proxy_util.c revision 92d311b27a6182c2eed67317990c8c168584ee75
/* ====================================================================
* The Apache Software License, Version 1.1
*
* Copyright (c) 2000-2001 The Apache Software Foundation. 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
* distribution.
*
* 3. The end-user documentation included with the redistribution,
* if any, must include the following acknowledgment:
* "This product includes software developed by the
* Apache Software Foundation (http://www.apache.org/)."
* Alternately, this acknowledgment may appear in the software itself,
* if and wherever such third-party acknowledgments normally appear.
*
* 4. The names "Apache" and "Apache Software Foundation" must
* not be used to endorse or promote products derived from this
* software without prior written permission. For written
* permission, please contact apache@apache.org.
*
* 5. Products derived from this software may not be called "Apache",
* nor may "Apache" appear in their name, without prior written
* permission of the Apache Software Foundation.
*
* THIS SOFTWARE IS PROVIDED ``AS IS'' AND ANY EXPRESSED 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 APACHE SOFTWARE FOUNDATION OR
* ITS CONTRIBUTORS 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.
* ====================================================================
*
* This software consists of voluntary contributions made by many
* individuals on behalf of the Apache Software Foundation. For more
* information on the Apache Software Foundation, please see
*
* Portions of this software are based upon public domain software
* originally written at the National Center for Supercomputing Applications,
* University of Illinois, Urbana-Champaign.
*/
/* Utility routines for Apache proxy */
#include "mod_proxy.h"
/* already called in the knowledge that the characters are hex digits */
PROXY_DECLARE(int) ap_proxy_hex2c(const char *x)
{
int i, ch;
#if !APR_CHARSET_EBCDIC
ch = x[0];
if (apr_isdigit(ch))
i = ch - '0';
else if (apr_isupper(ch))
else
i <<= 4;
ch = x[1];
if (apr_isdigit(ch))
i += ch - '0';
else if (apr_isupper(ch))
else
return i;
#else /*APR_CHARSET_EBCDIC*/
#endif /*APR_CHARSET_EBCDIC*/
}
{
#if !APR_CHARSET_EBCDIC
int i;
x[0] = '%';
if (i >= 10)
x[1] = ('A' - 10) + i;
else
x[1] = '0' + i;
i = ch & 0x0F;
if (i >= 10)
x[2] = ('A' - 10) + i;
else
x[2] = '0' + i;
#else /*APR_CHARSET_EBCDIC*/
static const char ntoa[] = { "0123456789ABCDEF" };
ch &= 0xFF;
x[0] = '%';
x[3] = '\0';
#endif /*APR_CHARSET_EBCDIC*/
}
/*
* canonicalise a URL-encoded string
*/
/*
* Convert a URL-encoded string to canonical form.
* It decodes characters which need not be encoded,
* and encodes those which must be encoded, and does not touch
* those which must not be touched.
*/
int isenc)
{
int i, j, ch;
char *y;
char *allowed; /* characters which should not be encoded */
/* N.B. in addition to :@&=, this allows ';' in an http path
* and '?' in an ftp path -- this may be revised
*
* Also, it makes a '+' character in a search string reserved, as
* it may be form-encoded. (Although RFC 1738 doesn't allow this -
* it only permits ; / ? : @ = & as reserved chars.)
*/
if (t == enc_path)
allowed = "$-_.+!*'(),;:@&=";
else if (t == enc_search)
allowed = "$-_.!*'(),;:@&=";
else if (t == enc_user)
allowed = "$-_.+!*'(),;@&=";
else if (t == enc_fpath)
allowed = "$-_.+!*'(),?:@&=";
else /* if (t == enc_parm) */
allowed = "$-_.+!*'(),?/:@&=";
if (t == enc_path)
reserved = "/";
else if (t == enc_search)
reserved = "+";
else
reserved = "";
for (i = 0, j = 0; i < len; i++, j++) {
/* always handle '/' first */
ch = x[i];
y[j] = ch;
continue;
}
/* decode it if not already done */
return NULL;
i += 2;
ap_proxy_c2hex(ch, &y[j]);
j += 2;
continue;
}
}
/* recode it, if necessary */
ap_proxy_c2hex(ch, &y[j]);
j += 2;
}
else
y[j] = ch;
}
y[j] = '\0';
return y;
}
/*
* Parses network-location.
* urlp on input the URL; on output the path, after the leading /
* password holder for password
* host holder for host
* port port number; only set if one is supplied.
*
* Returns an error string.
