lighttpd 1.4.x https://www.lighttpd.net/
You can not select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

1196 lines
31 KiB

#include "first.h"
#include "request.h"
#include "keyvalue.h"
#include "log.h"
#include <sys/stat.h>
#include <limits.h>
#include <stdlib.h>
#include <string.h>
#include <stdio.h>
#include <ctype.h>
static int request_check_hostname(server *srv, connection *con, buffer *host) {
enum { DOMAINLABEL, TOPLABEL } stage = TOPLABEL;
size_t i;
int label_len = 0;
size_t host_len;
char *colon;
int is_ip = -1; /* -1 don't know yet, 0 no, 1 yes */
int level = 0;
UNUSED(srv);
UNUSED(con);
/*
* hostport = host [ ":" port ]
* host = hostname | IPv4address | IPv6address
* hostname = *( domainlabel "." ) toplabel [ "." ]
* domainlabel = alphanum | alphanum *( alphanum | "-" ) alphanum
* toplabel = alpha | alpha *( alphanum | "-" ) alphanum
* IPv4address = 1*digit "." 1*digit "." 1*digit "." 1*digit
* IPv6address = "[" ... "]"
* port = *digit
*/
/* no Host: */
if (buffer_is_empty(host)) return 0;
host_len = buffer_string_length(host);
/* IPv6 adress */
if (host->ptr[0] == '[') {
char *c = host->ptr + 1;
int colon_cnt = 0;
/* check the address inside [...] */
for (; *c && *c != ']'; c++) {
if (*c == ':') {
if (++colon_cnt > 7) {
return -1;
}
} else if (!light_isxdigit(*c) && '.' != *c) {
return -1;
}
}
/* missing ] */
if (!*c) {
return -1;
}
/* check port */
if (*(c+1) == ':') {
for (c += 2; *c; c++) {
if (!light_isdigit(*c)) {
return -1;
}
}
}
else if ('\0' != *(c+1)) {
/* only a port is allowed to follow [...] */
return -1;
}
return 0;
}
if (NULL != (colon = memchr(host->ptr, ':', host_len))) {
char *c = colon + 1;
/* check portnumber */
for (; *c; c++) {
if (!light_isdigit(*c)) return -1;
}
/* remove the port from the host-len */
host_len = colon - host->ptr;
}
/* Host is empty */
if (host_len == 0) return -1;
/* if the hostname ends in a "." strip it */
if (host->ptr[host_len-1] == '.') {
/* shift port info one left */
if (NULL != colon) memmove(colon-1, colon, buffer_string_length(host) - host_len);
buffer_string_set_length(host, buffer_string_length(host) - 1);
host_len -= 1;
}
if (host_len == 0) return -1;
/* scan from the right and skip the \0 */
for (i = host_len; i-- > 0; ) {
const char c = host->ptr[i];
switch (stage) {
case TOPLABEL:
if (c == '.') {
/* only switch stage, if this is not the last character */
if (i != host_len - 1) {
if (label_len == 0) {
return -1;
}
/* check the first character at right of the dot */
if (is_ip == 0) {
if (!light_isalnum(host->ptr[i+1])) {
return -1;
}
} else if (!light_isdigit(host->ptr[i+1])) {
is_ip = 0;
} else if ('-' == host->ptr[i+1]) {
return -1;
} else {
/* just digits */
is_ip = 1;
}
stage = DOMAINLABEL;
label_len = 0;
level++;
} else if (i == 0) {
/* just a dot and nothing else is evil */
return -1;
}
} else if (i == 0) {
/* the first character of the hostname */
if (!light_isalnum(c)) {
return -1;
}
label_len++;
} else {
if (c != '-' && !light_isalnum(c)) {
return -1;
}
if (is_ip == -1) {
if (!light_isdigit(c)) is_ip = 0;
}
label_len++;
}
break;
case DOMAINLABEL:
if (is_ip == 1) {
if (c == '.') {
if (label_len == 0) {
return -1;
}
label_len = 0;
level++;
} else if (!light_isdigit(c)) {
return -1;
} else {
label_len++;
}
} else {
if (c == '.') {
if (label_len == 0) {
return -1;
}
/* c is either - or alphanum here */
if ('-' == host->ptr[i+1]) {
return -1;
}
label_len = 0;
level++;
} else if (i == 0) {
if (!light_isalnum(c)) {
return -1;
}
label_len++;
} else {
if (c != '-' && !light_isalnum(c)) {
return -1;
}
label_len++;
}
}
break;
}
}
/* a IP has to consist of 4 parts */
if (is_ip == 1 && level != 3) {
return -1;
}
if (label_len == 0) {
return -1;
}
return 0;
}
#if 0
#define DUMP_HEADER
#endif
static int http_request_split_value(array *vals, buffer *b) {
size_t i, len;
int state = 0;
const char *current;
const char *token_start = NULL, *token_end = NULL;
/*
* parse
*
* val1, val2, val3, val4
*
* into a array (more or less a explode() incl. striping of whitespaces
*/
if (buffer_string_is_empty(b)) return 0;
