freebsd-dev/usr.sbin/ppp/link.c
Pedro F. Giffuni 1de7b4b805 various: general adoption of SPDX licensing ID tags.
Mainly focus on files that use BSD 2-Clause license, however the tool I
was using misidentified many licenses so this was mostly a manual - error
prone - task.

The Software Package Data Exchange (SPDX) group provides a specification
to make it easier for automated tools to detect and summarize well known
opensource licenses. We are gradually adopting the specification, noting
that the tags are considered only advisory and do not, in any way,
superceed or replace the license texts.

No functional change intended.
2017-11-27 15:37:16 +00:00

415 lines
10 KiB
C

/*-
* SPDX-License-Identifier: BSD-2-Clause-FreeBSD
*
* Copyright (c) 1998 Brian Somers <brian@Awfulhak.org>
* All rights reserved.
*
* Redistribution and use in source and binary forms, with or without
* modification, are permitted provided that the following conditions
* are met:
* 1. Redistributions of source code must retain the above copyright
* notice, this list of conditions and the following disclaimer.
* 2. Redistributions in binary form must reproduce the above copyright
* notice, this list of conditions and the following disclaimer in the
* documentation and/or other materials provided with the distribution.
*
* THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND
* ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
* IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
* ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE
* FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
* DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
* OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
* HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
* LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
* OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
* SUCH DAMAGE.
*
* $FreeBSD$
*
*/
#include <sys/types.h>
#include <netinet/in_systm.h>
#include <sys/socket.h>
#include <sys/un.h>
#include <netinet/in.h>
#include <netinet/ip.h>
#include <stdarg.h>
#include <stdio.h>
#include <string.h>
#include <termios.h>
#include "defs.h"
#include "layer.h"
#include "mbuf.h"
#include "log.h"
#include "timer.h"
#include "lqr.h"
#include "hdlc.h"
#include "throughput.h"
#include "proto.h"
#include "fsm.h"
#include "descriptor.h"
#include "lcp.h"
#include "ccp.h"
#include "link.h"
#include "prompt.h"
#include "async.h"
#include "physical.h"
#include "mp.h"
#include "iplist.h"
#include "slcompress.h"
#include "ncpaddr.h"
#include "ip.h"
#include "ipcp.h"
#include "ipv6cp.h"
#include "auth.h"
#include "pap.h"
#include "chap.h"
#include "cbcp.h"
#include "command.h"
static void Despatch(struct bundle *, struct link *, struct mbuf *, u_short);
static inline void
link_AddInOctets(struct link *l, int n)
{
if (l->stats.gather) {
throughput_addin(&l->stats.total, n);
if (l->stats.parent)
throughput_addin(l->stats.parent, n);
}
}
static inline void
link_AddOutOctets(struct link *l, int n)
{
if (l->stats.gather) {
throughput_addout(&l->stats.total, n);
if (l->stats.parent)
throughput_addout(l->stats.parent, n);
}
}
void
link_SequenceQueue(struct link *l)
{
struct mqueue *queue, *highest;
log_Printf(LogDEBUG, "link_SequenceQueue\n");
highest = LINK_HIGHQ(l);
for (queue = l->Queue; queue < highest; queue++)
while (queue->len)
m_enqueue(highest, m_dequeue(queue));
}
void
link_DeleteQueue(struct link *l)
{
struct mqueue *queue, *highest;
highest = LINK_HIGHQ(l);
for (queue = l->Queue; queue <= highest; queue++)
while (queue->top)
m_freem(m_dequeue(queue));
}
size_t
link_QueueLen(struct link *l)
{
unsigned i;
size_t len;
for (i = 0, len = 0; i < LINK_QUEUES(l); i++)
len += l->Queue[i].len;
return len;
}
size_t
link_QueueBytes(struct link *l)
{
unsigned i;
size_t len, bytes;
struct mbuf *m;
bytes = 0;
for (i = 0, len = 0; i < LINK_QUEUES(l); i++) {
len = l->Queue[i].len;
m = l->Queue[i].top;
while (len--) {
bytes += m_length(m);
m = m->m_nextpkt;
}
}
return bytes;
}
void
link_PendingLowPriorityData(struct link *l, size_t *pkts, size_t *octets)
{
struct mqueue *queue, *highest;
struct mbuf *m;
size_t len;
/*
* This is all rfc1989 stuff... because our LQR packet is going to bypass
* everything that's not in the highest priority queue, we must be able to
* subtract that data from our outgoing packet/octet counts. However,
* we've already async-encoded our data at this point, but the async
* encodings MUSTn't be a part of the LQR-reported payload :( So, we have
* the async layer record how much it's padded the packet in the mbuf's
* priv field, and when we calculate our outgoing LQR values we subtract
* this value for each packet from the octet count sent.