*/
PROXY_DECLARE(char *)
{
apr_port_t i;
return "Malformed URL";
url = "";
else
*strp = '\0';
/* find password */
*strp = '\0';
return "Bad %-escape in URL (password)";
}
return "Bad %-escape in URL (username)";
}
}
}
*(strp++) = '\0';
for (i = 0; strp[i] != '\0'; i++)
if (!apr_isdigit(strp[i]))
break;
/* if (i == 0) the no port was given; keep default */
if (strp[i] != '\0') {
return "Bad port number in URL";
} else if (i > 0) {
if (*port > 65535)
return "Port number in URL > 65535";
}
}
if (*host == '\0')
return "Missing host in URL";
/* check hostname syntax */
for (i = 0; host[i] != '\0'; i++)
break;
/* must be an IP address */
{
return "Bad IP address in URL";
}
/* if (strchr(host,'.') == NULL && domain != NULL)
host = pstrcat(p, host, domain, NULL);
*/
return NULL;
}
static const char * const lwday[7] =
{"Sunday", "Monday", "Tuesday", "Wednesday", "Thursday", "Friday", "Saturday"};
/*
* If the date is a valid RFC 850 date or asctime() date, then it
* is converted to the RFC 1123 format, otherwise it is not modified.
* This routine is not very fast at doing conversions, as it uses
* sscanf and sprintf. However, if the date is already correctly
* formatted, then it exits very quickly.
*/
PROXY_DECLARE(const char *)
{
char *x = apr_pstrdup(p, x1);
q = strchr(x, ',');
/* check for RFC 850 date */
*q = '\0';
break;
*q = ',';
if (wk == 7)
return x; /* not a valid date */
if (q[4] != '-' || q[8] != '-' || q[11] != ' ' || q[14] != ':' ||
return x;
return x;
if (year < 70)
year += 2000;
else
year += 1900;
}
else {
/* check for acstime() date */
if (x[3] != ' ' || x[7] != ' ' || x[10] != ' ' || x[13] != ':' ||
x[16] != ':' || x[19] != ' ' || x[24] != '\0')
return x;
return x;
break;
if (wk == 7)
return x;
}
/* check date */
break;
if (mon == 12)
return x;
q = apr_palloc(p, 30);
return q;
}
{
rp->connection = c;
proxy_run_create_req(r, rp);
return rp;
}
/*
* Reads headers from a buffer and returns an array of headers.
* Returns NULL on file error
* This routine tries to deal with too long lines and continuation lines.
* @@@: XXX: FIXME: currently the headers are passed thru un-merged.
* Is that okay, or should they be collapsed where possible?
*/
PROXY_DECLARE(apr_table_t *)ap_proxy_read_headers(request_rec *r, request_rec *rr, char *buffer, int size, conn_rec *c)
{
int len;
char field[MAX_STRING_LEN];
/*
* Read header lines until we get the empty separator line, a read error,
* the connection closes (EOF), or we timeout.
*/
/* Buggy MS IIS servers sometimes return invalid headers
* (an extra "HTTP/1.0 200, OK" line sprinkled in between
* the usual MIME headers). Try to deal with it in a sensible
* way, but log the fact.
* XXX: The mask check is buggy if we ever see an HTTP/1.10 */
/* Nope, it wasn't even an extra HTTP header. Give up. */
return NULL;
}
"proxy: Ignoring duplicate HTTP header "
continue;
}
*value = '\0';
++value;
/* XXX: RFC2068 defines only SP and HT as whitespace, this test is
* wrong... and so are many others probably.
*/
while (apr_isspace(*value))
++value; /* Skip to start of value */
/* should strip trailing whitespace as well */
*end = '\0';
/* make sure we merge so as not to destroy duplicated headers */
/* the header was too long; at the least we should skip extra data */
>= MAX_STRING_LEN - 1) {
/* soak up the extra data */
}
if (len == 0) /* time to exit the larger loop as well */
break;
}
}
return headers_out;
}
/*
* list is a comma-separated list of case-insensitive tokens, with
* optional whitespace around the tokens.
* The return returns 1 if the token val is found in the list, or 0
* otherwise.
*/
{
int len, i;
const char *p;
if (p != NULL) {
i = p - list;
do
p++;
while (apr_isspace(*p));
}
else
i--;
return 1;
list = p;
}
return 0;
}
/*
* list is a comma-separated list of case-insensitive tokens, with
* optional whitespace around the tokens.
* The return returns 1 if the token val is found in the list, or 0
* otherwise.
