diff --git a/src/include/lwip/opt.h b/src/include/lwip/opt.h index d4078520..fcda2e3e 100644 --- a/src/include/lwip/opt.h +++ b/src/include/lwip/opt.h @@ -1759,8 +1759,12 @@ * VJ_SUPPORT==1: Support VJ header compression. */ #ifndef VJ_SUPPORT -#define VJ_SUPPORT 0 +#define VJ_SUPPORT 1 #endif +#if !PPPOS_SUPPORT +#undef VJ_SUPPORT +#define VJ_SUPPORT 0 /* Only PPPoS may need VJ compression */ +#endif /* !PPPOS_SUPPORT */ /** * PPP_MD5_RANDM==1: Use MD5 for better randomness. Automatically enabled if CHAP support is enabled. diff --git a/src/netif/ppp/ppp.c b/src/netif/ppp/ppp.c index c5c35cf0..af706987 100644 --- a/src/netif/ppp/ppp.c +++ b/src/netif/ppp/ppp.c @@ -89,6 +89,7 @@ #include "lwip/api.h" #include "lwip/snmp.h" #include "lwip/sio.h" +#include "lwip/sys.h" #include "lwip/ip.h" /* for ip_input() */ #include "ppp_impl.h" @@ -113,6 +114,9 @@ #if ECP_SUPPORT #include "ecp.h" #endif /* EAP_SUPPORT */ +#if VJ_SUPPORT +#include "vj.h" +#endif /* VJ_SUPPORT */ #if PPPOE_SUPPORT #include "netif/ppp_oe.h" diff --git a/src/netif/ppp/vj.c b/src/netif/ppp/vj.c new file mode 100644 index 00000000..0fed7aab --- /dev/null +++ b/src/netif/ppp/vj.c @@ -0,0 +1,647 @@ +/* + * Routines to compress and uncompess tcp packets (for transmission + * over low speed serial lines. + * + * Copyright (c) 1989 Regents of the University of California. + * All rights reserved. + * + * Redistribution and use in source and binary forms are permitted + * provided that the above copyright notice and this paragraph are + * duplicated in all such forms and that any documentation, + * advertising materials, and other materials related to such + * distribution and use acknowledge that the software was developed + * by the University of California, Berkeley. The name of the + * University may not be used to endorse or promote products derived + * from this software without specific prior written permission. + * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED + * WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE. + * + * Van Jacobson (van@helios.ee.lbl.gov), Dec 31, 1989: + * Initial distribution. + * + * Modified June 1993 by Paul Mackerras, paulus@cs.anu.edu.au, + * so that the entire packet being decompressed doesn't have + * to be in contiguous memory (just the compressed header). + * + * Modified March 1998 by Guy Lancaster, glanca@gesn.com, + * for a 16 bit processor. + */ + +#include "lwip/opt.h" +#if PPP_SUPPORT && VJ_SUPPORT /* don't build if not configured for use in lwipopts.h */ + +#include "ppp_impl.h" +#include "pppdebug.h" + +#include "vj.h" + +#include + +#if LINK_STATS +#define INCR(counter) ++comp->stats.counter +#else +#define INCR(counter) +#endif + +void +vj_compress_init(struct vjcompress *comp) +{ + register u_char i; + register struct cstate *tstate = comp->tstate; + +#if MAX_SLOTS == 0 + memset((char *)comp, 0, sizeof(*comp)); +#endif + comp->maxSlotIndex = MAX_SLOTS - 1; + comp->compressSlot = 0; /* Disable slot ID compression by default. */ + for (i = MAX_SLOTS - 1; i > 0; --i) { + tstate[i].cs_id = i; + tstate[i].cs_next = &tstate[i - 1]; + } + tstate[0].cs_next = &tstate[MAX_SLOTS - 1]; + tstate[0].cs_id = 0; + comp->last_cs = &tstate[0]; + comp->last_recv = 255; + comp->last_xmit = 255; + comp->flags = VJF_TOSS; +} + + +/* ENCODE encodes a number that is known to be non-zero. ENCODEZ + * checks for zero (since zero has to be encoded in the long, 3 byte + * form). + */ +#define ENCODE(n) { \ + if ((u_short)(n) >= 256) { \ + *cp++ = 0; \ + cp[1] = (u_char)(n); \ + cp[0] = (u_char)((n) >> 8); \ + cp += 2; \ + } else { \ + *cp++ = (u_char)(n); \ + } \ +} +#define ENCODEZ(n) { \ + if ((u_short)(n) >= 256 || (u_short)(n) == 0) { \ + *cp++ = 0; \ + cp[1] = (u_char)(n); \ + cp[0] = (u_char)((n) >> 8); \ + cp += 2; \ + } else { \ + *cp++ = (u_char)(n); \ + } \ +} + +#define DECODEL(f) { \ + if (*cp == 0) {\ + u32_t tmp = ntohl(f) + ((cp[1] << 8) | cp[2]); \ + (f) = htonl(tmp); \ + cp += 3; \ + } else { \ + u32_t tmp = ntohl(f) + (u32_t)*cp++; \ + (f) = htonl(tmp); \ + } \ +} + +#define DECODES(f) { \ + if (*cp == 0) {\ + u_short tmp = ntohs(f) + (((u_short)cp[1] << 8) | cp[2]); \ + (f) = htons(tmp); \ + cp += 3; \ + } else { \ + u_short tmp = ntohs(f) + (u_short)*cp++; \ + (f) = htons(tmp); \ + } \ +} + +#define DECODEU(f) { \ + if (*cp == 0) {\ + (f) = htons(((u_short)cp[1] << 8) | cp[2]); \ + cp += 3; \ + } else { \ + (f) = htons((u_short)*cp++); \ + } \ +} + +/* + * vj_compress_tcp - Attempt to do Van Jacobson header compression on a + * packet. This assumes that nb and comp are not null and that the first + * buffer of the chain contains a valid IP header. + * Return the VJ type code indicating whether or not the packet was + * compressed. + */ +u_int +vj_compress_tcp(struct vjcompress *comp, struct pbuf *pb) +{ + register struct ip_hdr *ip = (struct ip_hdr *)pb->payload; + register struct cstate *cs = comp->last_cs->cs_next; + register u_short hlen = IPH_HL(ip); + register struct tcp_hdr *oth; + register struct tcp_hdr *th; + register u_short deltaS, deltaA = 0; + register u_long deltaL; + register u_int changes = 0; + u_char new_seq[16]; + register u_char *cp = new_seq; + + /* + * Check that the packet is IP proto TCP. + */ + if (IPH_PROTO(ip) != IP_PROTO_TCP) { + return (TYPE_IP); + } + + /* + * Bail if this is an IP fragment or if the TCP packet isn't + * `compressible' (i.e., ACK isn't set or some other control bit is + * set). + */ + if ((IPH_OFFSET(ip) & PP_HTONS(0x3fff)) || pb->tot_len < 40) { + return (TYPE_IP); + } + th = (struct tcp_hdr *)&((long *)ip)[hlen]; + if ((TCPH_FLAGS(th) & (TCP_SYN|TCP_FIN|TCP_RST|TCP_ACK)) != TCP_ACK) { + return (TYPE_IP); + } + /* + * Packet is compressible -- we're going to send either a + * COMPRESSED_TCP or UNCOMPRESSED_TCP packet. Either way we need + * to locate (or create) the connection state. Special case the + * most recently used connection since it's most likely to be used + * again & we don't have to do any reordering if it's used. + */ + INCR(vjs_packets); + if (!ip_addr_cmp(&ip->src, &cs->cs_ip.src) + || !ip_addr_cmp(&ip->dest, &cs->cs_ip.dest) + || *(long *)th != ((long *)&cs->cs_ip)[IPH_HL(&cs->cs_ip)]) { + /* + * Wasn't the first -- search for it. + * + * States are kept in a circularly linked list with + * last_cs pointing to the end of the list. The + * list is kept in lru order by moving a state to the + * head