haiwan
/
http-parser
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Compare commits

base: haiwan:master
Branches Tags
haiwan:master
haiwan:make-http-max-header-size-gyp-configurable
haiwan:v0.10
haiwan:v0.6
haiwan:event_stream
haiwan:version0.2
..
compare: haiwan:make-http-max-header-size-gyp-configurable
Branches Tags
haiwan:master
haiwan:make-http-max-header-size-gyp-configurable
haiwan:v0.10
haiwan:v0.6
haiwan:event_stream
haiwan:version0.2

2 Commits
master ... make-http-

Author SHA1 Message Date
Matteo Collina 1ee68c071c style nit
6 years ago
Matteo Collina a0d4cd119e Make HTTP_MAX_HEADER_SIZE configurable via gyp
6 years ago
5 changed files with 69 additions and 260 deletions
Show Stats

12
Makefile
Unescape View File

@ -23,8 +23,8 @@ HELPER ?=
BINEXT ?=
SOLIBNAME = libhttp_parser
SOMAJOR = 2
SOMINOR = 9
SOREV = 4
SOMINOR = 8
SOREV = 1
ifeq (darwin,$(PLATFORM))
SOEXT ?= dylib
SONAME ?= $(SOLIBNAME).$(SOMAJOR).$(SOMINOR).$(SOEXT)
@ -133,14 +133,14 @@ tags: http_parser.c http_parser.h test.c
install: library
$(INSTALL) -D http_parser.h $(DESTDIR)$(INCLUDEDIR)/http_parser.h
$(INSTALL) -D $(LIBNAME) $(DESTDIR)$(LIBDIR)/$(LIBNAME)
ln -sf $(LIBNAME) $(DESTDIR)$(LIBDIR)/$(SONAME)
ln -sf $(LIBNAME) $(DESTDIR)$(LIBDIR)/$(SOLIBNAME).$(SOEXT)
ln -s $(LIBNAME) $(DESTDIR)$(LIBDIR)/$(SONAME)
ln -s $(LIBNAME) $(DESTDIR)$(LIBDIR)/$(SOLIBNAME).$(SOEXT)
install-strip: library
$(INSTALL) -D http_parser.h $(DESTDIR)$(INCLUDEDIR)/http_parser.h
$(INSTALL) -D -s $(LIBNAME) $(DESTDIR)$(LIBDIR)/$(LIBNAME)
ln -sf $(LIBNAME) $(DESTDIR)$(LIBDIR)/$(SONAME)
ln -sf $(LIBNAME) $(DESTDIR)$(LIBDIR)/$(SOLIBNAME).$(SOEXT)
ln -s $(LIBNAME) $(DESTDIR)$(LIBDIR)/$(SONAME)
ln -s $(LIBNAME) $(DESTDIR)$(LIBDIR)/$(SOLIBNAME).$(SOEXT)
uninstall:
rm $(DESTDIR)$(INCLUDEDIR)/http_parser.h