current = b->ptr;
len = buffer_string_length(b);
for (i = 0; i <= len; ++i, ++current) {
data_string *ds;
switch (state) {
case 0: /* find start of a token */
switch (*current) {
case ' ':
case '\t': /* skip white space */
case ',': /* skip empty token */
break;
case '\0': /* end of string */
return 0;
default:
/* found real data, switch to state 1 to find the end of the token */
token_start = token_end = current;
state = 1;
break;
}
break;
case 1: /* find end of token and last non white space character */
switch (*current) {
case ' ':
case '\t':
/* space - don't update token_end */
break;
case ',':
case '\0': /* end of string also marks the end of a token */
if (NULL == (ds = (data_string *)array_get_unused_element(vals, TYPE_STRING))) {
ds = data_string_init();
}
buffer_copy_string_len(ds->value, token_start, token_end-token_start+1);
array_insert_unique(vals, (data_unset *)ds);
state = 0;
break;
default:
/* no white space, update token_end to include current character */
token_end = current;
break;
}
break;
}
}
return 0;
}
static int request_uri_is_valid_char(unsigned char c) {
if (c <= 32) return 0;
if (c == 127) return 0;
if (c == 255) return 0;
return 1;
}
int http_request_parse(server *srv, connection *con) {
char *uri = NULL, *proto = NULL, *method = NULL, con_length_set;
int is_key = 1, key_len = 0, is_ws_after_key = 0, in_folding;
char *value = NULL, *key = NULL;
char *reqline_host = NULL;
int reqline_hostlen = 0;
enum { HTTP_CONNECTION_UNSET, HTTP_CONNECTION_KEEPALIVE, HTTP_CONNECTION_CLOSE } keep_alive_set = HTTP_CONNECTION_UNSET;
int line = 0;
int request_line_stage = 0;
size_t i, first, ilen;
int done = 0;
/*
* Request: "^(GET|POST|HEAD) ([^ ]+(\\?[^ ]+|)) (HTTP/1\\.[01])$"
* Option : "^([-a-zA-Z]+): (.+)$"
* End : "^$"
*/
if (con->conf.log_request_header) {
log_error_write(srv, __FILE__, __LINE__, "sdsdSb",
"fd:", con->fd,
"request-len:", buffer_string_length(con->request.request),
"\n", con->request.request);
}
if (con->request_count > 1 &&
con->request.request->ptr[0] == '\r' &&
con->request.request->ptr[1] == '\n') {
/* we are in keep-alive and might get \r\n after a previous POST request.*/
buffer_copy_string_len(con->parse_request, con->request.request->ptr + 2, buffer_string_length(con->request.request) - 2);
} else {
/* fill the local request buffer */
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
8 years ago
buffer_copy_buffer(con->parse_request, con->request.request);
}
keep_alive_set = 0;
con_length_set = 0;
/* parse the first line of the request
*
* should be:
*
* <method> <uri> <protocol>\r\n
* */
ilen = buffer_string_length(con->parse_request);
for (i = 0, first = 0; i < ilen && line == 0; i++) {
switch(con->parse_request->ptr[i]) {
case '\r':
if (con->parse_request->ptr[i+1] == '\n') {
http_method_t r;
char *nuri = NULL;
size_t j, jlen;
/* \r\n -> \0\0 */
con->parse_request->ptr[i] = '\0';
con->parse_request->ptr[i+1] = '\0';
buffer_copy_string_len(con->request.request_line, con->parse_request->ptr, i);
if (request_line_stage != 2) {
con->http_status = 400;
con->response.keep_alive = 0;
con->keep_alive = 0;
if (srv->srvconf.log_request_header_on_error) {
log_error_write(srv, __FILE__, __LINE__, "s", "incomplete request line -> 400");
log_error_write(srv, __FILE__, __LINE__, "Sb",
"request-header:\n",
con->request.request);
}
return 0;
}
proto = con->parse_request->ptr + first;
*(uri - 1) = '\0';
*(proto - 1) = '\0';
/* we got the first one :) */
if (HTTP_METHOD_UNSET == (r = get_http_method_key(method))) {
con->http_status = 501;
con->response.keep_alive = 0;
con->keep_alive = 0;
if (srv->srvconf.log_request_header_on_error) {
log_error_write(srv, __FILE__, __LINE__, "s", "unknown http-method -> 501");
log_error_write(srv, __FILE__, __LINE__, "Sb",
"request-header:\n",
con->request.request);
}
return 0;
}
con->request.http_method = r;
/*
* RFC2616 says:
*
* HTTP-Version = "HTTP" "/" 1*DIGIT "." 1*DIGIT
*
* */
if (0 == strncmp(proto, "HTTP/", sizeof("HTTP/") - 1)) {
char * major = proto + sizeof("HTTP/") - 1;
