pair_tokenize.c

Go to the documentation of this file.

/*
 *   This library is free software; you can redistribute it and/or
 *   modify it under the terms of the GNU Lesser General Public
 *   License as published by the Free Software Foundation; either
 *   version 2.1 of the License, or (at your option) any later version.
 *
 *   This library 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
 *   Lesser General Public License for more details.
 *
 *   You should have received a copy of the GNU Lesser General Public
 *   License along with this library; if not, write to the Free Software
 *   Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301, USA
 */
 
/** AVP parsing
 *
 * @file src/lib/util/pair_tokenize.c
 *
 * @copyright 2020 Network RADIUS SAS (legal@networkradius.com)
 */
RCSID("$Id: 08b15f75098ddcd4de97f865174a80d643ad41e6 $")
 
#include <freeradius-devel/util/skip.h>
#include <freeradius-devel/util/pair.h>
#include <freeradius-devel/util/proto.h>
#include <freeradius-devel/util/regex.h>
 
#include <freeradius-devel/protocol/radius/rfc2865.h>
#include <freeradius-devel/protocol/freeradius/freeradius.internal.h>
 
 
static ssize_t op_to_token(fr_token_t *token, char const *op, size_t oplen)
{
        char const *p = op;
 
        if (!token || !op || !oplen) return 0;
 
        switch (*p) {
        default:
                fr_strerror_const("Invalid text. Expected comparison operator");
                return -(p - op);
 
        case '!':
                if (oplen < 2) goto invalid_operator;
 
                if (p[1] == '=') {
                        *token = T_OP_NE;
                        p += 2;
 
#ifdef HAVE_REGEX
                } else if (p[1] == '~') {
                        *token = T_OP_REG_NE;
                        p += 2;
#endif
 
                } else if (p[1] == '*') {
                        *token = T_OP_CMP_FALSE;
                        p += 2;
 
                } else {
                invalid_operator:
                        fr_strerror_const("Invalid operator");
                        return -(p - op);
                }
                break;
 
        case '=':
                /*
                 *      Bare '=' is allowed.
                 */
                if (oplen == 1) {
                        *token = T_OP_EQ;
                        p++;
                        break;
                }
 
                if (p[1] == '=') {
                        *token = T_OP_CMP_EQ;
                        p += 2;
 
#ifdef HAVE_REGEX
                } else if (p[1] == '~') {
                        *token = T_OP_REG_EQ;
                        p += 2;
#endif
 
                } else if (p[1] == '*') {
                        *token = T_OP_CMP_TRUE;
                        p += 2;
 
                } else {
                        /*
                         *      Ignore whatever is after the '=' sign.
                         */
                        *token = T_OP_EQ;
                        p++;
                }
                break;
 
        case '<':
                if ((oplen > 1) && (p[1] == '=')) {
                        *token = T_OP_LE;
                        p += 2;
 
                } else {
                        *token = T_OP_LT;
                        p++;
                }
                break;
 
        case '>':
                if ((oplen > 1) && (p[1] == '=')) {
                        *token = T_OP_GE;
                        p += 2;
 
                } else {
                        *token = T_OP_GT;
                        p++;
                }
                break;
        }
 
        return p - op;
}
 
/** Allocate a fr_pair_t based on pre-parsed fields.
 *
 * @param ctx   the talloc ctx
 * @param da    the da for the vp
 * @param op    the operator
 * @param value the value to parse
 * @param value_len length of the value string
 * @param uerules used for unescaping.
 * @return
 *      - fr_pair_t* on success
 *      - NULL on error
 *
 *  It's just better for this function to take the broken-out /
 *  pre-parsed fields.  That way the caller can do any necessary
 *  parsing.
 */
static fr_pair_t *fr_pair_afrom_fields(TALLOC_CTX *ctx, fr_dict_attr_t const *da,
                                        fr_token_t op,
                                        char const *value, size_t value_len,
                                        fr_sbuff_unescape_rules_t const *uerules)
{
        fr_pair_t *vp;
 
        if (!da || !value || (value_len == 0)) return NULL;
 
        vp = fr_pair_afrom_da(ctx, da);
        if (!vp) return NULL;
 
        vp->op = op;
        PAIR_ALLOCED(vp);
 
        if (fr_pair_value_from_str(vp, value, value_len, uerules, false) < 0) {
                talloc_free(vp);
                return NULL;
        }
 
        return vp;
}
 
 
/** Allocate one fr_pair_t from a string, and add it to the pair_ctx cursor.
 *
 * @param[in,out] pair_ctx      the parsing context
 * @param[in] start             Where to create relative error offsets in relation to.
 * @param in    String to parse
 * @param inlen length of string to parse
 * @return
 *      - <= 0 on error (offset as negative integer)
 *      - > 0 on success (number of bytes parsed).
 */
static ssize_t fr_pair_afrom_str(fr_pair_ctx_t *pair_ctx, char const *start, char const *in, size_t inlen)
{
        char const *end = in + inlen;
        char const *p = in;
        ssize_t slen;
        fr_dict_attr_t const *da;
        char quote;
        char const *value;
        size_t value_len;
        fr_pair_t *vp;
        fr_token_t op;
 
