io.c 45.2 KB
Newer Older
1 2 3
/*
 *	BIRD Internet Routing Daemon -- Unix I/O
 *
4
 *	(c) 1998--2004 Martin Mares <mj@ucw.cz>
5
 *      (c) 2004       Ondrej Filip <feela@network.cz>
6 7 8 9
 *
 *	Can be freely distributed and used under the terms of the GNU GPL.
 */

10 11 12 13
/* Unfortunately, some glibc versions hide parts of RFC 3542 API
   if _GNU_SOURCE is not defined. */
#define _GNU_SOURCE 1

14 15
#include <stdio.h>
#include <stdlib.h>
16
#include <time.h>
17 18 19
#include <sys/time.h>
#include <sys/types.h>
#include <sys/socket.h>
20
#include <sys/uio.h>
21
#include <sys/un.h>
22
#include <poll.h>
23
#include <unistd.h>
24
#include <fcntl.h>
25
#include <errno.h>
26
#include <net/if.h>
Ondřej Zajíček's avatar
Ondřej Zajíček committed
27
#include <netinet/in.h>
28 29
#include <netinet/tcp.h>
#include <netinet/udp.h>
30
#include <netinet/icmp6.h>
31 32 33 34 35 36

#include "nest/bird.h"
#include "lib/lists.h"
#include "lib/resource.h"
#include "lib/timer.h"
#include "lib/socket.h"
37
#include "lib/event.h"
38
#include "lib/string.h"
39 40 41
#include "nest/iface.h"

#include "lib/unix.h"
42
#include "lib/sysio.h"
43

44
/* Maximum number of calls of tx handler for one socket in one
45
 * poll iteration. Should be small enough to not monopolize CPU by
46 47 48 49
 * one protocol instance.
 */
#define MAX_STEPS 4

50
/* Maximum number of calls of rx handler for all sockets in one poll
51 52 53 54
   iteration. RX callbacks are often much more costly so we limit
   this to gen small latencies */
#define MAX_RX_STEPS 4

55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83
/*
 *	Tracked Files
 */

struct rfile {
  resource r;
  FILE *f;
};

static void
rf_free(resource *r)
{
  struct rfile *a = (struct rfile *) r;

  fclose(a->f);
}

static void
rf_dump(resource *r)
{
  struct rfile *a = (struct rfile *) r;

  debug("(FILE *%p)\n", a->f);
}

static struct resclass rf_class = {
  "FILE",
  sizeof(struct rfile),
  rf_free,
84
  rf_dump,
85
  NULL,
86
  NULL
87 88 89
};

void *
90
tracked_fopen(pool *p, char *name, char *mode)
91 92 93 94 95 96 97 98 99 100 101
{
  FILE *f = fopen(name, mode);

  if (f)
    {
      struct rfile *r = ralloc(p, &rf_class);
      r->f = f;
    }
  return f;
}

102 103 104 105 106
/**
 * DOC: Timers
 *
 * Timers are resources which represent a wish of a module to call
 * a function at the specified time. The platform dependent code
Martin Mareš's avatar
Martin Mareš committed
107
 * doesn't guarantee exact timing, only that a timer function
108 109
 * won't be called before the requested time.
 *
110 111 112 113 114
 * In BIRD, time is represented by values of the &bird_clock_t type
 * which are integral numbers interpreted as a relative number of seconds since
 * some fixed time point in past. The current time can be read
 * from variable @now with reasonable accuracy and is monotonic. There is also
 * a current 'absolute' time in variable @now_real reported by OS.
115 116 117 118 119
 *
 * Each timer is described by a &timer structure containing a pointer
 * to the handler function (@hook), data private to this function (@data),
 * time the function should be called at (@expires, 0 for inactive timers),
 * for the other fields see |timer.h|.
120 121 122 123 124 125 126
 */

#define NEAR_TIMER_LIMIT 4

static list near_timers, far_timers;
static bird_clock_t first_far_timer = TIME_INFINITY;

127
/* now must be different from 0, because 0 is a special value in timer->expires */
128
bird_clock_t now = 1, now_real, boot_time;
129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182

static void
update_times_plain(void)
{
  bird_clock_t new_time = time(NULL);
  int delta = new_time - now_real;

  if ((delta >= 0) && (delta < 60))
    now += delta;
  else if (now_real != 0)
   log(L_WARN "Time jump, delta %d s", delta);

  now_real = new_time;
}

static void
update_times_gettime(void)
{
  struct timespec ts;
  int rv;

  rv = clock_gettime(CLOCK_MONOTONIC, &ts);
  if (rv != 0)
    die("clock_gettime: %m");

  if (ts.tv_sec != now) {
    if (ts.tv_sec < now)
      log(L_ERR "Monotonic timer is broken");

    now = ts.tv_sec;
    now_real = time(NULL);
  }
}

static int clock_monotonic_available;

static inline void
update_times(void)
{
  if (clock_monotonic_available)
    update_times_gettime();
  else
    update_times_plain();
}

static inline void
init_times(void)
{
 struct timespec ts;
 clock_monotonic_available = (clock_gettime(CLOCK_MONOTONIC, &ts) == 0);
 if (!clock_monotonic_available)
   log(L_WARN "Monotonic timer is missing");
}

183 184 185 186 187 188 189 190 191 192 193 194 195 196

static void
tm_free(resource *r)
{
  timer *t = (timer *) r;

  tm_stop(t);
}

static void
tm_dump(resource *r)
{
  timer *t = (timer *) r;

197
  debug("(code %p, data %p, ", t->hook, t->data);
198 199 200 201
  if (t->randomize)
    debug("rand %d, ", t->randomize);
  if (t->recurrent)
    debug("recur %d, ", t->recurrent);
202 203 204 205 206 207 208 209 210 211
  if (t->expires)
    debug("expires in %d sec)\n", t->expires - now);
  else
    debug("inactive)\n");
}

static struct resclass tm_class = {
  "Timer",
  sizeof(timer),
  tm_free,
212
  tm_dump,
213
  NULL,
214
  NULL
215 216
};

217 218 219 220 221 222 223 224
/**
 * tm_new - create a timer
 * @p: pool
 *
 * This function creates a new timer resource and returns
 * a pointer to it. To use the timer, you need to fill in
 * the structure fields and call tm_start() to start timing.
 */
225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241
timer *
tm_new(pool *p)
{
  timer *t = ralloc(p, &tm_class);
  return t;
}

static inline void
tm_insert_near(timer *t)
{
  node *n = HEAD(near_timers);

  while (n->next && (SKIP_BACK(timer, n, n)->expires < t->expires))
    n = n->next;
  insert_node(&t->n, n->prev);
}

