parse.c

使用最广泛的radius的linux的源码

/*
 * parse.c		Parse a policy language
 *
 * Version:	$Id: parse.c,v 1.25 2008/03/07 10:03:35 aland Exp $
 *
 *   This program 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 2 of the License, or
 *   (at your option) any later version.
 *
 *   This program 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 this program; if not, write to the Free Software
 *   Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA
 *
 * Copyright 2004  Alan DeKok <[e-mail]>
 * Copyright 2006  The FreeRADIUS server project
 */

#include <freeradius-devel/ident.h>
RCSID("$Id: parse.c,v 1.25 2008/03/07 10:03:35 aland Exp $")

#include "rlm_policy.h"

#ifdef HAVE_DIRENT_H
#include <dirent.h>

#ifdef HAVE_SYS_STAT_H
#include <sys/stat.h>
#endif
#endif

#include <freeradius-devel/modules.h>

const FR_NAME_NUMBER policy_return_codes[] = {
	{ "reject", RLM_MODULE_REJECT },
	{ "fail", RLM_MODULE_FAIL },
	{ "ok", RLM_MODULE_OK },
	{ "handled", RLM_MODULE_HANDLED },
	{ "invalid", RLM_MODULE_INVALID },
	{ "userlock", RLM_MODULE_USERLOCK },
	{ "notfound", RLM_MODULE_NOTFOUND },
	{ "noop", RLM_MODULE_NOOP },
	{ "updated", RLM_MODULE_UPDATED },
	{ NULL, RLM_MODULE_NUMCODES }
};

/*
 *	Explanations of what the lexical tokens are.
 */
static const FR_NAME_NUMBER policy_explanations[] = {
	{ "invalid input", POLICY_LEX_BAD },
	{ "end of file", POLICY_LEX_EOF },
	{ "end of line", POLICY_LEX_EOL },
	{ "whitespace", POLICY_LEX_WHITESPACE },
	{ "hash mark", POLICY_LEX_HASH },
	{ "left bracket", POLICY_LEX_L_BRACKET },
	{ "right bracket", POLICY_LEX_R_BRACKET },
	{ "{", POLICY_LEX_LC_BRACKET },
	{ "}", POLICY_LEX_RC_BRACKET },
	{ "comma", POLICY_LEX_COMMA },
	{ "logical AND", POLICY_LEX_L_AND },
	{ "logical OR", POLICY_LEX_L_OR },
	{ "AND", POLICY_LEX_AND },
	{ "OR", POLICY_LEX_OR },
	{ "logical NOT", POLICY_LEX_L_NOT },
	{ "assignment", POLICY_LEX_ASSIGN },
	{ "comparison", POLICY_LEX_CMP_EQUALS },
	{ "comparison", POLICY_LEX_CMP_NOT_EQUALS },
	{ "comparison", POLICY_LEX_LT },
	{ "comparison", POLICY_LEX_GT },
	{ "comparison", POLICY_LEX_LE },
	{ "comparison", POLICY_LEX_GT },
	{ "comparison", POLICY_LEX_RX_EQUALS },
	{ "comparison", POLICY_LEX_RX_NOT_EQUALS },
	{ "double quoted string", POLICY_LEX_DOUBLE_QUOTED_STRING },
	{ "single quoted string", POLICY_LEX_SINGLE_QUOTED_STRING },
	{ "back quoted string", POLICY_LEX_BACK_QUOTED_STRING },
	{ "bare word", POLICY_LEX_BARE_WORD },

	{ NULL, -1 }
};