of the list whenever it is referenced. Since + * the list is short and, empirically, the connection + * we want is almost always near the front, we locate + * states via linear search. If we don't find a state + * for the datagram, the oldest state is (re-)used. + */ + register struct cstate *lcs; + register struct cstate *lastcs = comp->last_cs; + + do { + lcs = cs; cs = cs->cs_next; + INCR(vjs_searches); + if (ip_addr_cmp(&ip->src, &cs->cs_ip.src) + && ip_addr_cmp(&ip->dest, &cs->cs_ip.dest) + && *(long *)th == ((long *)&cs->cs_ip)[IPH_HL(&cs->cs_ip)]) { + goto found; + } + } while (cs != lastcs); + + /* + * Didn't find it -- re-use oldest cstate. Send an + * uncompressed packet that tells the other side what + * connection number we're using for this conversation. + * Note that since the state list is circular, the oldest + * state points to the newest and we only need to set + * last_cs to update the lru linkage. + */ + INCR(vjs_misses); + comp->last_cs = lcs; + hlen += TCPH_HDRLEN(th); + hlen <<= 2; + /* Check that the IP/TCP headers are contained in the first buffer. */ + if (hlen > pb->len) { + return (TYPE_IP); + } + goto uncompressed; + + found: + /* + * Found it -- move to the front on the connection list. + */ + if (cs == lastcs) { + comp->last_cs = lcs; + } else { + lcs->cs_next = cs->cs_next; + cs->cs_next = lastcs->cs_next; + lastcs->cs_next = cs; + } + } + + oth = (struct tcp_hdr *)&((long *)&cs->cs_ip)[hlen]; + deltaS = hlen; + hlen += TCPH_HDRLEN(th); + hlen <<= 2; + /* Check that the IP/TCP headers are contained in the first buffer. */ + if (hlen > pb->len) { + PPPDEBUG(LOG_INFO, ("vj_compress_tcp: header len %d spans buffers\n", hlen)); + return (TYPE_IP); + } + + /* + * Make sure that only what we expect to change changed. The first + * line of the `if' checks the IP protocol version, header length & + * type of service. The 2nd line checks the "Don't fragment" bit. + * The 3rd line checks the time-to-live and protocol (the protocol + * check is unnecessary but costless). The 4th line checks the TCP + * header length. The 5th line checks IP options, if any. The 6th + * line checks TCP options, if any. If any of these things are + * different between the previous & current datagram, we send the + * current datagram `uncompressed'. + */ + if (((u_short *)ip)[0] != ((u_short *)&cs->cs_ip)[0] + || ((u_short *)ip)[3] != ((u_short *)&cs->cs_ip)[3] + || ((u_short *)ip)[4] != ((u_short *)&cs->cs_ip)[4] + || TCPH_HDRLEN(th) != TCPH_HDRLEN(oth) + || (deltaS > 5 && BCMP(ip + 1, &cs->cs_ip + 1, (deltaS - 5) << 2)) + || (TCPH_HDRLEN(th) > 5 && BCMP(th + 1, oth + 1, (TCPH_HDRLEN(th) - 5) << 2))) { + goto uncompressed; + } + + /* + * Figure out which of the changing fields changed. The + * receiver expects changes in the order: urgent, window, + * ack, seq (the order minimizes the number of temporaries + * needed in this section of code). + */ + if (TCPH_FLAGS(th) & TCP_URG) { + deltaS = ntohs(th->urgp); + ENCODEZ(deltaS); + changes |= NEW_U; + } else if (th->urgp != oth->urgp) { + /* argh! URG not set but urp changed -- a sensible + * implementation should never do this but RFC793 + * doesn't prohibit the change so we have to deal + * with it. */ + goto uncompressed; + } + + if ((deltaS = (u_short)(ntohs(th->wnd) - ntohs(oth->wnd))) != 0) { + ENCODE(deltaS); + changes |= NEW_W; + } + + if ((deltaL = ntohl(th->ackno) - ntohl(oth->ackno)) != 0) { + if (deltaL > 0xffff) { + goto uncompressed; + } + deltaA = (u_short)deltaL; + ENCODE(deltaA); + changes |= NEW_A; + } + + if ((deltaL = ntohl(th->seqno) - ntohl(oth->seqno)) != 0) { + if (deltaL > 0xffff) { + goto uncompressed; + } + deltaS = (u_short)deltaL; + ENCODE(deltaS); + changes |= NEW_S; + } + + switch(changes) { + case 0: + /* + * Nothing changed. If this packet contains data and the + * last one didn't, this is probably a data packet following + * an ack (normal on an interactive connection) and we send + * it compressed. Otherwise it's probably a retransmit, + * retransmitted ack or window probe. Send it uncompressed + * in case the other side missed the compressed version. + */ + if (IPH_LEN(ip) != IPH_LEN(&cs->cs_ip) && + ntohs(IPH_LEN(&cs->cs_ip)) == hlen) { + break; + } + + /* (fall through) */ + + case SPECIAL_I: + case SPECIAL_D: + /* + * actual changes match one of our special case encodings -- + * send packet uncompressed. + */ + goto uncompressed; + + case NEW_S|NEW_A: + if (deltaS == deltaA && deltaS == ntohs(IPH_LEN(&cs->cs_ip)) - hlen) { + /* special case for echoed terminal traffic */ + changes = SPECIAL_I; + cp = new_seq; + } + break; + + case NEW_S: + if (deltaS == ntohs(IPH_LEN(&cs->cs_ip)) - hlen) { + /* special case for data xfer */ + changes = SPECIAL_D; + cp = new_seq; + } + break; + } + + deltaS = (u_short)(ntohs(IPH_ID(ip)) - ntohs(IPH_ID(&cs->cs_ip))); + if (deltaS != 1) { + ENCODEZ(deltaS); + changes |= NEW_I; + } + if (TCPH_FLAGS(th) & TCP_PSH) { + changes |= TCP_PUSH_BIT; + } + /* + * Grab the cksum before we overwrite it below. Then update our + * state with this packet's header. + */ + deltaA = ntohs(th->chksum); + MEMCPY(&cs->cs_ip, ip, hlen); + + /* + * We want to use the original packet as our compressed packet. + * (cp - new_seq) is the number of bytes we need for compressed + * sequence numbers. In addition we need one byte for the change + * mask, one for the connection id and two for the tcp checksum. + * So, (cp - new_seq) + 4 bytes of header are needed. hlen is how + * many bytes of the original packet to toss so subtract the two to + * get the new packet size. + */ + deltaS = (u_short)(cp - new_seq); + if (!comp->compressSlot || comp->last_xmit != cs->cs_id) { + comp->last_xmit = cs->cs_id; + hlen -= deltaS + 4; + if(pbuf_header(pb, -hlen)){ + /* Can we cope with this failing? Just assert for now */ + LWIP_ASSERT("pbuf_header failed\n", 0); + } + cp = (u_char *)pb->payload; + *cp++ = (u_char)(changes | NEW_C); + *cp++ = cs->cs_id; + } else { + hlen -= deltaS + 3; + if(pbuf_header(pb, -hlen)) { + /* Can we cope with this failing? Just assert for now */ + LWIP_ASSERT("pbuf_header failed\n", 0); + } + cp = (u_char *)pb->payload; + *cp++ = (u_char)changes; + } + *cp++ = (u_char)(deltaA >> 8); + *cp++ = (u_char)deltaA; + MEMCPY(cp, new_seq, deltaS); + INCR(vjs_compressed); + return (TYPE_COMPRESSED_TCP); + + /* + * Update connection state cs & send uncompressed packet (that is, + * a