*/
{
int len, i;
const char *p;
if (p != NULL) {
i = p - list;
do
p++;
while (apr_isspace(*p));
}
else
i--;
/* do nothing */
}
else {
if (new)
else
}
list = p;
}
return new;
}
/*
* Converts 8 hex digits to a time integer
*/
PROXY_DECLARE(int) ap_proxy_hex2sec(const char *x)
{
int i, ch;
unsigned int j;
for (i = 0, j = 0; i < 8; i++) {
ch = x[i];
j <<= 4;
if (apr_isdigit(ch))
j |= ch - '0';
else if (apr_isupper(ch))
else
}
if (j == 0xffffffff)
return -1; /* so that it works with 8-byte ints */
else
return j;
}
/*
* Converts a time integer to 8 hex digits
*/
PROXY_DECLARE(void) ap_proxy_sec2hex(int t, char *y)
{
int i, ch;
unsigned int j = t;
for (i = 7; i >= 0; i--) {
ch = j & 0xF;
j >>= 4;
if (ch >= 10)
else
y[i] = ch + '0';
}
y[8] = '\0';
}
{
apr_pstrcat(r->pool,
"The proxy server could not handle the request "
" ",
"Reason: <STRONG>",
"</STRONG>", NULL));
/* Allow "error-notes" string to be printed by ap_send_error_response() */
return statuscode;
}
static const char *
{
return r->hostname;
/* Set url to the first char after "scheme://" */
return NULL;
url = apr_pstrdup(r->pool, &url[1]); /* make it point to "//", which is what proxy_canon_netloc expects */
"%s", err);
return host; /* ought to return the port, too */
}
/* Return TRUE if addr represents an IP address (or an IP network address) */
{
long ip_addr[4];
int i, quads;
long bits;
/* if the address is given with an explicit netmask, use that */
/* Due to a deficiency in apr_inet_addr(), it is impossible to parse */
/* "partial" addresses (with less than 4 quads) correctly, i.e. */
/* 192.168.123 is parsed as 192.168.0.123, which is not what I want. */
/* I therefore have to parse the IP address manually: */
/* addr and mask were set by proxy_readmask() */
/*return 1; */
/* Parse IP addr manually, optionally allowing */
/* abbreviated net addresses like 192.168. */
/* Iterate over up to 4 (dotted) quads. */
char *tmp;
break;
if (!apr_isdigit(*addr))
return 0; /* no digit at start of quad */
return 0;
/* invalid octet */
return 0;
}
++addr; /* after the 4th quad, a dot would be illegal */
}
char *tmp;
++addr;
return 0;
return 0;
}
else {
/* Determine (i.e., "guess") netmask by counting the */
/* number of trailing .0's; reduce #quads appropriately */
/* (so that 192.168.0.0 is equivalent to 192.168.) */
--quads;
/* "IP Address should be given in dotted-quad form, optionally followed by a netmask (e.g., 192.168.111.0/24)"; */
if (quads < 1)
return 0;
/* every zero-byte counts as 8 zero-bits */
"Warning: NetMask not supplied with IP-Addr; guessing: %s/%ld\n",
}
"Warning: NetMask and IP-Addr disagree in %s/%ld\n",
" Set to %s/%ld\n",
}
if (*addr == '\0') {
return 1;
}
else
}
/* Return TRUE if addr represents an IP address (or an IP network address) */
{
int i, ip_addr[4];
const char *host = proxy_get_host_of_request(r);
return 0;
#if DEBUGGING
#endif
return 1;
}
#if DEBUGGING
else {
}
#endif
}
else {
struct apr_sockaddr_t *reqaddr;
!= APR_SUCCESS) {
#if DEBUGGING
"2)IP-NoMatch: hostname=%s msg=Host not found",
host);
#endif
return 0;
}
/* Try to deal with multiple IP addr's for a host */
/* FIXME: This needs to be able to deal with IPv6 */
while (reqaddr) {
#if DEBUGGING
"3)IP-Match: %s[%s] <-> ", host,
#endif
return 1;
}
#if DEBUGGING
else {
"3)IP-NoMatch: %s[%s] <-> ", host,
}
#endif
}
}
return 0;
}
/* Return TRUE if addr represents a domain name */
{
int i;
/* Domain name must start with a '.' */
if (addr[0] != '.')