const FR_NAME_NUMBER rlm_policy_tokens[] = {
	{ "EOF", POLICY_LEX_EOF },
	{ "#", POLICY_LEX_HASH },
	{ "(", POLICY_LEX_L_BRACKET },
	{ ")", POLICY_LEX_R_BRACKET },
	{ "{", POLICY_LEX_LC_BRACKET },
	{ "}", POLICY_LEX_RC_BRACKET },
	{ ",", POLICY_LEX_COMMA },
	{ "&&", POLICY_LEX_L_AND },
	{ "||", POLICY_LEX_L_OR },
	{ "&", POLICY_LEX_AND },
	{ "|", POLICY_LEX_OR },
	{ "!", POLICY_LEX_L_NOT },
	{ "=", POLICY_LEX_ASSIGN },
	{ "==", POLICY_LEX_CMP_EQUALS },
	{ "!=", POLICY_LEX_CMP_NOT_EQUALS },
	{ "=*", POLICY_LEX_CMP_TRUE },
	{ "!*", POLICY_LEX_CMP_FALSE },
	{ "<", POLICY_LEX_LT },
	{ ">", POLICY_LEX_GT },
	{ "<=", POLICY_LEX_LE },
	{ ">=", POLICY_LEX_GT },
	{ "=~", POLICY_LEX_RX_EQUALS },
	{ "!~", POLICY_LEX_RX_NOT_EQUALS },
	{ "^=", POLICY_LEX_BEFORE_HEAD_ASSIGN },
	{ "^==", POLICY_LEX_BEFORE_WHERE_ASSIGN },
	{ "^.", POLICY_LEX_BEFORE_HEAD_EQUALS },
	{ "^.=", POLICY_LEX_BEFORE_WHERE_EQUALS },
	{ "$=", POLICY_LEX_AFTER_TAIL_ASSIGN },
	{ "$==", POLICY_LEX_AFTER_WHERE_ASSIGN },
	{ "$.", POLICY_LEX_AFTER_TAIL_EQUALS },
	{ "$.=", POLICY_LEX_AFTER_WHERE_EQUALS },
	{ ".=", POLICY_LEX_CONCAT_EQUALS },
	{ ":=", POLICY_LEX_SET_EQUALS },
	{ "double quoted string", POLICY_LEX_DOUBLE_QUOTED_STRING },
	{ "single quoted string", POLICY_LEX_SINGLE_QUOTED_STRING },
	{ "back quoted string", POLICY_LEX_BACK_QUOTED_STRING },
	{ "bare word", POLICY_LEX_BARE_WORD },

	{ NULL, -1 }
};


/*
 *	Hand-coded lexical analysis of a string.
 *	Handed input string, updates token, possible a decoded
 *	string in buffer, and returns the pointer to the next token.
 *
 *	Lexical tokens cannot cross a string boundary.
 */
static const char *policy_lex_string(const char *input,
				     policy_lex_t *token,
				     char *buffer, size_t buflen)
{
	rad_assert(input != NULL);

	if (buffer) *buffer = '\0';

	switch (*input) {
	case '\0':
		*token = POLICY_LEX_EOL;
		return NULL;	/* nothing more to do */

	case ' ':
	case '\t':
	case '\r':
	case '\n':
		/*
		 *	Skip over all of the whitespace in one swell foop.
		 */
		*token = POLICY_LEX_WHITESPACE;
		while ((*input == ' ') || (*input == '\t') ||
		       (*input == '\r') || (*input == '\n')) input++;
		return input;	/* point to next non-whitespace character */

	case '#':		/* ignore everything to the end of the line */
		*token = POLICY_LEX_EOL;
		return NULL;

	case '(':
		*token = POLICY_LEX_L_BRACKET;
		return input + 1;

	case ')':
		*token = POLICY_LEX_R_BRACKET;
		return input + 1;

	case '{':
		*token = POLICY_LEX_LC_BRACKET;
		return input + 1;

	case '}':
		*token = POLICY_LEX_RC_BRACKET;
		return input + 1;

	case ',':
		*token = POLICY_LEX_COMMA;
		return input + 1;

	case '+':
		switch (input[1]) {
		case '=':
			*token = POLICY_LEX_PLUS_EQUALS;
			input++;
			break;

		default:
			*token = POLICY_LEX_PLUS;
			break;
		}
		return input + 1;

	case '-':
		switch (input[1]) {
		case '=':
			*token = POLICY_LEX_MINUS_EQUALS;
			input++;
			break;

		default:
			*token = POLICY_LEX_MINUS;
			break;
		}
		return input + 1;

	case '.':
		if (input[1] == '=') {
			*token = POLICY_LEX_CONCAT_EQUALS;
			return input + 2;
		}
		*token = POLICY_LEX_BAD;
		return input + 1;
	
	case '^':
		if (input[1] == '.' ) {
			if (input[2] == '=') {
				*token = POLICY_LEX_BEFORE_WHERE_EQUALS;
				return input + 3;
			}
			else {
				*token = POLICY_LEX_BEFORE_HEAD_EQUALS;
				return input + 2;
			}
		}
		else if (input[1] == '=') {
			if (input[2] == '=') {
				*token = POLICY_LEX_BEFORE_WHERE_ASSIGN;
				return input + 3;
			}
			else {
				*token = POLICY_LEX_BEFORE_HEAD_ASSIGN;
				return input + 2;
			}
		}
			
