/* $OpenBSD: match.c,v 1.46 2026/05/31 04:19:16 djm Exp $ */

/*

 * Author: Tatu Ylonen <ylo@cs.hut.fi>

 * Copyright (c) 1995 Tatu Ylonen <ylo@cs.hut.fi>, Espoo, Finland

 *                    All rights reserved

 * Simple pattern matching, with '*' and '?' as wildcards.

 *

 * As far as I am concerned, the code I have written for this software

 * can be used freely for any purpose.  Any derived versions of this

 * software must be clearly marked as such, and if the derived work is

 * incompatible with the protocol description in the RFC file, it must be

 * called by a name other than "ssh" or "Secure Shell".

 */

/*

 * Copyright (c) 2000 Markus Friedl.  All rights reserved.

 * Copyright (c) 2026 Damien Miller.  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 ``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 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.

 */

#include "includes.h"

#include <sys/types.h>

#include <ctype.h>

#include <stdlib.h>

#include <string.h>

#include <stdarg.h>

#include <stdio.h>

#include "xmalloc.h"

#include "match.h"

#include "misc.h"

/*

 * Computes the epsilon closure of an NFA set.

 * In our wildcard grammar, epsilon transitions only exist for '*' wildcards,

 * allowing us to transition from state i to i+1 without consuming input.

 *

 * This function modifies 'states' in place.

 */

static void

epsilon_closure(char *states, const char *pattern, size_t M)

{

  size_t i;

  /* only need a forward pass as there are no back jumps in our grammar */

  for (i = 0; i < M; i++) {

    if (!states[i] || pattern[i] != '*')

      continue;

    /*

     * State i is active, and pattern[i] is '*', so we can

     * epsilon-transition to i+1.

     */

    states[i + 1] = 1;

  }

}

/*

 * Returns true if the given string matches the pattern (which may contain ?

 * and * as wildcards), and zero if it does not match. Uses an NFA internally.

 */

int

match_pattern(const char *s, const char *pattern)

{

  size_t M;

  size_t i;

  char *states, *next_states, *tmp;

  int active, matched = 0;

  /* trivial case: empty pattern vs empty input */

  if ((M = strlen(pattern)) == 0)

    return *s == '\0';

  /* A state for each pattern character, plus one final accepting state */

  states = xcalloc(M + 1, sizeof(*states));

  next_states = xcalloc(M + 1, sizeof(*next_states));

  /* Initial state: state 0 is active */

  states[0] = 1;

  /* Other states might be reachable now if the pattern starts with '*' */

  epsilon_closure(states, pattern, M);

  for (; *s; s++) {

    memset(next_states, 0, M + 1);

    /* Calculate the reachable next states given the input char */

    for (i = 0; i < M; i++) {

      if (!states[i])

        continue;

      if (pattern[i] == '*') {

        /*

         * '*' matches any character, so we can

         * stay in state i

         */

        next_states[i] = 1;

      } else if (pattern[i] == '?' || pattern[i] == *s) {

        /*

         * '?' matches any character, or we have

         * a literal match.

         */

        next_states[i + 1] = 1;

      }

    }

    /* Expand the reachable next states with epsilon transitions */

    epsilon_closure(next_states, pattern, M);

    /* Swap states and next_states */

    tmp = states;

    states = next_states;

    next_states = tmp;

    /* Check if we have any active pattern states left */

    active = 0;

    for (i = 0; i <= M; i++) {

      if (states[i]) {

        active = 1;

        break;

      }

    }

    if (!active)

      goto out; /* No active states, fail early */

  }

  /*

   * We matched only if we ended up in the final, accepting state

   * after consuming all the input.

   */

  matched = states[M];

 out:

  free(states);

  free(next_states);

  return matched;

}

/*

 * Tries to match the string against the

 * comma-separated sequence of subpatterns (each possibly preceded by ! to

 * indicate negation).  Returns -1 if negation matches, 1 if there is

 * a positive match, 0 if there is no match at all.

 */

int

match_pattern_list(const char *string, const char *pattern, int dolower)

{

  char sub[1024];

  int negated;

  int got_positive;

  u_int i, subi, len = strlen(pattern);

  got_positive = 0;

  for (i = 0; i < len;) {

    /* Check if the subpattern is negated. */

    if (pattern[i] == '!') {

      negated = 1;

      i++;

    } else

      negated = 0;

    /*

     * Extract the subpattern up to a comma or end.  Convert the

     * subpattern to lowercase.

     */

    for (subi = 0;

        i < len && subi < sizeof(sub) - 1 && pattern[i] != ',';

        subi++, i++)

      sub[subi] = dolower && isupper((u_char)pattern[i]) ?

          tolower((u_char)pattern[i]) : pattern[i];

    /* If subpattern too long, return failure (no match). */

    if (subi >= sizeof(sub) - 1)

      return 0;

    /* If the subpattern was terminated by a comma, then skip it. */

    if (i < len && pattern[i] == ',')

      i++;

    /* Null-terminate the subpattern. */

    sub[subi] = '\0';

    /* Try to match the subpattern against the string. */

    if (match_pattern(string, sub)) {

      if (negated)

        return -1;   /* Negative */

      else

        got_positive = 1; /* Positive */

    }

  }

  /*

   * Return success if got a positive match.  If there was a negative

   * match, we have already returned -1 and never get here.