        slen = fr_dict_attr_by_name_substr(NULL, &da, pair_ctx->parent, &FR_SBUFF_IN(p, end), NULL);
        if (slen <= 0) return slen - (in - start);
 
        if (da->parent != pair_ctx->parent) {
                fr_strerror_printf("Unexpected attribute %s is not a child of %s",
                                   da->name, pair_ctx->parent->name);
                return -(in - start);
        }
 
        p = in + slen;
        if (p >= end) {
                fr_strerror_const("Attribute name overflows the input buffer");
                return -(in - start);
        }
 
        while ((isspace((uint8_t) *p)) && (p < end)) p++;
 
        if (p >= end) {
                fr_strerror_const("No operator found in the input buffer");
                return -(p - start);
        }
 
        /*
         *      For now, the only allowed operator is equals.
         */
        slen = op_to_token(&op, p, (end - p));
        if (slen <= 0) {
                fr_strerror_const("Syntax error: expected '='");
                return slen - -(p - start);
        }
        p += slen;
 
        while ((isspace((uint8_t) *p)) && (p < end)) p++;
 
        if (p >= end) {
                fr_strerror_const("No value found in the input buffer");
                return -(p - start);
        }
 
        if (*p == '`') {
                fr_strerror_const("Invalid string quotation");
                return -(p - start);
        }
 
        if ((*p == '"') || (*p == '\'')) {
                quote = *p;
                value = p + 1;
 
                slen = fr_skip_string(p, end);
                if (slen <= 0) return slen - (p - start);
                p += slen;
                value_len = slen - 2; /* account for two "" */
 
        } else {
                quote = 0;
                value = p;
 
                /*
                 *      Skip bare words, but end at comma or end-of-buffer.
                 */
                while (!isspace((uint8_t) *p) && (*p != ',') && (p < end)) p++;
 
                value_len = p - value;
        }
 
        if (p > end) {
                fr_strerror_const("Value overflows the input buffer");
                return -(p - start);
        }
 
        vp = fr_pair_afrom_fields(pair_ctx->ctx, da, op, value, value_len, fr_value_unescape_by_char[(uint8_t)quote]);
        if (!vp) return -(in - start);
 
        PAIR_ALLOCED(vp);
        fr_pair_append(pair_ctx->list, vp);
 
        return p - start;
}
 
/** Set a new DA context based on the input string
 *
 * @param[in,out] pair_ctx      the parsing context
 * @param[in] in        String to parse
 * @param[in] inlen     length of string to parse
 * @return
 *      - <= 0 on error (offset as negative integer)
 *      - > 0 on success (number of bytes parsed).
 *
 *  pair_ctx->da is set to the new parsing context.
 *
 *  @todo - allow for child contexts, so that we can parse TLVs into vp->vp_children.
 *          This change requires nested cursors, but not necessarily nested contexts.
 *          We probably want to have a `fr_dlist_t` of pair_ctx, and we always
 *          operate on the last one.  When we change contexts to an attributes parent,
 *          we also change pair_ctx to a parent context.  This is like the da stack,
 *          but with child cursors, too.
 *
 *          We also want to emulate the previous behavior of group attributes based
 *          on parent, and increasing child_num.  i.e. if we're parsing a series of
 *          "attr-foo = bar", then we watch the parent context, and create a new
 *          parent VP if this child has a SMALLER attribute number than the previous
 *          child.  This allows the previous configurations to "just work".
 */
static ssize_t fr_pair_ctx_set(fr_pair_ctx_t *pair_ctx, char const *in, size_t inlen)
{
        char const *end = in + inlen;
        char const *p = in;
        ssize_t slen;
        fr_dict_attr_t const *da;
        fr_dict_attr_t const *parent = pair_ctx->parent;
 
        /*
         *      Parse the attribute name.
         */
        while (p < end) {
                slen = fr_dict_attr_by_name_substr(NULL, &da, parent, &FR_SBUFF_IN(p, end), NULL);
                if (slen <= 0) return slen - (p - in);
 
                if (da->parent != parent) {
                        fr_strerror_printf("Unexpected attribute %s is not a child of %s",
                                           da->name, parent->name);
                        return -(p - in);
                }
 
                if ((p + slen) > end) {
                        fr_strerror_const("Attribute name overflows the input buffer");
                        return -(p - in);
                }
 