		*token = POLICY_LEX_BAD;
		return input + 1;

	case '$':
		if (input[1] == '.' ) {
			if (input[2] == '=') {
				*token = POLICY_LEX_AFTER_WHERE_EQUALS;
				return input + 3;
			}
			else {
				*token = POLICY_LEX_AFTER_TAIL_EQUALS;
				return input + 2;
			}
		}
		else if (input[1] == '=') {
			if (input[2] == '=') {
				*token = POLICY_LEX_AFTER_WHERE_ASSIGN;
				return input + 3;
			}
			else {
				*token = POLICY_LEX_AFTER_TAIL_ASSIGN;
				return input + 2;
			}
		}

		*token = POLICY_LEX_BAD;
		return input + 1;

	case ':':
		if (input[1] == '=') {
			*token = POLICY_LEX_SET_EQUALS;
			return input + 2;
		}
		*token = POLICY_LEX_BAD;
		return input + 1;

	case '&':
		switch (input[1]) {
		case '&':
			*token = POLICY_LEX_L_AND;
			input++;
			break;

		case '=':
			*token = POLICY_LEX_AND_EQUALS;
			input++;
			break;

		default:
			*token = POLICY_LEX_AND;
		}
		return input + 1;

	case '|':
		switch (input[1]) {
		case '|':
			*token = POLICY_LEX_L_OR;
			input++;
			break;

		case '=':
			*token = POLICY_LEX_OR_EQUALS;
			input++;
			break;

		default:
			*token = POLICY_LEX_OR;
		}
		return input + 1;

	case '!':
		switch (input[1]) {
		case '=':
			input++;
			*token = POLICY_LEX_CMP_NOT_EQUALS;
			break;

		case '~':
			input++;
			*token = POLICY_LEX_RX_NOT_EQUALS;
			break;

		case '*':
			input++;
			*token = POLICY_LEX_CMP_FALSE;
			break;

		default:
			*token = POLICY_LEX_L_NOT;
		}
		return input + 1;

	case '=':
		switch (input[1]) {
		case '=':
			input++;
			*token = POLICY_LEX_CMP_EQUALS;
			break;

		case '~':
			input++;
			*token = POLICY_LEX_RX_EQUALS;
			break;

		case '*':
			input++;
			*token = POLICY_LEX_CMP_TRUE;
			break;

		default:
			*token = POLICY_LEX_ASSIGN;
		}
		return input + 1;

	case '<':
		if (input[1] == '=') {
			input++;
			*token = POLICY_LEX_LE;
		} else {
			*token = POLICY_LEX_LT;
		}
		return input + 1;

	case '>':
		if (input[1] == '=') {
			input++;
			*token = POLICY_LEX_GE;
		} else {
			*token = POLICY_LEX_GT;
		}
		return input + 1;

	case '"':
		if (buflen < 2) {
			*token = POLICY_LEX_BAD;
			return input + 1;
		}

		input++;
		while (*input != '"') {
			/*
			 *	Strings can't pass EOL.
			 */
			if (!*input) {
				return POLICY_LEX_BAD;
			}

			/*
			 *	FIXME: Embedded quotes?
			 */
			*(buffer++) = *(input++);
			buflen--;

			/*
			 *	FIXME: Print more warnings?
			 */
			if (buflen == 1) {
				break;
			}
		}
		*buffer = '\0';

		*token = POLICY_LEX_DOUBLE_QUOTED_STRING;
		return input + 1; /* skip trailing '"' */

	default:		/* bare word */
		break;
	}

	/*
	 *	It's a bare word, with nowhere to put it.  Die.
	 */
	if (!buffer) {
		*token = POLICY_LEX_BAD;
		return input + 1;
	}