242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258
/**
 * tm_start - start a timer
 * @t: timer
 * @after: number of seconds the timer should be run after
 *
 * This function schedules the hook function of the timer to
 * be called after @after seconds. If the timer has been already
 * started, it's @expire time is replaced by the new value.
 *
 * You can have set the @randomize field of @t, the timeout
 * will be increased by a random number of seconds chosen
 * uniformly from range 0 .. @randomize.
 *
 * You can call tm_start() from the handler function of the timer
 * to request another run of the timer. Also, you can set the @recurrent
 * field to have the timer re-added automatically with the same timeout.
 */
259 260 261 262 263 264
void
tm_start(timer *t, unsigned after)
{
  bird_clock_t when;

  if (t->randomize)
265
    after += random() % (t->randomize + 1);
266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281
  when = now + after;
  if (t->expires == when)
    return;
  if (t->expires)
    rem_node(&t->n);
  t->expires = when;
  if (after <= NEAR_TIMER_LIMIT)
    tm_insert_near(t);
  else
    {
      if (!first_far_timer || first_far_timer > when)
	first_far_timer = when;
      add_tail(&far_timers, &t->n);
    }
}

282 283 284 285 286 287 288
/**
 * tm_stop - stop a timer
 * @t: timer
 *
 * This function stops a timer. If the timer is already stopped,
 * nothing happens.
 */
289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335
void
tm_stop(timer *t)
{
  if (t->expires)
    {
      rem_node(&t->n);
      t->expires = 0;
    }
}

static void
tm_dump_them(char *name, list *l)
{
  node *n;
  timer *t;

  debug("%s timers:\n", name);
  WALK_LIST(n, *l)
    {
      t = SKIP_BACK(timer, n, n);
      debug("%p ", t);
      tm_dump(&t->r);
    }
  debug("\n");
}

void
tm_dump_all(void)
{
  tm_dump_them("Near", &near_timers);
  tm_dump_them("Far", &far_timers);
}

static inline time_t
tm_first_shot(void)
{
  time_t x = first_far_timer;

  if (!EMPTY_LIST(near_timers))
    {
      timer *t = SKIP_BACK(timer, n, HEAD(near_timers));
      if (t->expires < x)
	x = t->expires;
    }
  return x;
}

336 337
void io_log_event(void *hook, void *data);

338 339 340 341 342 343 344 345
static void
tm_shot(void)
{
  timer *t;
  node *n, *m;

  if (first_far_timer <= now)
    {
346
      bird_clock_t limit = now + NEAR_TIMER_LIMIT;
347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363
      first_far_timer = TIME_INFINITY;
      n = HEAD(far_timers);
      while (m = n->next)
	{
	  t = SKIP_BACK(timer, n, n);
	  if (t->expires <= limit)
	    {
	      rem_node(n);
	      tm_insert_near(t);
	    }
	  else if (t->expires < first_far_timer)
	    first_far_timer = t->expires;
	  n = m;
	}
    }
  while ((n = HEAD(near_timers)) -> next)
    {
364
      int delay;
365 366 367 368
      t = SKIP_BACK(timer, n, n);
      if (t->expires > now)
	break;
      rem_node(n);
369
      delay = t->expires - now;
370
      t->expires = 0;
371 372 373 374 375 376 377
      if (t->recurrent)
	{
	  int i = t->recurrent - delay;
	  if (i < 0)
	    i = 0;
	  tm_start(t, i);
	}
378
      io_log_event(t->hook, t->data);
379 380 381 382
      t->hook(t);
    }
}

383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406
/**
 * tm_parse_datetime - parse a date and time
 * @x: datetime string
 *
 * tm_parse_datetime() takes a textual representation of
 * a date and time (dd-mm-yyyy hh:mm:ss)
 * and converts it to the corresponding value of type &bird_clock_t.
 */
bird_clock_t
tm_parse_datetime(char *x)
{
  struct tm tm;
  int n;
  time_t t;

  if (sscanf(x, "%d-%d-%d %d:%d:%d%n", &tm.tm_mday, &tm.tm_mon, &tm.tm_year, &tm.tm_hour, &tm.tm_min, &tm.tm_sec, &n) != 6 || x[n])
    return tm_parse_date(x);
  tm.tm_mon--;
  tm.tm_year -= 1900;
  t = mktime(&tm);
  if (t == (time_t) -1)
    return 0;
  return t;
}
407 408 409 410 411 412 413
/**
 * tm_parse_date - parse a date
 * @x: date string
 *
 * tm_parse_date() takes a textual representation of a date (dd-mm-yyyy)
 * and converts it to the corresponding value of type &bird_clock_t.
 */
414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431
bird_clock_t
tm_parse_date(char *x)
{
  struct tm tm;
  int n;
  time_t t;

  if (sscanf(x, "%d-%d-%d%n", &tm.tm_mday, &tm.tm_mon, &tm.tm_year, &n) != 3 || x[n])
    return 0;
  tm.tm_mon--;
  tm.tm_year -= 1900;
  tm.tm_hour = tm.tm_min = tm.tm_sec = 0;
  t = mktime(&tm);
  if (t == (time_t) -1)
    return 0;
  return t;
}

432 433
static void
tm_format_reltime(char *x, struct tm *tm, bird_clock_t delta)
434
{
435 436
  static char *month_names[12] = { "Jan", "Feb", "Mar", "Apr", "May", "Jun",
				   "Jul", "Aug", "Sep", "Oct", "Nov", "Dec" };
437

438 439 440 441 442 443
  if (delta < 20*3600)
    bsprintf(x, "%02d:%02d", tm->tm_hour, tm->tm_min);
  else if (delta < 360*86400)
    bsprintf(x, "%s%02d", month_names[tm->tm_mon], tm->tm_mday);
  else
    bsprintf(x, "%d", tm->tm_year+1900);
444 445
}

446 447
#include "conf/conf.h"

448 449 450 451 452
/**
 * tm_format_datetime - convert date and time to textual representation
 * @x: destination buffer of size %TM_DATETIME_BUFFER_SIZE
 * @t: time
 *
453 454
 * This function formats the given relative time value @t to a textual
 * date/time representation (dd-mm-yyyy hh:mm:ss) in real time.
455
 */
456
void
457
tm_format_datetime(char *x, struct timeformat *fmt_spec, bird_clock_t t)
458
{
459
  const char *fmt_used;
460
  struct tm *tm;
461 462
  bird_clock_t delta = now - t;
  t = now_real - delta;
463 464
  tm = localtime(&t);

465 466
  if (fmt_spec->fmt1 == NULL)
    return tm_format_reltime(x, tm, delta);
467

468 469
  if ((fmt_spec->limit == 0) || (delta < fmt_spec->limit))
    fmt_used = fmt_spec->fmt1;
470
  else
471 472 473 474 475
    fmt_used = fmt_spec->fmt2;

  int rv = strftime(x, TM_DATETIME_BUFFER_SIZE, fmt_used, tm);
  if (((rv == 0) && fmt_used[0]) || (rv == TM_DATETIME_BUFFER_SIZE))
    strcpy(x, "<too-long>");
476 477
}

