host.c

一个从网络上自动下载文件的自由工具

/* Host name resolution and matching.
   Copyright (C) 1996, 1997, 1998, 1999, 2000, 2001, 2002, 2003,
   2004, 2005, 2006, 2007, 2008 Free Software Foundation, Inc.

This file is part of GNU Wget.

GNU Wget 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 3 of the License, or
 (at your option) any later version.

GNU Wget is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
GNU General Public License for more details.

You should have received a copy of the GNU General Public License
along with Wget.  If not, see <http://www.gnu.org/licenses/>.

Additional permission under GNU GPL version 3 section 7

If you modify this program, or any covered work, by linking or
combining it with the OpenSSL project's OpenSSL library (or a
modified version of that library), containing parts covered by the
terms of the OpenSSL or SSLeay licenses, the Free Software Foundation
grants you additional permission to convey the resulting work.
Corresponding Source for a non-source form of such a combination
shall include the source code for the parts of OpenSSL used as well
as that of the covered work.  */

#include <config.h>

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <assert.h>

#ifndef WINDOWS
# include <sys/socket.h>
# include <netinet/in.h>
# ifndef __BEOS__
#  include <arpa/inet.h>
# endif
# include <netdb.h>
# define SET_H_ERRNO(err) ((void)(h_errno = (err)))
#else  /* WINDOWS */
# define SET_H_ERRNO(err) WSASetLastError (err)
#endif /* WINDOWS */

#include <errno.h>

#include "wget.h"
#include "utils.h"
#include "host.h"
#include "url.h"
#include "hash.h"

#ifndef NO_ADDRESS
# define NO_ADDRESS NO_DATA
#endif

/* Lists of IP addresses that result from running DNS queries.  See
   lookup_host for details.  */

struct address_list {
  int count;                    /* number of adrresses */
  ip_address *addresses;        /* pointer to the string of addresses */

  int faulty;                   /* number of addresses known not to work. */
  bool connected;               /* whether we were able to connect to
                                   one of the addresses in the list,
                                   at least once. */

  int refcount;                 /* reference count; when it drops to
                                   0, the entry is freed. */
};

/* Get the bounds of the address list.  */

void
address_list_get_bounds (const struct address_list *al, int *start, int *end)
{
  *start = al->faulty;
  *end   = al->count;
}

/* Return a pointer to the address at position POS.  */

const ip_address *
address_list_address_at (const struct address_list *al, int pos)
{
  assert (pos >= al->faulty && pos < al->count);
  return al->addresses + pos;
}

/* Return true if AL contains IP, false otherwise.  */

bool
address_list_contains (const struct address_list *al, const ip_address *ip)
{
  int i;
  switch (ip->family)
    {
    case AF_INET:
      for (i = 0; i < al->count; i++)
        {
          ip_address *cur = al->addresses + i;
          if (cur->family == AF_INET
              && (cur->data.d4.s_addr == ip->data.d4.s_addr))
            return true;
        }
      return false;
#ifdef ENABLE_IPV6
    case AF_INET6:
      for (i = 0; i < al->count; i++)
        {
          ip_address *cur = al->addresses + i;
          if (cur->family == AF_INET6
#ifdef HAVE_SOCKADDR_IN6_SCOPE_ID
              && cur->ipv6_scope == ip->ipv6_scope
#endif
              && IN6_ARE_ADDR_EQUAL (&cur->data.d6, &ip->data.d6))
            return true;
        }
      return false;
#endif /* ENABLE_IPV6 */
    default:
      abort ();
    }
}

/* Mark the INDEXth element of AL as faulty, so that the next time
   this address list is used, the faulty element will be skipped.  */

void
address_list_set_faulty (struct address_list *al, int index)
{
  /* We assume that the address list is traversed in order, so that a
     "faulty" attempt is always preceded with all-faulty addresses,
     and this is how Wget uses it.  */
  assert (index == al->faulty);

  ++al->faulty;
  if (al->faulty >= al->count)
    /* All addresses have been proven faulty.  Since there's not much
       sense in returning the user an empty address list the next
       time, we'll rather make them all clean, so that they can be
       retried anew.  */
    al->faulty = 0;
}