                /*
                 *      Check for ending conditions.
                 */
                p += slen;
                parent = da;
 
                if ((p >= end) || (*p == ',')) {
                        break;
                }
 
                /*
                 *      We now MUST have FOO.BAR
                 */
                if (*p != '.') {
                        fr_strerror_const("Unexpected text after attribute");
                        return -(p - in);
                }
                p++;
        }
 
        pair_ctx->parent = parent;
        return p - in;
}
 
 
/** Parse a pair context from a string.
 *
 * @param pair_ctx      the parsing context
 * @param in    String to parse
 * @param inlen length of string to parse
 * @return
 *      - <= 0 on error (offset as negative integer)
 *      - > 0 on success (number of bytes parsed).
 *
 *  This function will parse fr_pair_ts, or context changes, up to
 *  end of string, or a trailing ','.  The caller is responsible for
 *  parsing the comma.
 *
 *  It accepts the following syntax:
 *
 *  - Attribute = value
 *      * reset to a new top-level context according to the parent of Attribute
 *      * parse the attribute and the value
 *
 *  - .Attribute = value
 *      * parse the attribute and the value in the CURRENT top-level context
 *
 *  - .Attribute
 *      * reset to a new context according to Attribute, relative to the current context
 *
 *  - ..Attribute
 *      * reset to a new context according to Attribute, relative to the current context
 *      * more '.' will walk back up the context tree.
 *
 *  - Attribute
 *      * reset to a new top-level context according to Attribute
 *
 */
ssize_t fr_pair_ctx_afrom_str(fr_pair_ctx_t *pair_ctx, char const *in, size_t inlen)
{
        char const *end = in + inlen;
        char const *p = in;
        ssize_t slen;
        fr_dict_attr_t const *da;
 
        while (isspace((uint8_t) *p) && (p < end)) p++;
        if (p >= end) return end - in;
 
        /*
         *      There may be one or more leading '.'
         */
        if (*p == '.') {
 
                /*
                 *      .ATTRIBUTE = VALUE
                 */
                if (p[1] != '.') {
                        p++;
                        return fr_pair_afrom_str(pair_ctx, in, p, end - p);
                }
 
                /*
                 *      '.' is our parent.
                 */
                da = pair_ctx->parent;
 
                /*
                 *      Loop until we find the end of the '.', resetting parent each time.
                 */
                while (p < end) {
                        if (*p == '.') {
                                pair_ctx->parent = da;
                                da = da->parent;
                                p++;
                                continue;
                        }
 
                        /*
                         *      Comma is the end of a reference.
                         */
                        if (*p == ',') {
                                return p - in;
                        }
 
                        /*
                         *      .FOO, must be a reference.
                         */
                        break;
                }
 
                if (p == end) return p - in;
 
        } else {
                /*
                 *      No leading '.', the reference MUST be from the root.
                 */
                pair_ctx->parent = fr_dict_root(pair_ctx->parent->dict);
 
                /*
                 *      We allow a leaf OR a reference here.
                 */
                slen = fr_dict_attr_by_name_substr(NULL, &da, pair_ctx->parent, &FR_SBUFF_IN(p, end), NULL);
                if (slen <= 0) return slen - (p - in);
 
                /*
                 *      Structural types do not have values.  So a
                 *      bare "Foo-Group" string MUST be changing the
                 *      reference to that da.
                 */
                switch (da->type) {
                case FR_TYPE_GROUP:
                case FR_TYPE_TLV:
                case FR_TYPE_STRUCT:
                        pair_ctx->parent = da;
                        p += slen;
 
                        if ((p >= end) || (*p == ',')) return p - in;
 
                        if (*p != '.') {
                                fr_strerror_const("Unexpected text");
                                return - (p - in);
                        }
 
                        p++;
                        break;
 
                default:
                        /*
                         *      Non-structural types MUST have values.
                         *      So change the parent to its parent,
                         *      and parse the leaf pair.
                         */
                        pair_ctx->parent = da->parent;
                        return fr_pair_afrom_str(pair_ctx, in, p, end - in);
                }
        }
 
        /*
         *      Set the new context based on the attribute
         */
        slen = fr_pair_ctx_set(pair_ctx, p, end - p);
        if (slen <= 0) return slen - (p - in);
 
        p += slen;
        return p - in;
}
 
/** Reset a pair_ctx to the dictionary root.
 *
 * @param pair_ctx      the parsing context
 * @param dict          the dictionary to reset to the root
 *
 *  This function is used in order to reset contexts when parsing
 *  strings that change attribute lists. i.e. &request.foo, &reply.bar
 *
 *  This function is simple for now, but will get complex once we
 *  start using vp->vp_children
 */
void fr_pair_ctx_reset(fr_pair_ctx_t *pair_ctx, fr_dict_t const *dict)
{
        pair_ctx->parent = fr_dict_root(dict);
}