478

479 480 481 482 483 484 485 486 487 488
/**
 * DOC: Sockets
 *
 * Socket resources represent network connections. Their data structure (&socket)
 * contains a lot of fields defining the exact type of the socket, the local and
 * remote addresses and ports, pointers to socket buffers and finally pointers to
 * hook functions to be called when new data have arrived to the receive buffer
 * (@rx_hook), when the contents of the transmit buffer have been transmitted
 * (@tx_hook) and when an error or connection close occurs (@err_hook).
 *
489
 * Freeing of sockets from inside socket hooks is perfectly safe.
490 491
 */

492 493 494 495
#ifndef SOL_IP
#define SOL_IP IPPROTO_IP
#endif

496 497 498 499
#ifndef SOL_IPV6
#define SOL_IPV6 IPPROTO_IPV6
#endif

500 501 502 503 504
#ifndef SOL_ICMPV6
#define SOL_ICMPV6 IPPROTO_ICMPV6
#endif


505 506 507
/*
 *	Sockaddr helper functions
 */
508

509 510 511 512 513
static inline int sockaddr_length(int af)
{ return (af == AF_INET) ? sizeof(struct sockaddr_in) : sizeof(struct sockaddr_in6); }

static inline void
sockaddr_fill4(struct sockaddr_in *sa, ip_addr a, struct iface *ifa, uint port)
514
{
515 516 517 518 519 520 521
  memset(sa, 0, sizeof(struct sockaddr_in));
#ifdef HAVE_SIN_LEN
  sa->sin_len = sizeof(struct sockaddr_in);
#endif
  sa->sin_family = AF_INET;
  sa->sin_port = htons(port);
  sa->sin_addr = ipa_to_in4(a);
522
}
523

524 525
static inline void
sockaddr_fill6(struct sockaddr_in6 *sa, ip_addr a, struct iface *ifa, uint port)
526
{
527 528 529 530 531 532 533 534 535 536 537
  memset(sa, 0, sizeof(struct sockaddr_in6));
#ifdef SIN6_LEN
  sa->sin6_len = sizeof(struct sockaddr_in6);
#endif
  sa->sin6_family = AF_INET6;
  sa->sin6_port = htons(port);
  sa->sin6_flowinfo = 0;
  sa->sin6_addr = ipa_to_in6(a);

  if (ifa && ipa_is_link_local(a))
    sa->sin6_scope_id = ifa->index;
538
}
539

540 541
void
sockaddr_fill(sockaddr *sa, int af, ip_addr a, struct iface *ifa, uint port)
542
{
543 544 545 546 547 548
  if (af == AF_INET)
    sockaddr_fill4((struct sockaddr_in *) sa, a, ifa, port);
  else if (af == AF_INET6)
    sockaddr_fill6((struct sockaddr_in6 *) sa, a, ifa, port);
  else
    bug("Unknown AF");
549 550
}

551 552
static inline void
sockaddr_read4(struct sockaddr_in *sa, ip_addr *a, struct iface **ifa, uint *port)
553
{
554 555
  *port = ntohs(sa->sin_port);
  *a = ipa_from_in4(sa->sin_addr);
556 557
}

558 559
static inline void
sockaddr_read6(struct sockaddr_in6 *sa, ip_addr *a, struct iface **ifa, uint *port)
560
{
561 562
  *port = ntohs(sa->sin6_port);
  *a = ipa_from_in6(sa->sin6_addr);
563

564 565
  if (ifa && ipa_is_link_local(*a))
    *ifa = if_find_by_index(sa->sin6_scope_id);
566 567
}

568 569
int
sockaddr_read(sockaddr *sa, int af, ip_addr *a, struct iface **ifa, uint *port)
570
{
571 572
  if (sa->sa.sa_family != af)
    goto fail;
573

574 575 576 577 578 579
  if (af == AF_INET)
    sockaddr_read4((struct sockaddr_in *) sa, a, ifa, port);
  else if (af == AF_INET6)
    sockaddr_read6((struct sockaddr_in6 *) sa, a, ifa, port);
  else
    goto fail;
580

581
  return 0;
582

583 584 585 586
 fail:
  *a = IPA_NONE;
  *port = 0;
  return -1;
587 588 589
}


590 591 592
/*
 *	IPv6 multicast syscalls
 */
593

594
/* Fortunately standardized in RFC 3493 */
595

596 597
#define INIT_MREQ6(maddr,ifa) \
  { .ipv6mr_multiaddr = ipa_to_in6(maddr), .ipv6mr_interface = ifa->index }
598

599 600
static inline int
sk_setup_multicast6(sock *s)
601
{
602 603 604
  int index = s->iface->index;
  int ttl = s->ttl;
  int n = 0;
605

606 607
  if (setsockopt(s->fd, SOL_IPV6, IPV6_MULTICAST_IF, &index, sizeof(index)) < 0)
    ERR("IPV6_MULTICAST_IF");
608

609 610
  if (setsockopt(s->fd, SOL_IPV6, IPV6_MULTICAST_HOPS, &ttl, sizeof(ttl)) < 0)
    ERR("IPV6_MULTICAST_HOPS");
611

612 613
  if (setsockopt(s->fd, SOL_IPV6, IPV6_MULTICAST_LOOP, &n, sizeof(n)) < 0)
    ERR("IPV6_MULTICAST_LOOP");
614

615
  return 0;
616 617
}

618 619
static inline int
sk_join_group6(sock *s, ip_addr maddr)
620
{
621
  struct ipv6_mreq mr = INIT_MREQ6(maddr, s->iface);
622

623 624
  if (setsockopt(s->fd, SOL_IPV6, IPV6_JOIN_GROUP, &mr, sizeof(mr)) < 0)
    ERR("IPV6_JOIN_GROUP");
625

626
  return 0;
627 628
}

629 630
static inline int
sk_leave_group6(sock *s, ip_addr maddr)
631
{
632
  struct ipv6_mreq mr = INIT_MREQ6(maddr, s->iface);
633

634 635 636 637 638
  if (setsockopt(s->fd, SOL_IPV6, IPV6_LEAVE_GROUP, &mr, sizeof(mr)) < 0)
    ERR("IPV6_LEAVE_GROUP");

  return 0;
}
639

640

641 642 643
/*
 *	IPv6 packet control messages
 */
644

645
/* Also standardized, in RFC 3542 */
646

647 648 649 650 651 652 653 654 655
/*
 * RFC 2292 uses IPV6_PKTINFO for both the socket option and the cmsg
 * type, RFC 3542 changed the socket option to IPV6_RECVPKTINFO. If we
 * don't have IPV6_RECVPKTINFO we suppose the OS implements the older
 * RFC and we use IPV6_PKTINFO.
 */
#ifndef IPV6_RECVPKTINFO
#define IPV6_RECVPKTINFO IPV6_PKTINFO
#endif
656 657 658 659 660 661
/*
 * Same goes for IPV6_HOPLIMIT -> IPV6_RECVHOPLIMIT.
 */
#ifndef IPV6_RECVHOPLIMIT
#define IPV6_RECVHOPLIMIT IPV6_HOPLIMIT
#endif
662