char * minor = strchr(major, '.');
char *err = NULL;
int major_num = 0, minor_num = 0;
int invalid_version = 0;
if (NULL == minor || /* no dot */
minor == major || /* no major */
*(minor + 1) == '\0' /* no minor */) {
invalid_version = 1;
} else {
*minor = '\0';
major_num = strtol(major, &err, 10);
if (*err != '\0') invalid_version = 1;
*minor++ = '.';
minor_num = strtol(minor, &err, 10);
if (*err != '\0') invalid_version = 1;
}
if (invalid_version) {
con->http_status = 400;
con->keep_alive = 0;
if (srv->srvconf.log_request_header_on_error) {
log_error_write(srv, __FILE__, __LINE__, "s", "unknown protocol -> 400");
log_error_write(srv, __FILE__, __LINE__, "Sb",
"request-header:\n",
con->request.request);
}
return 0;
}
if (major_num == 1 && minor_num == 1) {
con->request.http_version = con->conf.allow_http11 ? HTTP_VERSION_1_1 : HTTP_VERSION_1_0;
} else if (major_num == 1 && minor_num == 0) {
con->request.http_version = HTTP_VERSION_1_0;
} else {
con->http_status = 505;
if (srv->srvconf.log_request_header_on_error) {
log_error_write(srv, __FILE__, __LINE__, "s", "unknown HTTP version -> 505");
log_error_write(srv, __FILE__, __LINE__, "Sb",
"request-header:\n",
con->request.request);
}
return 0;
}
} else {
con->http_status = 400;
con->keep_alive = 0;
if (srv->srvconf.log_request_header_on_error) {
log_error_write(srv, __FILE__, __LINE__, "s", "unknown protocol -> 400");
log_error_write(srv, __FILE__, __LINE__, "Sb",
"request-header:\n",
con->request.request);
}
return 0;
}
if (0 == strncmp(uri, "http://", 7) &&
NULL != (nuri = strchr(uri + 7, '/'))) {
reqline_host = uri + 7;
reqline_hostlen = nuri - reqline_host;
buffer_copy_string_len(con->request.uri, nuri, proto - nuri - 1);
} else if (0 == strncmp(uri, "https://", 8) &&
NULL != (nuri = strchr(uri + 8, '/'))) {
reqline_host = uri + 8;
reqline_hostlen = nuri - reqline_host;
buffer_copy_string_len(con->request.uri, nuri, proto - nuri - 1);
} else {
/* everything looks good so far */
buffer_copy_string_len(con->request.uri, uri, proto - uri - 1);
}
/* check uri for invalid characters */
jlen = buffer_string_length(con->request.uri);
for (j = 0; j < jlen; j++) {
if (!request_uri_is_valid_char(con->request.uri->ptr[j])) {
unsigned char buf[2];
con->http_status = 400;
con->keep_alive = 0;
if (srv->srvconf.log_request_header_on_error) {
buf[0] = con->request.uri->ptr[j];
buf[1] = '\0';
if (con->request.uri->ptr[j] > 32 &&
con->request.uri->ptr[j] != 127) {
/* the character is printable -> print it */
log_error_write(srv, __FILE__, __LINE__, "ss",
"invalid character in URI -> 400",
buf);
} else {
/* a control-character, print ascii-code */
log_error_write(srv, __FILE__, __LINE__, "sd",
"invalid character in URI -> 400",
con->request.uri->ptr[j]);
}
log_error_write(srv, __FILE__, __LINE__, "Sb",
"request-header:\n",
con->request.request);
}
return 0;
}
}
fix buffer, chunk and http_chunk API * remove unused structs and functions (buffer_array, read_buffer) * change return type from int to void for many functions, as the return value (indicating error/success) was never checked, and the function would only fail on programming errors and not on invalid input; changed functions to use force_assert instead of returning an error. * all "len" parameters now are the real size of the memory to be read. the length of strings is given always without the terminating 0. * the "buffer" struct still counts the terminating 0 in ->used, provide buffer_string_length() to get the length of a string in a buffer. unset config "strings" have used == 0, which is used in some places to distinguish unset values from "" (empty string) values. * most buffer usages should now use it as string container. * optimise some buffer copying by "moving" data to other buffers * use (u)intmax_t for generic int-to-string functions * remove unused enum values: UNUSED_CHUNK, ENCODING_UNSET * converted BUFFER_APPEND_SLASH to inline function (no macro feature needed) * refactor: create chunkqueue_steal: moving (partial) chunks into another queue * http_chunk: added separate function to terminate chunked body instead of magic handling in http_chunk_append_mem(). http_chunk_append_* now handle empty chunks, and never terminate the chunked body. From: Stefan Bühler <stbuehler@web.de> git-svn-id: svn://svn.lighttpd.net/lighttpd/branches/lighttpd-1.4.x@2975 152afb58-edef-0310-8abb-c4023f1b3aa9