*/
highest = LINK_HIGHQ(l);
*pkts = *octets = 0;
for (queue = l->Queue; queue < highest; queue++) {
len = queue->len;
*pkts += len;
for (m = queue->top; len--; m = m->m_nextpkt)
*octets += m_length(m) - m->priv;
}
}
struct mbuf *
link_Dequeue(struct link *l)
{
int pri;
struct mbuf *bp;
for (bp = NULL, pri = LINK_QUEUES(l) - 1; pri >= 0; pri--)
if (l->Queue[pri].len) {
bp = m_dequeue(l->Queue + pri);
log_Printf(LogDEBUG, "link_Dequeue: Dequeued from queue %d,"
" containing %lu more packets\n", pri,
(u_long)l->Queue[pri].len);
break;
}
return bp;
}
static struct protostatheader {
u_short number;
const char *name;
} ProtocolStat[NPROTOSTAT] = {
{ PROTO_IP, "IP" },
{ PROTO_VJUNCOMP, "VJ_UNCOMP" },
{ PROTO_VJCOMP, "VJ_COMP" },
{ PROTO_COMPD, "COMPD" },
{ PROTO_ICOMPD, "ICOMPD" },
{ PROTO_LCP, "LCP" },
{ PROTO_IPCP, "IPCP" },
{ PROTO_CCP, "CCP" },
{ PROTO_PAP, "PAP" },
{ PROTO_LQR, "LQR" },
{ PROTO_CHAP, "CHAP" },
{ PROTO_MP, "MULTILINK" },
{ 0, "Others" } /* must be last */
};
void
link_ProtocolRecord(struct link *l, u_short proto, int type)
{
int i;
for (i = 0; i < NPROTOSTAT; i++)
if (ProtocolStat[i].number == proto || ProtocolStat[i].number == 0) {
if (type == PROTO_IN)
l->proto_in[i]++;
else
l->proto_out[i]++;
break;
}
}
void
link_ReportProtocolStatus(struct link *l, struct prompt *prompt)
{
int i;
prompt_Printf(prompt, " Protocol in out "
"Protocol in out\n");
for (i = 0; i < NPROTOSTAT; i++) {
prompt_Printf(prompt, " %-9s: %8lu, %8lu",
ProtocolStat[i].name, l->proto_in[i], l->proto_out[i]);
if ((i % 2) == 0)
prompt_Printf(prompt, "\n");
}
if (!(i % 2))
prompt_Printf(prompt, "\n");
}
void
link_PushPacket(struct link *l, struct mbuf *bp, struct bundle *b, int pri,
u_short proto)
{
int layer;
/*
* When we ``push'' a packet into the link, it gets processed by the
* ``push'' function in each layer starting at the top.
* We never expect the result of a ``push'' to be more than one
* packet (as we do with ``pull''s).
*/
if(pri < 0 || (unsigned)pri >= LINK_QUEUES(l))
pri = 0;
bp->priv = 0; /* Adjusted by the async layer ! */
for (layer = l->nlayers; layer && bp; layer--)
if (l->layer[layer - 1]->push != NULL)
bp = (*l->layer[layer - 1]->push)(b, l, bp, pri, &proto);
if (bp) {
link_AddOutOctets(l, m_length(bp));
log_Printf(LogDEBUG, "link_PushPacket: Transmit proto 0x%04x\n", proto);
m_enqueue(l->Queue + pri, m_pullup(bp));
}
}
void
link_PullPacket(struct link *l, char *buf, size_t len, struct bundle *b)
{
struct mbuf *bp, *lbp[LAYER_MAX], *next;
u_short lproto[LAYER_MAX], proto;
int layer;
/*
* When we ``pull'' a packet from the link, it gets processed by the
* ``pull'' function in each layer starting at the bottom.