663

664 665
#define CMSG6_SPACE_PKTINFO CMSG_SPACE(sizeof(struct in6_pktinfo))
#define CMSG6_SPACE_TTL CMSG_SPACE(sizeof(int))
666

667 668 669 670
static inline int
sk_request_cmsg6_pktinfo(sock *s)
{
  int y = 1;
671

672 673 674 675
  if (setsockopt(s->fd, SOL_IPV6, IPV6_RECVPKTINFO, &y, sizeof(y)) < 0)
    ERR("IPV6_RECVPKTINFO");

  return 0;
676 677
}

678 679
static inline int
sk_request_cmsg6_ttl(sock *s)
680
{
681
  int y = 1;
682

683 684
  if (setsockopt(s->fd, SOL_IPV6, IPV6_RECVHOPLIMIT, &y, sizeof(y)) < 0)
    ERR("IPV6_RECVHOPLIMIT");
685

686 687
  return 0;
}
688

689 690 691 692
static inline void
sk_process_cmsg6_pktinfo(sock *s, struct cmsghdr *cm)
{
  if (cm->cmsg_type == IPV6_PKTINFO)
693
  {
694 695 696
    struct in6_pktinfo *pi = (struct in6_pktinfo *) CMSG_DATA(cm);
    s->laddr = ipa_from_in6(pi->ipi6_addr);
    s->lifindex = pi->ipi6_ifindex;
697
  }
698
}
699

700 701 702 703 704
static inline void
sk_process_cmsg6_ttl(sock *s, struct cmsghdr *cm)
{
  if (cm->cmsg_type == IPV6_HOPLIMIT)
    s->rcv_ttl = * (int *) CMSG_DATA(cm);
705 706
}

707 708
static inline void
sk_prepare_cmsgs6(sock *s, struct msghdr *msg, void *cbuf, size_t cbuflen)
709 710 711
{
  struct cmsghdr *cm;
  struct in6_pktinfo *pi;
712
  int controllen = 0;
713 714 715 716 717

  msg->msg_control = cbuf;
  msg->msg_controllen = cbuflen;

  cm = CMSG_FIRSTHDR(msg);
718
  cm->cmsg_level = SOL_IPV6;
719 720
  cm->cmsg_type = IPV6_PKTINFO;
  cm->cmsg_len = CMSG_LEN(sizeof(*pi));
721
  controllen += CMSG_SPACE(sizeof(*pi));
722 723 724

  pi = (struct in6_pktinfo *) CMSG_DATA(cm);
  pi->ipi6_ifindex = s->iface ? s->iface->index : 0;
725
  pi->ipi6_addr = ipa_to_in6(s->saddr);
726

727
  msg->msg_controllen = controllen;
728
}
729

730

731 732 733 734 735 736
/*
 *	Miscellaneous socket syscalls
 */

static inline int
sk_set_ttl4(sock *s, int ttl)
737
{
738 739 740 741
  if (setsockopt(s->fd, SOL_IP, IP_TTL, &ttl, sizeof(ttl)) < 0)
    ERR("IP_TTL");

  return 0;
742 743
}

744 745 746 747 748
static inline int
sk_set_ttl6(sock *s, int ttl)
{
  if (setsockopt(s->fd, SOL_IPV6, IPV6_UNICAST_HOPS, &ttl, sizeof(ttl)) < 0)
    ERR("IPV6_UNICAST_HOPS");
749

750 751 752 753 754
  return 0;
}

static inline int
sk_set_tos4(sock *s, int tos)
755
{
756 757
  if (setsockopt(s->fd, SOL_IP, IP_TOS, &tos, sizeof(tos)) < 0)
    ERR("IP_TOS");
758

759 760
  return 0;
}
761

762 763 764 765 766
static inline int
sk_set_tos6(sock *s, int tos)
{
  if (setsockopt(s->fd, SOL_IPV6, IPV6_TCLASS, &tos, sizeof(tos)) < 0)
    ERR("IPV6_TCLASS");
767

768 769
  return 0;
}
770

771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796
static inline int
sk_set_high_port(sock *s)
{
  /* Port range setting is optional, ignore it if not supported */

#ifdef IP_PORTRANGE
  if (sk_is_ipv4(s))
  {
    int range = IP_PORTRANGE_HIGH;
    if (setsockopt(s->fd, SOL_IP, IP_PORTRANGE, &range, sizeof(range)) < 0)
      ERR("IP_PORTRANGE");
  }
#endif

#ifdef IPV6_PORTRANGE
  if (sk_is_ipv6(s))
  {
    int range = IPV6_PORTRANGE_HIGH;
    if (setsockopt(s->fd, SOL_IPV6, IPV6_PORTRANGE, &range, sizeof(range)) < 0)
      ERR("IPV6_PORTRANGE");
  }
#endif

  return 0;
}

797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819
static inline byte *
sk_skip_ip_header(byte *pkt, int *len)
{
  if ((*len < 20) || ((*pkt & 0xf0) != 0x40))
    return NULL;

  int hlen = (*pkt & 0x0f) * 4;
  if ((hlen < 20) || (hlen > *len))
    return NULL;

  *len -= hlen;
  return pkt + hlen;
}

byte *
sk_rx_buffer(sock *s, int *len)
{
  if (sk_is_ipv4(s) && (s->type == SK_IP))
    return sk_skip_ip_header(s->rbuf, len);
  else
    return s->rbuf;
}

820

821 822 823
/*
 *	Public socket functions
 */
824

825 826 827 828 829 830 831 832 833
/**
 * sk_setup_multicast - enable multicast for given socket
 * @s: socket
 *
 * Prepare transmission of multicast packets for given datagram socket.
 * The socket must have defined @iface.
 *
 * Result: 0 for success, -1 for an error.
 */
834

835 836 837 838
int
sk_setup_multicast(sock *s)
{
  ASSERT(s->iface);
839

840 841 842 843 844
  if (sk_is_ipv4(s))
    return sk_setup_multicast4(s);
  else
    return sk_setup_multicast6(s);
}
845

846 847 848 849 850 851 852 853 854 855
/**
 * sk_join_group - join multicast group for given socket
 * @s: socket
 * @maddr: multicast address
 *
 * Join multicast group for given datagram socket and associated interface.
 * The socket must have defined @iface.
 *
 * Result: 0 for success, -1 for an error.
 */
856