return 0;
/* rfc1035 says DNS names must consist of "[-a-zA-Z0-9]" and '.' */
continue;
#if 0
if (addr[i] == ':') {
"@@@@ handle optional port in proxy_is_domainname()");
/* @@@@ handle optional port */
}
#endif
if (addr[i] != '\0')
return 0;
/* Strip trailing dots */
addr[i] = '\0';
return 1;
}
/* Return TRUE if host "host" is in domain "domain" */
{
const char *host = proxy_get_host_of_request(r);
return 0;
/* @@@ do this within the setup? */
/* Ignore trailing dots in domain comparison: */
--d_len;
--h_len;
}
/* Return TRUE if host represents a host name */
{
struct apr_sockaddr_t *addr;
int i;
/* Host names must not start with a '.' */
if (host[0] == '.')
return 0;
/* rfc1035 says DNS names must consist of "[-a-zA-Z0-9]" and '.' */
return 0;
/* Strip trailing dots */
host[i] = '\0';
return 1;
}
/* Return TRUE if host "host" is equal to host2 "host2" */
{
const char *host2 = proxy_get_host_of_request(r);
int h2_len;
int h1_len;
return 0; /* oops! */
#if 0
/* Try to deal with multiple IP addr's for a host */
while (addr) {
if (addr->ipaddr_ptr == ? ? ? ? ? ? ? ? ? ? ? ? ?)
return 1;
}
#endif
/* Ignore trailing dots in host2 comparison: */
--h2_len;
--h1_len;
}
/* Return TRUE if addr is to be matched as a word */
{
return 1;
}
/* Return TRUE if string "str2" occurs literally in "str1" */
{
const char *host = proxy_get_host_of_request(r);
}
/* checks whether a host in uri_addr matches proxyblock */
{
int j;
/* XXX FIXME: conf->noproxies->elts is part of an opaque structure */
return HTTP_FORBIDDEN;
}
while (conf_addr) {
while (uri_addr) {
char *conf_ip;
char *uri_ip;
return HTTP_FORBIDDEN;
}
}
}
}
return OK;
}
/* set up the minimal filter set */
{
return OK;
}
/* converts a series of buckets into a string
* XXX: BillS says this function performs essentially the same function as
* ap_rgetline() in protocol.c. Deprecate this function and use ap_rgetline()
* instead? I think ap_proxy_string_read() will not work properly on non ASCII
* (EBCDIC) machines either.
*/
{
apr_bucket *e;
char *response;
int found = 0;
/* start with an empty string */
buff[0] = 0;
*eos = 0;
/* loop through each brigade */
while (!found) {
/* get brigade from network one line at a time */
0))) {
return rv;
}
/* loop through each bucket */
while (!found) {
/* The connection aborted or timed out */
return APR_ECONNABORTED;
}
e = APR_BRIGADE_FIRST(bb);
if (APR_BUCKET_IS_EOS(e)) {
*eos = 1;
}
else {
return rv;
}
/* is string LF terminated?
* XXX: This check can be made more efficient by simply checking
* if the last character in the 'response' buffer is an ASCII_LF.
* See ap_rgetline() for an example.
*/
found = 1;
}
/* concat strings until buff is full - then throw the data away */
}
if (len > 0) {
}
}
}
}
return APR_SUCCESS;
}
/* remove other filters (like DECHUNK) from filter stack */
{
ap_filter_t *f = c->output_filters;
while (f) {
f = f->next;
continue;
}
else {
f = f->next;
}
}
}
/* unmerge an element in the table */
{
/* get the value to unmerge */
if (!initial) {
return;
}
/* remove the value from the headers */
apr_table_unset(t, key);
/* find each comma */
}
count++;
}
}
const char *proxy_function,
const char *backend_name,
server_rec *s,
apr_pool_t *p)
{
int connected = 0;
int loglevel;
while (backend_addr && !connected) {
SOCK_STREAM, p)) != APR_SUCCESS) {
"proxy: %s: error creating fam %d socket for target %s",
/* this could be an IPv6 address from the DNS but the
* local machine won't give us an IPv6 socket; hopefully the
* DNS returned an additional address to try
*/
continue;
}
if (conf->recv_buffer_size > 0 &&
conf->recv_buffer_size))) {
"setsockopt(SO_RCVBUF): Failed to set "
"ProxyReceiveBufferSize, using default");
}
#endif
/* Set a timeout on the socket */
}
else {
(int)(s->timeout * APR_USEC_PER_SEC));
}
"proxy: %s: fam %d socket created to connect to %s",
/* make the connection out of the socket */
/* if an error occurred, loop round and try again */
if (rv != APR_SUCCESS) {
"proxy: %s: attempt to connect to %pI (%s) failed",
continue;
}
connected = 1;
}
return connected ? 0 : 1;
}