regular ip/tcp packet but with the 'conversation id' we hope + * to use on future compressed packets in the protocol field). + */ +uncompressed: + MEMCPY(&cs->cs_ip, ip, hlen); + IPH_PROTO_SET(ip, cs->cs_id); + comp->last_xmit = cs->cs_id; + return (TYPE_UNCOMPRESSED_TCP); +} + +/* + * Called when we may have missed a packet. + */ +void +vj_uncompress_err(struct vjcompress *comp) +{ + comp->flags |= VJF_TOSS; + INCR(vjs_errorin); +} + +/* + * "Uncompress" a packet of type TYPE_UNCOMPRESSED_TCP. + * Return 0 on success, -1 on failure. + */ +int +vj_uncompress_uncomp(struct pbuf *nb, struct vjcompress *comp) +{ + register u_int hlen; + register struct cstate *cs; + register struct ip_hdr *ip; + + ip = (struct ip_hdr *)nb->payload; + hlen = IPH_HL(ip) << 2; + if (IPH_PROTO(ip) >= MAX_SLOTS + || hlen + sizeof(struct tcp_hdr) > nb->len + || (hlen += TCPH_HDRLEN(((struct tcp_hdr *)&((char *)ip)[hlen])) << 2) + > nb->len + || hlen > MAX_HDR) { + PPPDEBUG(LOG_INFO, ("vj_uncompress_uncomp: bad cid=%d, hlen=%d buflen=%d\n", + IPH_PROTO(ip), hlen, nb->len)); + comp->flags |= VJF_TOSS; + INCR(vjs_errorin); + return -1; + } + cs = &comp->rstate[comp->last_recv = IPH_PROTO(ip)]; + comp->flags &=~ VJF_TOSS; + IPH_PROTO_SET(ip, IP_PROTO_TCP); + MEMCPY(&cs->cs_ip, ip, hlen); + cs->cs_hlen = (u_short)hlen; + INCR(vjs_uncompressedin); + return 0; +} + +/* + * Uncompress a packet of type TYPE_COMPRESSED_TCP. + * The packet is composed of a buffer chain and the first buffer + * must contain an accurate chain length. + * The first buffer must include the entire compressed TCP/IP header. + * This procedure replaces the compressed header with the uncompressed + * header and returns the length of the VJ header. + */ +int +vj_uncompress_tcp(struct pbuf **nb, struct vjcompress *comp) +{ + u_char *cp; + struct tcp_hdr *th; + struct cstate *cs; + u_short *bp; + struct pbuf *n0 = *nb; + u32_t tmp; + u_int vjlen, hlen, changes; + + INCR(vjs_compressedin); + cp = (u_char *)n0->payload; + changes = *cp++; + if (changes & NEW_C) { + /* + * Make sure the state index is in range, then grab the state. + * If we have a good state index, clear the 'discard' flag. + */ + if (*cp >= MAX_SLOTS) { + PPPDEBUG(LOG_INFO, ("vj_uncompress_tcp: bad cid=%d\n", *cp)); + goto bad; + } + + comp->flags &=~ VJF_TOSS; + comp->last_recv = *cp++; + } else { + /* + * this packet has an implicit state index. If we've + * had a line error since the last time we got an + * explicit state index, we have to toss the packet. + */ + if (comp->flags & VJF_TOSS) { + PPPDEBUG(LOG_INFO, ("vj_uncompress_tcp: tossing\n")); + INCR(vjs_tossed); + return (-1); + } + } + cs = &comp->rstate[comp->last_recv]; + hlen = IPH_HL(&cs->cs_ip) << 2; + th = (struct tcp_hdr *)&((u_char *)&cs->cs_ip)[hlen]; + th->chksum = htons((*cp << 8) | cp[1]); + cp += 2; + if (changes & TCP_PUSH_BIT) { + TCPH_SET_FLAG(th, TCP_PSH); + } else { + TCPH_UNSET_FLAG(th, TCP_PSH); + } + + switch (changes & SPECIALS_MASK) { + case SPECIAL_I: + { + register u32_t i = ntohs(IPH_LEN(&cs->cs_ip)) - cs->cs_hlen; + /* some compilers can't nest inline assembler.. */ + tmp = ntohl(th->ackno) + i; + th->ackno = htonl(tmp); + tmp = ntohl(th->seqno) + i; + th->seqno = htonl(tmp); + } + break; + + case SPECIAL_D: + /* some compilers can't nest inline assembler.. */ + tmp = ntohl(th->seqno) + ntohs(IPH_LEN(&cs->cs_ip)) - cs->cs_hlen; + th->seqno = htonl(tmp); + break; + + default: + if (changes & NEW_U) { + TCPH_SET_FLAG(th, TCP_URG); + DECODEU(th->urgp); + } else { + TCPH_UNSET_FLAG(th, TCP_URG); + } + if (changes & NEW_W) { + DECODES(th->wnd); + } + if (changes & NEW_A) { + DECODEL(th->ackno); + } + if (changes & NEW_S) { + DECODEL(th->seqno); + } + break; + } + if (changes & NEW_I) { + DECODES(cs->cs_ip._id); + } else { + IPH_ID_SET(&cs->cs_ip, ntohs(IPH_ID(&cs->cs_ip)) + 1); + IPH_ID_SET(&cs->cs_ip, htons(IPH_ID(&cs->cs_ip))); + } + + /* + * At this point, cp points to the first byte of data in the + * packet. Fill in the IP total length and update the IP + * header checksum. + */ + vjlen = (u_short)(cp - (u_char*)n0->payload); + if (n0->len < vjlen) { + /* + * We must have dropped some characters (crc should detect + * this but the old slip framing won't) + */ + PPPDEBUG(LOG_INFO, ("vj_uncompress_tcp: head buffer %d too short %d\n", + n0->len, vjlen)); + goto bad; + } + +#if BYTE_ORDER == LITTLE_ENDIAN + tmp = n0->tot_len - vjlen + cs->cs_hlen; + IPH_LEN_SET(&cs->cs_ip, htons((u_short)tmp)); +#else + IPH_LEN_SET(&cs->cs_ip, htons(n0->tot_len - vjlen + cs->cs_hlen)); +#endif + + /* recompute the ip header checksum */ + bp = (u_short *) &cs->cs_ip; + IPH_CHKSUM_SET(&cs->cs_ip, 0); + for (tmp = 0; hlen > 0; hlen -= 2) { + tmp += *bp++; + } + tmp = (tmp & 0xffff) + (tmp >> 16); + tmp = (tmp & 0xffff) + (tmp >> 16); + IPH_CHKSUM_SET(&cs->cs_ip, (u_short)(~tmp)); + + /* Remove the compressed header and prepend the uncompressed header. */ + if(pbuf_header(n0, -((s16_t)(vjlen)))) { + /* Can we cope with this failing? Just assert for now */ + LWIP_ASSERT("pbuf_header failed\n", 0); + goto bad; + } + + if(LWIP_MEM_ALIGN(n0->payload) != n0->payload) { + struct pbuf *np, *q; + u8_t *bufptr; + + np = pbuf_alloc(PBUF_RAW, n0->len + cs->cs_hlen, PBUF_POOL); + if(!np) { + PPPDEBUG(LOG_WARNING, ("vj_uncompress_tcp: realign failed\n")); + goto bad; + } + + if(pbuf_header(np, -cs->cs_hlen)) { + /* Can we cope with this failing? Just assert for now */ + LWIP_ASSERT("pbuf_header failed\n", 0); + goto bad; + } + + bufptr = n0->payload; + for(q = np; q != NULL; q = q->next) { + MEMCPY(q->payload, bufptr, q->len); + bufptr += q->len; + } + + if(n0->next) { + pbuf_chain(np, n0->next); + pbuf_dechain(n0); + } + pbuf_free(n0); + n0 = np; + } + + if(pbuf_header(n0, cs->cs_hlen)) { + struct pbuf *np; + + LWIP_ASSERT("vj_uncompress_tcp: cs->cs_hlen <= PBUF_POOL_BUFSIZE", cs->cs_hlen <= PBUF_POOL_BUFSIZE); + np = pbuf_alloc(PBUF_RAW, cs->cs_hlen, PBUF_POOL); + if(!np) { + PPPDEBUG(LOG_WARNING, ("vj_uncompress_tcp: prepend failed\n")); + goto bad; + } + pbuf_cat(np, n0); + n0 = np; + } + LWIP_ASSERT("n0->len >= cs->cs_hlen", n0->len >= cs->cs_hlen); + MEMCPY(n0->payload, &cs->cs_ip, cs->cs_hlen); + + *nb = n0; + + return vjlen; + +bad: + comp->flags |= VJF_TOSS; + INCR(vjs_errorin); + return (-1); +} + +#endif /* PPP_SUPPORT && VJ_SUPPORT */ diff --git a/src/netif/ppp/vj.h b/src/netif/ppp/vj.h new file mode 100644 