857 858 859 860 861 862 863 864
int
sk_join_group(sock *s, ip_addr maddr)
{
  if (sk_is_ipv4(s))
    return sk_join_group4(s, maddr);
  else
    return sk_join_group6(s, maddr);
}
865

866 867 868 869 870 871 872 873 874 875
/**
 * sk_leave_group - leave multicast group for given socket
 * @s: socket
 * @maddr: multicast address
 *
 * Leave multicast group for given datagram socket and associated interface.
 * The socket must have defined @iface.
 *
 * Result: 0 for success, -1 for an error.
 */
876

877 878 879 880 881 882 883
int
sk_leave_group(sock *s, ip_addr maddr)
{
  if (sk_is_ipv4(s))
    return sk_leave_group4(s, maddr);
  else
    return sk_leave_group6(s, maddr);
884 885
}

886
/**
887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909
 * sk_setup_broadcast - enable broadcast for given socket
 * @s: socket
 *
 * Allow reception and transmission of broadcast packets for given datagram
 * socket. The socket must have defined @iface. For transmission, packets should
 * be send to @brd address of @iface.
 *
 * Result: 0 for success, -1 for an error.
 */

int
sk_setup_broadcast(sock *s)
{
  int y = 1;

  if (setsockopt(s->fd, SOL_SOCKET, SO_BROADCAST, &y, sizeof(y)) < 0)
    ERR("SO_BROADCAST");

  return 0;
}

/**
 * sk_set_ttl - set transmit TTL for given socket
910 911 912
 * @s: socket
 * @ttl: TTL value
 *
913 914
 * Set TTL for already opened connections when TTL was not set before. Useful
 * for accepted connections when different ones should have different TTL.
915 916 917 918 919 920 921 922 923
 *
 * Result: 0 for success, -1 for an error.
 */

int
sk_set_ttl(sock *s, int ttl)
{
  s->ttl = ttl;

924 925 926 927
  if (sk_is_ipv4(s))
    return sk_set_ttl4(s, ttl);
  else
    return sk_set_ttl6(s, ttl);
928 929
}

930
/**
931
 * sk_set_min_ttl - set minimal accepted TTL for given socket
932 933 934
 * @s: socket
 * @ttl: TTL value
 *
935 936
 * Set minimal accepted TTL for given socket. Can be used for TTL security.
 * implementations.
937 938 939 940 941 942 943
 *
 * Result: 0 for success, -1 for an error.
 */

int
sk_set_min_ttl(sock *s, int ttl)
{
944 945 946 947
  if (sk_is_ipv4(s))
    return sk_set_min_ttl4(s, ttl);
  else
    return sk_set_min_ttl6(s, ttl);
948
}
949

950
#if 0
951
/**
952
 * sk_set_md5_auth - add / remove MD5 security association for given socket
953 954
 * @s: socket
 * @a: IP address of the other side
955
 * @ifa: Interface for link-local IP address
956 957
 * @passwd: password used for MD5 authentication
 *
958 959 960 961
 * In TCP MD5 handling code in kernel, there is a set of pairs (address,
 * password) used to choose password according to address of the other side.
 * This function is useful for listening socket, for active sockets it is enough
 * to set s->password field.
962 963 964 965 966 967 968 969
 *
 * When called with passwd != NULL, the new pair is added,
 * When called with passwd == NULL, the existing pair is removed.
 *
 * Result: 0 for success, -1 for an error.
 */

int
970
sk_set_md5_auth(sock *s, ip_addr a, struct iface *ifa, char *passwd)
971 972
{ DUMMY; }
#endif
973

974 975 976 977 978 979 980 981 982 983 984 985
/**
 * sk_set_ipv6_checksum - specify IPv6 checksum offset for given socket
 * @s: socket
 * @offset: offset
 *
 * Specify IPv6 checksum field offset for given raw IPv6 socket. After that, the
 * kernel will automatically fill it for outgoing packets and check it for
 * incoming packets. Should not be used on ICMPv6 sockets, where the position is
 * known to the kernel.
 *
 * Result: 0 for success, -1 for an error.
 */
986

987 988 989
int
sk_set_ipv6_checksum(sock *s, int offset)
{
990
  if (setsockopt(s->fd, SOL_IPV6, IPV6_CHECKSUM, &offset, sizeof(offset)) < 0)
991
    ERR("IPV6_CHECKSUM");
992 993 994 995

  return 0;
}

996
int
997
sk_set_icmp6_filter(sock *s, int p1, int p2)
998 999 1000 1001 1002 1003 1004 1005
{
  /* a bit of lame interface, but it is here only for Radv */
  struct icmp6_filter f;

  ICMP6_FILTER_SETBLOCKALL(&f);
  ICMP6_FILTER_SETPASS(p1, &f);
  ICMP6_FILTER_SETPASS(p2, &f);

1006
  if (setsockopt(s->fd, SOL_ICMPV6, ICMP6_FILTER, &f, sizeof(f)) < 0)
1007
    ERR("ICMP6_FILTER");
1008 1009 1010 1011

  return 0;
}

1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133
void
sk_log_error(sock *s, const char *p)
{
  log(L_ERR "%s: Socket error: %s%#m", p, s->err);
}


/*
 *	Actual struct birdsock code
 */

static list sock_list;
static struct birdsock *current_sock;
static struct birdsock *stored_sock;

static inline sock *
sk_next(sock *s)
{
  if (!s->n.next->next)
    return NULL;
  else
    return SKIP_BACK(sock, n, s->n.next);
}

static void
sk_alloc_bufs(sock *s)
{
  if (!s->rbuf && s->rbsize)
    s->rbuf = s->rbuf_alloc = xmalloc(s->rbsize);
  s->rpos = s->rbuf;
  if (!s->tbuf && s->tbsize)
    s->tbuf = s->tbuf_alloc = xmalloc(s->tbsize);
  s->tpos = s->ttx = s->tbuf;
}

static void
sk_free_bufs(sock *s)
{
  if (s->rbuf_alloc)
  {
    xfree(s->rbuf_alloc);
    s->rbuf = s->rbuf_alloc = NULL;
  }
  if (s->tbuf_alloc)
  {
    xfree(s->tbuf_alloc);
    s->tbuf = s->tbuf_alloc = NULL;
  }
}

static void
sk_free(resource *r)
{
  sock *s = (sock *) r;

  sk_free_bufs(s);
  if (s->fd >= 0)
  {
    close(s->fd);