8 years ago
buffer_copy_buffer(con->request.orig_uri, con->request.uri);
con->http_status = 0;
i++;
line++;
first = i+1;
}
break;
case ' ':
switch(request_line_stage) {
case 0:
/* GET|POST|... */
method = con->parse_request->ptr + first;
first = i + 1;
break;
case 1:
/* /foobar/... */
uri = con->parse_request->ptr + first;
first = i + 1;
break;
default:
/* ERROR, one space to much */
con->http_status = 400;
con->response.keep_alive = 0;
con->keep_alive = 0;
if (srv->srvconf.log_request_header_on_error) {
log_error_write(srv, __FILE__, __LINE__, "s", "overlong request line -> 400");
log_error_write(srv, __FILE__, __LINE__, "Sb",
"request-header:\n",
con->request.request);
}
return 0;
}
request_line_stage++;
break;
}
}
in_folding = 0;
if (buffer_string_is_empty(con->request.uri)) {
con->http_status = 400;
con->response.keep_alive = 0;
con->keep_alive = 0;
if (srv->srvconf.log_request_header_on_error) {
log_error_write(srv, __FILE__, __LINE__, "s", "no uri specified -> 400");
log_error_write(srv, __FILE__, __LINE__, "Sb",
"request-header:\n",
con->request.request);
}
return 0;
}
if (reqline_host) {
/* Insert as host header */
data_string *ds;
if (NULL == (ds = (data_string *)array_get_unused_element(con->request.headers, TYPE_STRING))) {
ds = data_string_init();
}
buffer_copy_string_len(ds->key, CONST_STR_LEN("Host"));
buffer_copy_string_len(ds->value, reqline_host, reqline_hostlen);
array_insert_unique(con->request.headers, (data_unset *)ds);
con->request.http_host = ds->value;
}
for (; i <= ilen && !done; i++) {
char *cur = con->parse_request->ptr + i;
if (is_key) {
size_t j;
int got_colon = 0;
/**
* 1*<any CHAR except CTLs or separators>
* CTLs == 0-31 + 127, CHAR = 7-bit ascii (0..127)
*
*/
switch(*cur) {
case ':':
is_key = 0;
value = cur + 1;
if (is_ws_after_key == 0) {
key_len = i - first;
}
is_ws_after_key = 0;
break;
case '(':
case ')':
case '<':
case '>':
case '@':
case ',':
case ';':
case '\\':
case '\"':
case '/':
case '[':
case ']':
case '?':
case '=':
case '{':
case '}':
con->http_status = 400;
con->keep_alive = 0;
con->response.keep_alive = 0;
if (srv->srvconf.log_request_header_on_error) {
log_error_write(srv, __FILE__, __LINE__, "sbsds",
"invalid character in key", con->request.request, cur, *cur, "-> 400");
log_error_write(srv, __FILE__, __LINE__, "Sb",
"request-header:\n",
con->request.request);
}
return 0;
case ' ':
case '\t':
if (i == first) {
is_key = 0;
in_folding = 1;
value = cur;
break;
}
key_len = i - first;
/* skip every thing up to the : */
for (j = 1; !got_colon; j++) {
switch(con->parse_request->ptr[j + i]) {
case ' ':
case '\t':
/* skip WS */
continue;
case ':':
/* ok, done; handle the colon the usual way */
i += j - 1;
got_colon = 1;
is_ws_after_key = 1; /* we already know the key length */
break;
default:
/* error */
if (srv->srvconf.log_request_header_on_error) {
log_error_write(srv, __FILE__, __LINE__, "s", "WS character in key -> 400");
log_error_write(srv, __FILE__, __LINE__, "Sb",
"request-header:\n",
con->request.request);
}
con->http_status = 400;
con->response.keep_alive = 0;
con->keep_alive = 0;
return 0;
}
}
break;
case '\r':
if (con->parse_request->ptr[i+1] == '\n' && i == first) {
/* End of Header */
con->parse_request->ptr[i] = '\0';
con->parse_request->ptr[i+1] = '\0';
i++;
done = 1;
} else {
if (srv->srvconf.log_request_header_on_error) {
log_error_write(srv, __FILE__, __LINE__, "s", "CR without LF -> 400");
log_error_write(srv, __FILE__, __LINE__, "Sb",
"request-header:\n",
con->request.request);
}
con->http_status = 400;
con->keep_alive = 0;
con->response.keep_alive = 0;
return 0;
}
break;
default:
if (*cur < 32 || ((unsigned char)*cur) >= 127) {
con->http_status = 400;
con->keep_alive = 0;