   */

  return got_positive;

}

/* Match a list representing users or groups. */

int

match_usergroup_pattern_list(const char *string, const char *pattern)

{

#ifdef HAVE_CYGWIN

  /* Windows usernames may be Unicode and are not case sensitive */

  return cygwin_ug_match_pattern_list(string, pattern);

#else

  /* Case sensitive match */

  return match_pattern_list(string, pattern, 0);

#endif

}

/*

 * Tries to match the host name (which must be in all lowercase) against the

 * comma-separated sequence of subpatterns (each possibly preceded by ! to

 * indicate negation).  Returns -1 if negation matches, 1 if there is

 * a positive match, 0 if there is no match at all.

 */

int

match_hostname(const char *host, const char *pattern)

{

  char *hostcopy = xstrdup(host);

  int r;

  lowercase(hostcopy);

  r = match_pattern_list(hostcopy, pattern, 1);

  free(hostcopy);

  return r;

}

/*

 * returns 0 if we get a negative match for the hostname or the ip

 * or if we get no match at all.  returns -1 on error, or 1 on

 * successful match.

 */

int

match_host_and_ip(const char *host, const char *ipaddr,

    const char *patterns)

{

  int mhost, mip;

  if ((mip = addr_match_list(ipaddr, patterns)) == -2)

    return -1; /* error in ipaddr match */

  else if (host == NULL || ipaddr == NULL || mip == -1)

    return 0; /* negative ip address match, or testing pattern */

  /* negative hostname match */

  if ((mhost = match_hostname(host, patterns)) == -1)

    return 0;

  /* no match at all */

  if (mhost == 0 && mip == 0)

    return 0;

  return 1;

}

/*

 * Match user, user@host_or_ip, user@host_or_ip_list against pattern.

 * If user, host and ipaddr are all NULL then validate pattern/

 * Returns -1 on invalid pattern, 0 on no match, 1 on match.

 */

int

match_user(const char *user, const char *host, const char *ipaddr,

    const char *pattern)

{

  char *p, *pat;

  int ret;

  /* test mode */

  if (user == NULL && host == NULL && ipaddr == NULL) {

    if ((p = strrchr(pattern, '@')) != NULL &&

        match_host_and_ip(NULL, NULL, p + 1) < 0)

      return -1;

    return 0;

  }

  if (user == NULL)

    return 0; /* shouldn't happen */

  if (strrchr(pattern, '@') == NULL)

    return match_pattern(user, pattern);

  pat = xstrdup(pattern);

  p = strrchr(pat, '@');

  *p++ = '\0';

  if ((ret = match_pattern(user, pat)) == 1)

    ret = match_host_and_ip(host, ipaddr, p);

  free(pat);

  return ret;

}

/*

 * Returns first item from client-list that is also supported by server-list,

 * caller must free the returned string.

 */

#define MAX_PROP  40

#define SEP ","

char *

match_list(const char *client, const char *server, u_int *next)

{

  char *sproposals[MAX_PROP];

  char *c, *s, *p, *ret, *cp, *sp;

  int i, j, nproposals;

  c = cp = xstrdup(client);

  s = sp = xstrdup(server);

  for ((p = strsep(&sp, SEP)), i=0; p && *p != '\0';

      (p = strsep(&sp, SEP)), i++) {

    if (i < MAX_PROP)

      sproposals[i] = p;

    else

      break;

  }

  nproposals = i;

  for ((p = strsep(&cp, SEP)), i=0; p && *p != '\0';

      (p = strsep(&cp, SEP)), i++) {

    for (j = 0; j < nproposals; j++) {

      if (strcmp(p, sproposals[j]) == 0) {

        ret = xstrdup(p);

        if (next != NULL)

          *next = (cp == NULL) ?

              strlen(c) : (u_int)(cp - c);

        free(c);

        free(s);

        return ret;

      }

    }

  }

  if (next != NULL)

    *next = strlen(c);

  free(c);

  free(s);

  return NULL;

}

/*

 * Filter proposal using pattern-list filter.

 * "denylist" determines sense of filter:

 * non-zero indicates that items matching filter should be excluded.

 * zero indicates that only items matching filter should be included.

 * returns NULL on allocation error, otherwise caller must free result.

 */

static char *

filter_list(const char *proposal, const char *filter, int denylist)

{

  size_t len = strlen(proposal) + 1;

  char *fix_prop = malloc(len);

  char *orig_prop = strdup(proposal);

  char *cp, *tmp;

  int r;

  if (fix_prop == NULL || orig_prop == NULL) {

    free(orig_prop);

    free(fix_prop);

    return NULL;

  }

  tmp = orig_prop;

  *fix_prop = '\0';

  while ((cp = strsep(&tmp, ",")) != NULL) {

    r = match_pattern_list(cp, filter, 0);

    if ((denylist && r != 1) || (!denylist && r == 1)) {

      if (*fix_prop != '\0')

        strlcat(fix_prop, ",", len);

      strlcat(fix_prop, cp, len);

    }

  }

  free(orig_prop);

  return fix_prop;

}

/*

 * Filters a comma-separated list of strings, excluding any entry matching

 * the 'filter' pattern list. Caller must free returned string.

 */

char *

match_filter_denylist(const char *proposal, const char *filter)

{

  return filter_list(proposal, filter, 1);

}

/*

 * Filters a comma-separated list of strings, including only entries matching

 * the 'filter' pattern list. Caller must free returned string.

 */

char *

match_filter_allowlist(const char *proposal, const char *filter)

{

  return filter_list(proposal, filter, 0);

}