    /* FIXME: we should call sk_stop() for SKF_THREAD sockets */
    if (s->flags & SKF_THREAD)
      return;

    if (s == current_sock)
      current_sock = sk_next(s);
    if (s == stored_sock)
      stored_sock = sk_next(s);
    rem_node(&s->n);
  }
}

void
sk_set_rbsize(sock *s, uint val)
{
  ASSERT(s->rbuf_alloc == s->rbuf);

  if (s->rbsize == val)
    return;

  s->rbsize = val;
  xfree(s->rbuf_alloc);
  s->rbuf_alloc = xmalloc(val);
  s->rpos = s->rbuf = s->rbuf_alloc;
}

void
sk_set_tbsize(sock *s, uint val)
{
  ASSERT(s->tbuf_alloc == s->tbuf);

  if (s->tbsize == val)
    return;

  byte *old_tbuf = s->tbuf;

  s->tbsize = val;
  s->tbuf = s->tbuf_alloc = xrealloc(s->tbuf_alloc, val);
  s->tpos = s->tbuf + (s->tpos - old_tbuf);
  s->ttx  = s->tbuf + (s->ttx  - old_tbuf);
}

void
sk_set_tbuf(sock *s, void *tbuf)
{
  s->tbuf = tbuf ?: s->tbuf_alloc;
  s->ttx = s->tpos = s->tbuf;
}

void
sk_reallocate(sock *s)
{
  sk_free_bufs(s);
  sk_alloc_bufs(s);
}

static void
sk_dump(resource *r)
{
  sock *s = (sock *) r;
  static char *sk_type_names[] = { "TCP<", "TCP>", "TCP", "UDP", NULL, "IP", NULL, "MAGIC", "UNIX<", "UNIX", "DEL!" };

1134
  debug("(%s, ud=%p, sa=%I, sp=%d, da=%I, dp=%d, tos=%d, ttl=%d, if=%s)\n",
1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178
	sk_type_names[s->type],
	s->data,
	s->saddr,
	s->sport,
	s->daddr,
	s->dport,
	s->tos,
	s->ttl,
	s->iface ? s->iface->name : "none");
}

static struct resclass sk_class = {
  "Socket",
  sizeof(sock),
  sk_free,
  sk_dump,
  NULL,
  NULL
};

/**
 * sk_new - create a socket
 * @p: pool
 *
 * This function creates a new socket resource. If you want to use it,
 * you need to fill in all the required fields of the structure and
 * call sk_open() to do the actual opening of the socket.
 *
 * The real function name is sock_new(), sk_new() is a macro wrapper
 * to avoid collision with OpenSSL.
 */
sock *
sock_new(pool *p)
{
  sock *s = ralloc(p, &sk_class);
  s->pool = p;
  // s->saddr = s->daddr = IPA_NONE;
  s->tos = s->priority = s->ttl = -1;
  s->fd = -1;
  return s;
}

static int
sk_setup(sock *s)
1179
{
1180 1181
  int y = 1;
  int fd = s->fd;
1182

1183 1184
  if (fcntl(fd, F_SETFL, O_NONBLOCK) < 0)
    ERR("O_NONBLOCK");
1185

1186 1187
  if (!s->af)
    return 0;
1188

1189 1190
  if (ipa_nonzero(s->saddr) && !(s->flags & SKF_BIND))
    s->flags |= SKF_PKTINFO;
1191

1192 1193 1194 1195 1196 1197 1198 1199
#ifdef CONFIG_USE_HDRINCL
  if (sk_is_ipv4(s) && (s->type == SK_IP) && (s->flags & SKF_PKTINFO))
  {
    s->flags &= ~SKF_PKTINFO;
    s->flags |= SKF_HDRINCL;
    if (setsockopt(fd, SOL_IP, IP_HDRINCL, &y, sizeof(y)) < 0)
      ERR("IP_HDRINCL");
  }
1200 1201
#endif

1202 1203 1204 1205 1206 1207 1208 1209
  if (s->iface)
  {
#ifdef SO_BINDTODEVICE
    struct ifreq ifr;
    strcpy(ifr.ifr_name, s->iface->name);
    if (setsockopt(s->fd, SOL_SOCKET, SO_BINDTODEVICE, &ifr, sizeof(ifr)) < 0)
      ERR("SO_BINDTODEVICE");
#endif
1210

1211 1212 1213 1214 1215
#ifdef CONFIG_UNIX_DONTROUTE
    if (setsockopt(s->fd, SOL_SOCKET, SO_DONTROUTE, &y, sizeof(y)) < 0)
      ERR("SO_DONTROUTE");
#endif
  }
1216

1217 1218
  if (s->priority >= 0)
    if (sk_set_priority(s, s->priority) < 0)
1219 1220
      return -1;

1221 1222 1223 1224 1225
  if (sk_is_ipv4(s))
  {
    if (s->flags & SKF_LADDR_RX)
      if (sk_request_cmsg4_pktinfo(s) < 0)
	return -1;
1226

1227 1228 1229
    if (s->flags & SKF_TTL_RX)
      if (sk_request_cmsg4_ttl(s) < 0)
	return -1;
1230

1231 1232 1233
    if ((s->type == SK_UDP) || (s->type == SK_IP))
      if (sk_disable_mtu_disc4(s) < 0)
	return -1;
1234

1235 1236 1237
    if (s->ttl >= 0)
      if (sk_set_ttl4(s, s->ttl) < 0)
	return -1;
1238

1239 1240 1241 1242
    if (s->tos >= 0)
      if (sk_set_tos4(s, s->tos) < 0)
	return -1;
  }
1243

1244 1245 1246 1247 1248
  if (sk_is_ipv6(s))
  {
    if (s->flags & SKF_V6ONLY)
      if (setsockopt(fd, SOL_IPV6, IPV6_V6ONLY, &y, sizeof(y)) < 0)
	ERR("IPV6_V6ONLY");
1249

1250 1251 1252
    if (s->flags & SKF_LADDR_RX)
      if (sk_request_cmsg6_pktinfo(s) < 0)
	return -1;
1253

1254 1255 1256
    if (s->flags & SKF_TTL_RX)
      if (sk_request_cmsg6_ttl(s) < 0)
	return -1;
1257

1258 1259 1260
    if ((s->type == SK_UDP) || (s->type == SK_IP))
      if (sk_disable_mtu_disc6(s) < 0)
	return -1;
1261

1262 1263 1264
    if (s->ttl >= 0)
      if (sk_set_ttl6(s, s->ttl) < 0)
	return -1;
1265

1266 1267 1268 1269
    if (s->tos >= 0)
      if (sk_set_tos6(s, s->tos) < 0)
	return -1;
  }
1270 1271 1272 1273

  return 0;
}

1274 1275
static void
sk_insert(sock *s)
1276
{
1277
  add_tail(&sock_list, &s->n);
1278 1279
}

1280
static void
1281 1282
sk_tcp_connected(sock *s)
{
1283 1284 1285 1286 1287 1288
  sockaddr sa;
  int sa_len = sizeof(sa);

  if ((getsockname(s->fd, &sa.sa, &sa_len) < 0) ||
      (sockaddr_read(&sa, s->af, &s->saddr, &s->iface, &s->sport) < 0))
    log(L_WARN "SOCK: Cannot get local IP address for TCP>");
1289