	/*
	 *	Getting one character is stupid.
	 */
	if (buflen < 2) {
		*token = POLICY_LEX_BAD;
		return input + 1;
	}

	/*
	 *	Bare words are [-a-zA-Z0-9.]+
	 */
	while (*input) {
		if (!(((*input >= '0') && (*input <= '9')) ||
		      ((*input >= 'a') && (*input <= 'z')) ||
		      ((*input >= 'A') && (*input <= 'Z')) ||
		      (*input == '-') || (*input == '.') ||
		      (*input == ':') || (*input == '_'))) {
			break;
		}
		*(buffer++) = *(input++);
		buflen--;

		/*
		 *	FIXME: Print more warnings?
		 */
		if (buflen == 1) {
			break;
		}
	}
	*buffer = '\0';

	*token = POLICY_LEX_BARE_WORD;
	return input;
}


/*
 *	We want to lexically analyze a file, so we need a wrapper
 *	around the lexical analysis of strings.
 */
typedef struct policy_lex_file_t {
	FILE		*fp;
	const char	*parse;
	const char	*filename;
	int		lineno;
	int		debug;
	rbtree_t	*policies;
	policy_lex_t	token;
	char		buffer[1024];
} policy_lex_file_t;


#define POLICY_LEX_FLAG_RETURN_EOL  (1 << 0)
#define POLICY_LEX_FLAG_PEEK        (1 << 1)
#define POLICY_LEX_FLAG_PRINT_TOKEN (1 << 2)

#define debug_tokens if ((lexer->debug & POLICY_DEBUG_PRINT_TOKENS) && fr_log_fp) fr_printf_log


/*
 *	Function to return a token saying what it read, and possibly
 *	a buffer of the quoted string or bare word.
 */
static policy_lex_t policy_lex_file(policy_lex_file_t *lexer,
				    int flags,
				    char *mystring, size_t mystringlen)
{
	policy_lex_t token = POLICY_LEX_BARE_WORD; /* to prime it */

	if (lexer->debug & POLICY_DEBUG_PRINT_TOKENS) {
		flags |= POLICY_LEX_FLAG_PRINT_TOKEN;
	}

	if (!lexer->fp) {
		return POLICY_LEX_EOF;
	}

	/*
	 *	Starting off, the buffer needs to be primed.
	 */
	if (!lexer->parse) {
		lexer->parse = fgets(lexer->buffer,
				     sizeof(lexer->buffer),
				     lexer->fp);

		if (!lexer->parse) {
			return POLICY_LEX_EOF;
		}

		lexer->lineno = 1;
	} /* buffer is primed, read stuff */

	if (lexer->token != POLICY_LEX_BAD) {
		token = lexer->token;
		lexer->token = POLICY_LEX_BAD;
		return token;
	}

	/*
	 *	Ignore whitespace, and keep filling the buffer
	 */
	while (lexer->parse) {
		const char *next;

		next = policy_lex_string(lexer->parse, &token,
					 mystring, mystringlen);
		switch (token) {
		case POLICY_LEX_WHITESPACE: /* skip whitespace */
			lexer->parse = next;
			continue;

		case POLICY_LEX_EOL: /* read another line */
			lexer->parse = fgets(lexer->buffer,
					     sizeof(lexer->buffer),
					     lexer->fp);
			lexer->lineno++;
			if (flags & POLICY_LEX_FLAG_RETURN_EOL) {
				return POLICY_LEX_EOL;
			}
			break;	/* read another token */

		default:	/* return the token */
			if (!(flags & POLICY_LEX_FLAG_PEEK)) {
				lexer->parse = next;
			}
			if (flags & POLICY_LEX_FLAG_PRINT_TOKEN) {
				debug_tokens("[%s token %s] ",
					     (flags & POLICY_LEX_FLAG_PEEK) ? "peek " : "",
					     fr_int2str(rlm_policy_tokens,
							  token, "?"));
			}
			return token;
			break;
		}
	} /* loop until EOF */

	/*
	 *	Close it for the user.
	 */
	fclose(lexer->fp);
	lexer->fp = NULL;

	return POLICY_LEX_EOF;
}


/*
 *	Push a token back onto the input.
 *
 *	FIXME: Push words, too?