index 00000000..3780667b --- /dev/null +++ b/src/netif/ppp/vj.h @@ -0,0 +1,161 @@ +/* + * Definitions for tcp compression routines. + * + * $Id: vj.h,v 1.7 2010/02/22 17:52:09 goldsimon Exp $ + * + * Copyright (c) 1989 Regents of the University of California. + * All rights reserved. + * + * Redistribution and use in source and binary forms are permitted + * provided that the above copyright notice and this paragraph are + * duplicated in all such forms and that any documentation, + * advertising materials, and other materials related to such + * distribution and use acknowledge that the software was developed + * by the University of California, Berkeley. The name of the + * University may not be used to endorse or promote products derived + * from this software without specific prior written permission. + * THIS SOFTWARE IS PROVIDED ``AS IS'' AND WITHOUT ANY EXPRESS OR + * IMPLIED WARRANTIES, INCLUDING, WITHOUT LIMITATION, THE IMPLIED + * WARRANTIES OF MERCHANTIBILITY AND FITNESS FOR A PARTICULAR PURPOSE. + * + * Van Jacobson (van@helios.ee.lbl.gov), Dec 31, 1989: + * - Initial distribution. + */ + +#include "lwip/opt.h" +#if PPP_SUPPORT && VJ_SUPPORT /* don't build if not configured for use in lwipopts.h */ + +#ifndef VJ_H +#define VJ_H + +#include "lwip/ip.h" +#include "lwip/tcp_impl.h" + +#define MAX_SLOTS 16 /* must be > 2 and < 256 */ +#define MAX_HDR 128 + +/* + * Compressed packet format: + * + * The first octet contains the packet type (top 3 bits), TCP + * 'push' bit, and flags that indicate which of the 4 TCP sequence + * numbers have changed (bottom 5 bits). The next octet is a + * conversation number that associates a saved IP/TCP header with + * the compressed packet. The next two octets are the TCP checksum + * from the original datagram. The next 0 to 15 octets are + * sequence number changes, one change per bit set in the header + * (there may be no changes and there are two special cases where + * the receiver implicitly knows what changed -- see below). + * + * There are 5 numbers which can change (they are always inserted + * in the following order): TCP urgent pointer, window, + * acknowlegement, sequence number and IP ID. (The urgent pointer + * is different from the others in that its value is sent, not the + * change in value.) Since typical use of SLIP links is biased + * toward small packets (see comments on MTU/MSS below), changes + * use a variable length coding with one octet for numbers in the + * range 1 - 255 and 3 octets (0, MSB, LSB) for numbers in the + * range 256 - 65535 or 0. (If the change in sequence number or + * ack is more than 65535, an uncompressed packet is sent.) + */ + +/* + * Packet types (must not conflict with IP protocol version) + * + * The top nibble of the first octet is the packet type. There are + * three possible types: IP (not proto TCP or tcp with one of the + * control flags set); uncompressed TCP (a normal IP/TCP packet but + * with the 8-bit protocol field replaced by an 8-bit connection id -- + * this type of packet syncs the sender & receiver); and compressed + * TCP (described above). + * + * LSB of 4-bit field is TCP "PUSH" bit (a worthless anachronism) and + * is logically part of the 4-bit "changes" field that follows. Top + * three bits are actual packet type. For backward compatibility + * and in the interest of conserving bits, numbers are chosen so the + * IP protocol version number (4) which normally appears in this nibble + * means "IP packet". + */ + +/* packet types */ +#define TYPE_IP 0x40 +#define TYPE_UNCOMPRESSED_TCP 0x70 +#define TYPE_COMPRESSED_TCP 0x80 +#define TYPE_ERROR 0x00 + +/* Bits in first octet of compressed packet */ +#define NEW_C 0x40 /* flag bits for what changed in a packet */ +#define NEW_I 0x20 +#define NEW_S 0x08 +#define NEW_A 0x04 +#define NEW_W 0x02 +#define NEW_U 0x01 + +/* reserved, special-case values of above */ +#define SPECIAL_I (NEW_S|NEW_W|NEW_U) /* echoed interactive traffic */ +#define SPECIAL_D (NEW_S|NEW_A|NEW_W|NEW_U) /* unidirectional data */ +#define SPECIALS_MASK (NEW_S|NEW_A|NEW_W|NEW_U) + +#define TCP_PUSH_BIT 0x10 + + +/* + * "state" data for each active tcp conversation on the wire. This is + * basically a copy of the entire IP/TCP header from the last packet + * we saw from the conversation together with a small identifier + * the transmit & receive ends of the line use to locate saved header. + */ +struct cstate { + struct cstate *cs_next; /* next most recently used state (xmit only) */ + u_short cs_hlen; /* size of hdr (receive only) */ + u_char cs_id; /* connection # associated with this state */ + u_char cs_filler; + union { + char csu_hdr[MAX_HDR]; + struct ip_hdr csu_ip; /* ip/tcp hdr from most recent packet */ + } vjcs_u; +}; +#define cs_ip vjcs_u.csu_ip +#define cs_hdr vjcs_u.csu_hdr + + +struct vjstat { + unsigned long vjs_packets; /* outbound packets */ + unsigned long vjs_compressed; /* outbound compressed packets */ + unsigned long vjs_searches; /* searches for connection state */ + unsigned long vjs_misses; /* times couldn't find conn. state */ + unsigned long vjs_uncompressedin; /* inbound uncompressed packets */ + unsigned long vjs_compressedin; /* inbound compressed packets */ + unsigned long vjs_errorin; /* inbound unknown type packets */ + unsigned long vjs_tossed; /* inbound packets tossed because of error */ +}; + +/* + * all the state data for one serial line (we need one of these per line). + */ +struct vjcompress { + struct cstate *last_cs; /* most recently used tstate */ + u_char last_recv; /* last rcvd conn. id */ + u_char last_xmit; /* last sent conn. id */ + u_short flags; + u_char maxSlotIndex; + u_char compressSlot; /* Flag indicating OK to compress slot ID. */ +#if LINK_STATS + struct vjstat stats; +#endif + struct cstate tstate[MAX_SLOTS]; /* xmit connection states */ + struct cstate rstate[MAX_SLOTS]; /* receive connection states */ +}; + +/* flag values */ +#define VJF_TOSS 1U /* tossing rcvd frames because of input err */ + +extern void vj_compress_init (struct vjcompress *comp); +extern u_int vj_compress_tcp (struct vjcompress *comp, struct pbuf *pb); +extern void vj_uncompress_err (struct vjcompress *comp); +extern int vj_uncompress_uncomp(struct pbuf *nb, struct vjcompress *comp); +extern int vj_uncompress_tcp (struct pbuf **nb, struct vjcompress *comp); + +#endif /* VJ_H */ + +#endif /* PPP_SUPPORT && VJ_SUPPORT */