1290 1291
  s->type = SK_TCP;
  sk_alloc_bufs(s);
1292
  s->tx_hook(s);
1293 1294
}

1295
static int
1296
sk_passive_connected(sock *s, int type)
1297
{
1298 1299 1300
  sockaddr loc_sa, rem_sa;
  int loc_sa_len = sizeof(loc_sa);
  int rem_sa_len = sizeof(rem_sa);
1301

1302 1303 1304 1305
  int fd = accept(s->fd, ((type == SK_TCP) ? &rem_sa.sa : NULL), &rem_sa_len);
  if (fd < 0)
  {
    if ((errno != EINTR) && (errno != EAGAIN))
1306
      s->err_hook(s, errno);
1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334
    return 0;
  }

  sock *t = sk_new(s->pool);
  t->type = type;
  t->fd = fd;
  t->af = s->af;
  t->ttl = s->ttl;
  t->tos = s->tos;
  t->rbsize = s->rbsize;
  t->tbsize = s->tbsize;

  if (type == SK_TCP)
  {
    if ((getsockname(fd, &loc_sa.sa, &loc_sa_len) < 0) ||
	(sockaddr_read(&loc_sa, s->af, &t->saddr, &t->iface, &t->sport) < 0))
      log(L_WARN "SOCK: Cannot get local IP address for TCP<");

    if (sockaddr_read(&rem_sa, s->af, &t->daddr, &t->iface, &t->dport) < 0)
      log(L_WARN "SOCK: Cannot get remote IP address for TCP<");
  }

  if (sk_setup(t) < 0)
  {
    /* FIXME: Call err_hook instead ? */
    log(L_ERR "SOCK: Incoming connection: %s%#m", t->err);

    /* FIXME: handle it better in rfree() */
1335
    close(t->fd);
1336 1337 1338 1339 1340 1341 1342 1343 1344
    t->fd = -1;
    rfree(t);
    return 1;
  }

  sk_insert(t);
  sk_alloc_bufs(t);
  s->rx_hook(t, 0);
  return 1;
1345 1346
}

1347 1348 1349 1350 1351 1352 1353 1354 1355 1356
/**
 * sk_open - open a socket
 * @s: socket
 *
 * This function takes a socket resource created by sk_new() and
 * initialized by the user and binds a corresponding network connection
 * to it.
 *
 * Result: 0 for success, -1 for an error.
 */
1357 1358 1359
int
sk_open(sock *s)
{
1360 1361
  int af = BIRD_AF;
  int fd = -1;
1362 1363 1364 1365
  int do_bind = 0;
  int bind_port = 0;
  ip_addr bind_addr = IPA_NONE;
  sockaddr sa;
1366

1367
  switch (s->type)
1368 1369 1370 1371 1372 1373 1374 1375 1376 1377
  {
  case SK_TCP_ACTIVE:
    s->ttx = "";			/* Force s->ttx != s->tpos */
    /* Fall thru */
  case SK_TCP_PASSIVE:
    fd = socket(af, SOCK_STREAM, IPPROTO_TCP);
    bind_port = s->sport;
    bind_addr = s->saddr;
    do_bind = bind_port || ipa_nonzero(bind_addr);
    break;
1378

1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401
  case SK_UDP:
    fd = socket(af, SOCK_DGRAM, IPPROTO_UDP);
    bind_port = s->sport;
    bind_addr = (s->flags & SKF_BIND) ? s->saddr : IPA_NONE;
    do_bind = 1;
    break;

  case SK_IP:
    fd = socket(af, SOCK_RAW, s->dport);
    bind_port = 0;
    bind_addr = (s->flags & SKF_BIND) ? s->saddr : IPA_NONE;
    do_bind = ipa_nonzero(bind_addr);
    break;

  case SK_MAGIC:
    af = 0;
    fd = s->fd;
    break;

  default:
    bug("sk_open() called for invalid sock type %d", s->type);
  }

1402
  if (fd < 0)
1403 1404 1405
    ERR("socket");

  s->af = af;
1406 1407
  s->fd = fd;

1408 1409
  if (sk_setup(s) < 0)
    goto err;
1410

1411
  if (do_bind)
1412 1413
  {
    if (bind_port)
1414
    {
1415 1416 1417 1418
      int y = 1;

      if (setsockopt(fd, SOL_SOCKET, SO_REUSEADDR, &y, sizeof(y)) < 0)
	ERR2("SO_REUSEADDR");
1419

1420
#ifdef CONFIG_NO_IFACE_BIND
1421 1422 1423 1424 1425 1426
      /* Workaround missing ability to bind to an iface */
      if ((s->type == SK_UDP) && s->iface && ipa_zero(bind_addr))
      {
	if (setsockopt(fd, SOL_SOCKET, SO_REUSEPORT, &y, sizeof(y)) < 0)
	  ERR2("SO_REUSEPORT");
      }
1427
#endif
1428
    }
1429 1430 1431 1432
    else
      if (s->flags & SKF_HIGH_PORT)
	if (sk_set_high_port(s) < 0)
	  log(L_WARN "Socket error: %s%#m", s->err);
1433

1434 1435 1436 1437
    sockaddr_fill(&sa, af, bind_addr, s->iface, bind_port);
    if (bind(fd, &sa.sa, SA_LEN(sa)) < 0)
      ERR2("bind");
  }
1438 1439

  if (s->password)
1440 1441
    if (sk_set_md5_auth(s, s->daddr, s->iface, s->password) < 0)
      goto err;
1442

1443
  switch (s->type)
1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464
  {
  case SK_TCP_ACTIVE:
    sockaddr_fill(&sa, af, s->daddr, s->iface, s->dport);
    if (connect(fd, &sa.sa, SA_LEN(sa)) >= 0)
      sk_tcp_connected(s);
    else if (errno != EINTR && errno != EAGAIN && errno != EINPROGRESS &&
	     errno != ECONNREFUSED && errno != EHOSTUNREACH && errno != ENETUNREACH)
      ERR2("connect");
    break;

  case SK_TCP_PASSIVE:
    if (listen(fd, 8) < 0)
      ERR2("listen");
    break;

  case SK_MAGIC:
    break;

  default:
    sk_alloc_bufs(s);
  }
1465

1466 1467
  if (!(s->flags & SKF_THREAD))
    sk_insert(s);
1468 1469
  return 0;

1470
err:
1471 1472 1473 1474 1475
  close(fd);
  s->fd = -1;
  return -1;
}

1476
int
1477 1478 1479
sk_open_unix(sock *s, char *name)
{
  struct sockaddr_un sa;
1480 1481 1482
  int fd;