* Each ``pull'' may produce multiple packets, chained together using
* bp->m_nextpkt.
* Each packet that results from each pull has to be pulled through
* all of the higher layers before the next resulting packet is pulled
* through anything; this ensures that packets that depend on the
* fsm state resulting from the receipt of the previous packet aren't
* surprised.
*/
link_AddInOctets(l, len);
memset(lbp, '\0', sizeof lbp);
lbp[0] = m_get(len, MB_UNKNOWN);
memcpy(MBUF_CTOP(lbp[0]), buf, len);
lproto[0] = 0;
layer = 0;
while (layer || lbp[layer]) {
if (lbp[layer] == NULL) {
layer--;
continue;
}
bp = lbp[layer];
lbp[layer] = bp->m_nextpkt;
bp->m_nextpkt = NULL;
proto = lproto[layer];
if (l->layer[layer]->pull != NULL)
bp = (*l->layer[layer]->pull)(b, l, bp, &proto);
if (layer == l->nlayers - 1) {
/* We've just done the top layer, despatch the packet(s) */
while (bp) {
next = bp->m_nextpkt;
bp->m_nextpkt = NULL;
log_Printf(LogDEBUG, "link_PullPacket: Despatch proto 0x%04x\n", proto);
Despatch(b, l, bp, proto);
bp = next;
}
} else {
lbp[++layer] = bp;
lproto[layer] = proto;
}
}
}
int
link_Stack(struct link *l, struct layer *layer)
{
if (l->nlayers == sizeof l->layer / sizeof l->layer[0]) {
log_Printf(LogERROR, "%s: Oops, cannot stack a %s layer...\n",
l->name, layer->name);
return 0;
}
l->layer[l->nlayers++] = layer;
return 1;
}
void
link_EmptyStack(struct link *l)
{
l->nlayers = 0;
}
static const struct {
u_short proto;
struct mbuf *(*fn)(struct bundle *, struct link *, struct mbuf *);
} despatcher[] = {
{ PROTO_IP, ipv4_Input },
#ifndef NOINET6
{ PROTO_IPV6, ipv6_Input },
#endif
{ PROTO_MP, mp_Input },
{ PROTO_LCP, lcp_Input },
{ PROTO_IPCP, ipcp_Input },
#ifndef NOINET6
{ PROTO_IPV6CP, ipv6cp_Input },
#endif
{ PROTO_PAP, pap_Input },
{ PROTO_CHAP, chap_Input },
{ PROTO_CCP, ccp_Input },
{ PROTO_LQR, lqr_Input },
{ PROTO_CBCP, cbcp_Input }
};
#define DSIZE (sizeof despatcher / sizeof despatcher[0])
static void
Despatch(struct bundle *bundle, struct link *l, struct mbuf *bp, u_short proto)
{
unsigned f;
for (f = 0; f < DSIZE; f++)
if (despatcher[f].proto == proto) {
bp = (*despatcher[f].fn)(bundle, l, bp);
break;
}
if (bp) {
struct physical *p = link2physical(l);
log_Printf(LogPHASE, "%s protocol 0x%04x (%s)\n",
f == DSIZE ? "Unknown" : "Unexpected", proto,
hdlc_Protocol2Nam(proto));
bp = m_pullup(proto_Prepend(bp, proto, 0, 0));
lcp_SendProtoRej(&l->lcp, MBUF_CTOP(bp), bp->m_len);
if (p) {
p->hdlc.lqm.ifInDiscards++;
p->hdlc.stats.unknownproto++;
}
m_freem(bp);
}
}
int
link_ShowLayers(struct cmdargs const *arg)
{
struct link *l = command_ChooseLink(arg);
int layer;
for (layer = l->nlayers; layer; layer--)
prompt_Printf(arg->prompt, "%s%s", layer == l->nlayers ? "" : ", ",
l->layer[layer - 1]->name);
if (l->nlayers)
prompt_Printf(arg->prompt, "\n");
return 0;
}