/* Set the "connected" flag to true.  This flag used by connect.c to
   see if the host perhaps needs to be resolved again.  */

void
address_list_set_connected (struct address_list *al)
{
  al->connected = true;
}

/* Return the value of the "connected" flag. */

bool
address_list_connected_p (const struct address_list *al)
{
  return al->connected;
}

#ifdef ENABLE_IPV6

/* Create an address_list from the addresses in the given struct
   addrinfo.  */

static struct address_list *
address_list_from_addrinfo (const struct addrinfo *ai)
{
  struct address_list *al;
  const struct addrinfo *ptr;
  int cnt;
  ip_address *ip;

  cnt = 0;
  for (ptr = ai; ptr != NULL ; ptr = ptr->ai_next)
    if (ptr->ai_family == AF_INET || ptr->ai_family == AF_INET6)
      ++cnt;
  if (cnt == 0)
    return NULL;

  al = xnew0 (struct address_list);
  al->addresses = xnew_array (ip_address, cnt);
  al->count     = cnt;
  al->refcount  = 1;

  ip = al->addresses;
  for (ptr = ai; ptr != NULL; ptr = ptr->ai_next)
    if (ptr->ai_family == AF_INET6) 
      {
        const struct sockaddr_in6 *sin6 =
          (const struct sockaddr_in6 *)ptr->ai_addr;
        ip->family = AF_INET6;
        ip->data.d6 = sin6->sin6_addr;
#ifdef HAVE_SOCKADDR_IN6_SCOPE_ID
        ip->ipv6_scope = sin6->sin6_scope_id;
#endif
        ++ip;
      } 
    else if (ptr->ai_family == AF_INET)
      {
        const struct sockaddr_in *sin =
          (const struct sockaddr_in *)ptr->ai_addr;
        ip->family = AF_INET;
        ip->data.d4 = sin->sin_addr;
        ++ip;
      }
  assert (ip - al->addresses == cnt);
  return al;
}

#define IS_IPV4(addr) (((const ip_address *) addr)->family == AF_INET)

/* Compare two IP addresses by family, giving preference to the IPv4
   address (sorting it first).  In other words, return -1 if ADDR1 is
   IPv4 and ADDR2 is IPv6, +1 if ADDR1 is IPv6 and ADDR2 is IPv4, and
   0 otherwise.

   This is intended to be used as the comparator arg to a qsort-like
   sorting function, which is why it accepts generic pointers.  */

static int
cmp_prefer_ipv4 (const void *addr1, const void *addr2)
{
  return !IS_IPV4 (addr1) - !IS_IPV4 (addr2);
}

#define IS_IPV6(addr) (((const ip_address *) addr)->family == AF_INET6)

/* Like the above, but give preference to the IPv6 address.  */

static int
cmp_prefer_ipv6 (const void *addr1, const void *addr2)
{
  return !IS_IPV6 (addr1) - !IS_IPV6 (addr2);
}

#else  /* not ENABLE_IPV6 */

/* Create an address_list from a NULL-terminated vector of IPv4
   addresses.  This kind of vector is returned by gethostbyname.  */

static struct address_list *
address_list_from_ipv4_addresses (char **vec)
{
  int count, i;
  struct address_list *al = xnew0 (struct address_list);

  count = 0;
  while (vec[count])
    ++count;
  assert (count > 0);

  al->addresses = xnew_array (ip_address, count);
  al->count     = count;
  al->refcount  = 1;

  for (i = 0; i < count; i++)
    {
      ip_address *ip = &al->addresses[i];
      ip->family = AF_INET;
      memcpy (IP_INADDR_DATA (ip), vec[i], 4);
    }

  return al;
}

#endif /* not ENABLE_IPV6 */

static void
address_list_delete (struct address_list *al)
{
  xfree (al->addresses);
  xfree (al);
}

/* Mark the address list as being no longer in use.  This will reduce
   its reference count which will cause the list to be freed when the
   count reaches 0.  */

void
address_list_release (struct address_list *al)
{
  --al->refcount;
  DEBUGP (("Releasing 0x%0*lx (new refcount %d).\n", PTR_FORMAT (al),
           al->refcount));
  if (al->refcount <= 0)
    {
      DEBUGP (("Deleting unused 0x%0*lx.\n", PTR_FORMAT (al)));
      address_list_delete (al);
    }
}

/* Versions of gethostbyname and getaddrinfo that support timeout. */

#ifndef ENABLE_IPV6

struct ghbnwt_context {
  const char *host_name;
  struct hostent *hptr;
};

static void
gethostbyname_with_timeout_callback (void *arg)
{
  struct ghbnwt_context *ctx = (struct ghbnwt_context *)arg;
  ctx->hptr = gethostbyname (ctx->host_name);
}

/* Just like gethostbyname, except it times out after TIMEOUT seconds.
   In case of timeout, NULL is returned and errno is set to ETIMEDOUT.
   The function makes sure that when NULL is returned for reasons
   other than timeout, errno is reset.  */

static struct hostent *
gethostbyname_with_timeout (const char *host_name, double timeout)
{
  struct ghbnwt_context ctx;
  ctx.host_name = host_name;
  if (run_with_timeout (timeout, gethostbyname_with_timeout_callback, &ctx))
    {
      SET_H_ERRNO (HOST_NOT_FOUND);
      errno = ETIMEDOUT;
      return NULL;
    }
  if (!ctx.hptr)
    errno = 0;
  return ctx.hptr;
}

/* Print error messages for host errors.  */
static char *
host_errstr (int error)
{
  /* Can't use switch since some of these constants can be equal,
     which makes the compiler complain about duplicate case
     values.  */
  if (error == HOST_NOT_FOUND
      || error == NO_RECOVERY
      || error == NO_DATA
      || error == NO_ADDRESS)
    return _("Unknown host");
  else if (error == TRY_AGAIN)
    /* Message modeled after what gai_strerror returns in similar
       circumstances.  */
    return _("Temporary failure in name resolution");
  else
    return _("Unknown error");
}

#else  /* ENABLE_IPV6 */

struct gaiwt_context {
  const char *node;
  const char *service;
  const struct addrinfo *hints;
  struct addrinfo **res;
  int exit_code;
};

static void
getaddrinfo_with_timeout_callback (void *arg)
{
  struct gaiwt_context *ctx = (struct gaiwt_context *)arg;
  ctx->exit_code = getaddrinfo (ctx->node, ctx->service, ctx->hints, ctx->res);
}

/* Just like getaddrinfo, except it times out after TIMEOUT seconds.
   In case of timeout, the EAI_SYSTEM error code is returned and errno
   is set to ETIMEDOUT.  */

static int
getaddrinfo_with_timeout (const char *node, const char *service,
                          const struct addrinfo *hints, struct addrinfo **res,
                          double timeout)
{
  struct gaiwt_context ctx;
  ctx.node = node;
  ctx.service = service;
  ctx.hints = hints;
  ctx.res = res;

  if (run_with_timeout (timeout, getaddrinfo_with_timeout_callback, &ctx))
    {
      errno = ETIMEDOUT;
      return EAI_SYSTEM;
    }
  return ctx.exit_code;
}

#endif /* ENABLE_IPV6 */

/* Return a textual representation of ADDR, i.e. the dotted quad for
   IPv4 addresses, and the colon-separated list of hex words (with all
   zeros omitted, etc.) for IPv6 addresses.  */

const char *
print_address (const ip_address *addr)
{
#ifdef ENABLE_IPV6
  static char buf[64];
  if (!inet_ntop (addr->family, IP_INADDR_DATA (addr), buf, sizeof buf))
    snprintf (buf, sizeof buf, "<error: %s>", strerror (errno));
  return buf;
#else
  return inet_ntoa (addr->data.d4);
#endif
}

/* The following two functions were adapted from glibc's
   implementation of inet_pton, written by Paul Vixie. */

static bool
is_valid_ipv4_address (const char *str, const char *end)
{
  bool saw_digit = false;
  int octets = 0;
  int val = 0;

  while (str < end)
    {
      int ch = *str++;

      if (ch >= '0' && ch <= '9')
        {
          val = val * 10 + (ch - '0');

          if (val > 255)
            return false;
          if (!saw_digit)
            {
              if (++octets > 4)
                return false;
              saw_digit = true;