  /* We are sloppy during error (leak fd and not set s->err), but we die anyway */
1483 1484 1485

  fd = socket(AF_UNIX, SOCK_STREAM, 0);
  if (fd < 0)
1486 1487 1488 1489
    return -1;

  if (fcntl(fd, F_SETFL, O_NONBLOCK) < 0)
    return -1;
1490

1491
  /* Path length checked in test_old_bird() */
1492
  sa.sun_family = AF_UNIX;
1493
  strcpy(sa.sun_path, name);
1494

1495
  if (bind(fd, (struct sockaddr *) &sa, SUN_LEN(&sa)) < 0)
1496 1497 1498 1499 1500 1501
    return -1;

  if (listen(fd, 8) < 0)
    return -1;

  s->fd = fd;
1502
  sk_insert(s);
1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535
  return 0;
}


#define CMSG_RX_SPACE MAX(CMSG4_SPACE_PKTINFO+CMSG4_SPACE_TTL, \
			  CMSG6_SPACE_PKTINFO+CMSG6_SPACE_TTL)
#define CMSG_TX_SPACE MAX(CMSG4_SPACE_PKTINFO,CMSG6_SPACE_PKTINFO)

static void
sk_prepare_cmsgs(sock *s, struct msghdr *msg, void *cbuf, size_t cbuflen)
{
  if (sk_is_ipv4(s))
    sk_prepare_cmsgs4(s, msg, cbuf, cbuflen);
  else
    sk_prepare_cmsgs6(s, msg, cbuf, cbuflen);
}

static void
sk_process_cmsgs(sock *s, struct msghdr *msg)
{
  struct cmsghdr *cm;

  s->laddr = IPA_NONE;
  s->lifindex = 0;
  s->rcv_ttl = -1;

  for (cm = CMSG_FIRSTHDR(msg); cm != NULL; cm = CMSG_NXTHDR(msg, cm))
  {
    if ((cm->cmsg_level == SOL_IP) && sk_is_ipv4(s))
    {
      sk_process_cmsg4_pktinfo(s, cm);
      sk_process_cmsg4_ttl(s, cm);
    }
1536

1537 1538 1539 1540 1541 1542
    if ((cm->cmsg_level == SOL_IPV6) && sk_is_ipv6(s))
    {
      sk_process_cmsg6_pktinfo(s, cm);
      sk_process_cmsg6_ttl(s, cm);
    }
  }
1543 1544
}

1545 1546 1547 1548 1549 1550 1551 1552

static inline int
sk_sendmsg(sock *s)
{
  struct iovec iov = {s->tbuf, s->tpos - s->tbuf};
  byte cmsg_buf[CMSG_TX_SPACE];
  sockaddr dst;

1553
  sockaddr_fill(&dst, s->af, s->daddr, s->iface, s->dport);
1554 1555

  struct msghdr msg = {
1556 1557
    .msg_name = &dst.sa,
    .msg_namelen = SA_LEN(dst),
1558 1559 1560 1561 1562 1563 1564 1565 1566 1567
    .msg_iov = &iov,
    .msg_iovlen = 1
  };

#ifdef CONFIG_USE_HDRINCL
  byte hdr[20];
  struct iovec iov2[2] = { {hdr, 20}, iov };

  if (s->flags & SKF_HDRINCL)
  {
1568
    sk_prepare_ip_header(s, hdr, iov.iov_len);
1569 1570 1571 1572 1573 1574
    msg.msg_iov = iov2;
    msg.msg_iovlen = 2;
  }
#endif

  if (s->flags & SKF_PKTINFO)
1575
    sk_prepare_cmsgs(s, &msg, cmsg_buf, sizeof(cmsg_buf));
1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587

  return sendmsg(s->fd, &msg, 0);
}

static inline int
sk_recvmsg(sock *s)
{
  struct iovec iov = {s->rbuf, s->rbsize};
  byte cmsg_buf[CMSG_RX_SPACE];
  sockaddr src;

  struct msghdr msg = {
1588 1589
    .msg_name = &src.sa,
    .msg_namelen = sizeof(src), // XXXX ??
1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605
    .msg_iov = &iov,
    .msg_iovlen = 1,
    .msg_control = cmsg_buf,
    .msg_controllen = sizeof(cmsg_buf),
    .msg_flags = 0
  };

  int rv = recvmsg(s->fd, &msg, 0);
  if (rv < 0)
    return rv;

  //ifdef IPV4
  //  if (cf_type == SK_IP)
  //    rv = ipv4_skip_header(pbuf, rv);
  //endif

1606 1607
  sockaddr_read(&src, s->af, &s->faddr, NULL, &s->fport);
  sk_process_cmsgs(s, &msg);
1608 1609 1610 1611 1612 1613 1614 1615 1616 1617

  if (msg.msg_flags & MSG_TRUNC)
    s->flags |= SKF_TRUNCATED;
  else
    s->flags &= ~SKF_TRUNCATED;

  return rv;
}


1618 1619
static inline void reset_tx_buffer(sock *s) { s->ttx = s->tpos = s->tbuf; }

1620 1621 1622 1623 1624 1625
static int
sk_maybe_write(sock *s)
{
  int e;

  switch (s->type)
1626 1627 1628 1629 1630
  {
  case SK_TCP:
  case SK_MAGIC:
  case SK_UNIX:
    while (s->ttx != s->tpos)
1631
    {
1632 1633 1634 1635 1636
      e = write(s->fd, s->ttx, s->tpos - s->ttx);

      if (e < 0)
      {
	if (errno != EINTR && errno != EAGAIN)
1637
	{
1638 1639 1640 1641
	  reset_tx_buffer(s);
	  /* EPIPE is just a connection close notification during TX */
	  s->err_hook(s, (errno != EPIPE) ? errno : 0);
	  return -1;
1642
	}
1643 1644 1645 1646 1647 1648 1649 1650 1651 1652 1653
	return 0;
      }
      s->ttx += e;
    }
    reset_tx_buffer(s);
    return 1;

  case SK_UDP:
  case SK_IP:
    {
      if (s->tbuf == s->tpos)
1654
	return 1;
1655 1656 1657 1658 1659 1660 1661 1662 1663 1664 1665 1666 1667 1668 1669

      e = sk_sendmsg(s);

      if (e < 0)
      {
	if (errno != EINTR && errno != EAGAIN)
	{
	  reset_tx_buffer(s);
	  s->err_hook(s, errno);
	  return -1;
	}

	if (!s->tx_hook)
	  reset_tx_buffer(s);
	return 0;
1670
      }
1671 1672
      reset_tx_buffer(s);
      return 1;
1673
    }
1674 1675 1676
  default:
    bug("sk_maybe_write: unknown socket type %d", s->type);
  }
1677 1678
}

1679 1680 1681 1682
int
sk_rx_ready(sock *s)
{
  int rv;
1683 1684
  struct pollfd pfd = { .fd = s->fd };
  pfd.events |= POLLIN;
1685 1686

 redo: