACK seen on SYN packet              */
        char    saw_tstamp;     /* Saw TIMESTAMP on last packet         */
        __u8    snd_wscale;     /* Window scaling received from sender  */
        __u8    rcv_wscale;     /* Window scaling to send to receiver   */
        __u8    nonagle;        /* Disable Nagle algorithm?             */
        __u8    keepalive_probes; /* num of allowed keep alive probes   */

/*      PAWS/RTTM data  */
        __u32   rcv_tsval;      /* Time stamp value                     */
        __u32   rcv_tsecr;      /* Time stamp echo reply                */
        __u32   ts_recent;      /* Time stamp to echo next              */
        long    ts_recent_stamp;/* Time we stored ts_recent (for aging) */

/*      SACKs data      */
        __u16   user_mss;       /* mss requested by user in ioctl */
        __u8    dsack;          /* D-SACK is scheduled                  */
        __u8    eff_sacks;      /* Size of SACK array to send with next packet */
        struct tcp_sack_block duplicate_sack[1]; /* D-SACK block */
        struct tcp_sack_block selective_acks[4]; /* The SACKS themselves*/

        __u32   window_clamp;   /* Maximal window to advertise          */
        __u32   rcv_ssthresh;   /* Current window clamp                 */
        __u8    probes_out;     /* unanswered 0 window probes           */
        __u8    num_sacks;      /* Number of SACK blocks                */
        __u16   advmss;         /* Advertised MSS                       */

        __u8    syn_retries;    /* num of allowed syn retries */
        __u8    ecn_flags;      /* ECN status bits.                     */
        __u16   prior_ssthresh; /* ssthresh saved at recovery start     */
        __u32   lost_out;       /* Lost packets                         */
        __u32   sacked_out;     /* SACK'd packets                       */
        __u32   fackets_out;    /* FACK'd packets                       */
        __u32   high_seq;       /* snd_nxt at onset of congestion       */

        __u32   retrans_stamp;  /* Timestamp of the last retransmit,
                                 * also used in SYN-SENT to remember stamp of
                                 * the first SYN. */

        __u32   undo_marker;    /* tracking retrans started here. */
        int     undo_retrans;   /* number of undoable retransmissions. */
        __u32   urg_seq;        /* Seq of received urgent pointer */
        __u16   urg_data;       /* Saved octet of OOB data and control flags */
        __u8    pending;        /* Scheduled timer event        */
        __u8    urg_mode;       /* In urgent mode               */
        __u32   snd_up;         /* Urgent pointer               */

        /* The syn_wait_lock is necessary only to avoid tcp_get_info having
         * to grab the main lock sock while browsing the listening hash
         * (otherwise it's deadlock prone).
         * This lock is acquired in read mode only from tcp_get_info() and
         * it's acquired in write mode _only_ from code that is actively
         * changing the syn_wait_queue. All readers that are holding
         * the master sock lock don't need to grab this lock in read mode
         * too as the syn_wait_queue writes are always protected from
         * the main sock lock.
         */

        rwlock_t                syn_wait_lock;
        struct tcp_listen_opt   *listen_opt;

        /* FIFO of established children */
        struct open_request     *accept_queue;
        struct open_request     *accept_queue_tail;

        int                     write_pending;  /* A write to socket waits to start. */

        unsigned int            keepalive_time;   /* time before keep alive takes place */
        unsigned int            keepalive_intvl;  /* time interval between keep alive probes */
        int                     linger2;

        int                     frto_counter; /* Number of new acks after RTO */
        __u32                   frto_highmark; /* snd_nxt when RTO occurred */

        unsigned long last_synq_overflow;
};

/* WARNING: don't change the layout of the members in tcp_sock! */
struct tcp_sock {
        struct sock       sk;
#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
        struct ipv6_pinfo *pinet6;
#endif
        struct inet_opt   inet;
        struct tcp_opt    tcp;
};

#define tcp_sk(__sk) (&((struct tcp_sock *)__sk)->tcp)

#endif

#endif  /* _LINUX_TCP_H */