160
http_parser.c
Unescape View File

@ -25,8 +25,6 @@
#include <string.h>
#include <limits.h>
static uint32_t max_header_size = HTTP_MAX_HEADER_SIZE;
#ifndef ULLONG_MAX
# define ULLONG_MAX ((uint64_t) -1) /* 2^64-1 */
#endif
@ -141,20 +139,20 @@ do { \
} while (0)
/* Don't allow the total size of the HTTP headers (including the status
* line) to exceed max_header_size. This check is here to protect
* line) to exceed HTTP_MAX_HEADER_SIZE. This check is here to protect
* embedders against denial-of-service attacks where the attacker feeds
* us a never-ending header that the embedder keeps buffering.
*
* This check is arguably the responsibility of embedders but we're doing
* it on the embedder's behalf because most won't bother and this way we
* make the web a little safer. max_header_size is still far bigger
* make the web a little safer. HTTP_MAX_HEADER_SIZE is still far bigger
* than any reasonable request or response so this should never affect
* day-to-day operation.
*/
#define COUNT_HEADER_SIZE(V) \
do { \
nread += (uint32_t)(V); \
if (UNLIKELY(nread > max_header_size)) { \
nread += (V); \
if (UNLIKELY(nread > (HTTP_MAX_HEADER_SIZE))) { \
SET_ERRNO(HPE_HEADER_OVERFLOW); \
goto error; \
} \
@ -381,10 +379,7 @@ enum header_states
, h_transfer_encoding
, h_upgrade
, h_matching_transfer_encoding_token_start
, h_matching_transfer_encoding_chunked
, h_matching_transfer_encoding_token
, h_matching_connection_token_start
, h_matching_connection_keep_alive
, h_matching_connection_close
@ -731,7 +726,6 @@ reexecute:
if (ch == CR || ch == LF)
break;
parser->flags = 0;
parser->extra_flags = 0;
parser->content_length = ULLONG_MAX;
if (ch == 'H') {
@ -769,7 +763,6 @@ reexecute:
if (ch == CR || ch == LF)
break;
parser->flags = 0;
parser->extra_flags = 0;
parser->content_length = ULLONG_MAX;
if (ch == 'H') {
@ -927,7 +920,6 @@ reexecute:
if (ch == CR || ch == LF)
break;
parser->flags = 0;
parser->extra_flags = 0;
parser->content_length = ULLONG_MAX;
if (UNLIKELY(!IS_ALPHA(ch))) {
@ -1263,9 +1255,9 @@ reexecute:
switch (parser->header_state) {
case h_general: {
size_t left = data + len - p;
const char* pe = p + MIN(left, max_header_size);
while (p+1 < pe && TOKEN(p[1])) {
size_t limit = data + len - p;
limit = MIN(limit, HTTP_MAX_HEADER_SIZE);
while (p+1 < data + limit && TOKEN(p[1])) {
p++;
}
break;
@ -1341,7 +1333,6 @@ reexecute:
parser->header_state = h_general;
} else if (parser->index == sizeof(TRANSFER_ENCODING)-2) {
parser->header_state = h_transfer_encoding;
parser->extra_flags |= F_TRANSFER_ENCODING >> 8;
}
break;
@ -1423,14 +1414,10 @@ reexecute:
if ('c' == c) {
parser->header_state = h_matching_transfer_encoding_chunked;
} else {
parser->header_state = h_matching_transfer_encoding_token;
parser->header_state = h_general;
}
break;
/* Multi-value `Transfer-Encoding` header */
case h_matching_transfer_encoding_token_start:
break;
case h_content_length:
if (UNLIKELY(!IS_NUM(ch))) {
SET_ERRNO(HPE_INVALID_CONTENT_LENGTH);
@ -1447,11 +1434,6 @@ reexecute:
parser->header_state = h_content_length_num;
break;
/* when obsolete line folding is encountered for content length
* continue to the s_header_value state */
case h_content_length_ws:
break;
case h_connection:
/* looking for 'Connection: keep-alive' */
if (c == 'k') {
@ -1508,21 +1490,24 @@ reexecute:
switch (h_state) {
case h_general:
{
size_t left = data + len - p;
const char* pe = p + MIN(left, max_header_size);
for (; p != pe; p++) {
ch = *p;
if (ch == CR || ch == LF) {
--p;
break;
}
if (!lenient && !IS_HEADER_CHAR(ch)) {
SET_ERRNO(HPE_INVALID_HEADER_TOKEN);
goto error;
}
const char* p_cr;
const char* p_lf;
size_t limit = data + len - p;
limit = MIN(limit, HTTP_MAX_HEADER_SIZE);
p_cr = (const char*) memchr(p, CR, limit);
p_lf = (const char*) memchr(p, LF, limit);
if (p_cr != NULL) {
if (p_lf != NULL && p_cr >= p_lf)
p = p_lf;
else
p = p_cr;
} else if (UNLIKELY(p_lf != NULL)) {
p = p_lf;
} else {
p = data + len;
}
if (p == data + len)
--p;
break;
}
@ -1574,41 +1559,16 @@ reexecute:
goto error;
/* Transfer-Encoding: chunked */
case h_matching_transfer_encoding_token_start:
/* looking for 'Transfer-Encoding: chunked' */
if ('c' == c) {
h_state = h_matching_transfer_encoding_chunked;
} else if (STRICT_TOKEN(c)) {
/* TODO(indutny): similar code below does this, but why?
* At the very least it seems to be inconsistent given that
* h_matching_transfer_encoding_token does not check for
* `STRICT_TOKEN`
*/
h_state = h_matching_transfer_encoding_token;
} else if (c == ' ' || c == '\t') {
/* Skip lws */
} else {
h_state = h_general;
}
break;
case h_matching_transfer_encoding_chunked:
parser->index++;
if (parser->index > sizeof(CHUNKED)-1
|| c != CHUNKED[parser->index]) {
h_state = h_matching_transfer_encoding_token;
h_state = h_general;
} else if (parser->index == sizeof(CHUNKED)-2) {
h_state = h_transfer_encoding_chunked;
}
break;
case h_matching_transfer_encoding_token:
if (ch == ',') {
h_state = h_matching_transfer_encoding_token_start;
parser->index = 0;
}
break;
case h_matching_connection_token_start:
/* looking for 'Connection: keep-alive' */
if (c == 'k') {
@ -1667,7 +1627,7 @@ reexecute:
break;
case h_transfer_encoding_chunked:
if (ch != ' ') h_state = h_matching_transfer_encoding_token;
if (ch != ' ') h_state = h_general;
break;
case h_connection_keep_alive:
@ -1717,10 +1677,6 @@ reexecute:
case s_header_value_lws:
{
if (ch == ' ' || ch == '\t') {
if (parser->header_state == h_content_length_num) {
/* treat obsolete line folding as space */
parser->header_state = h_content_length_ws;
}
UPDATE_STATE(s_header_value_start);
REEXECUTE();
}
@ -1773,11 +1729,6 @@ reexecute:
case h_transfer_encoding_chunked:
parser->flags |= F_CHUNKED;
break;
case h_content_length:
/* do not allow empty content length */
SET_ERRNO(HPE_INVALID_CONTENT_LENGTH);
goto error;
break;
default:
break;
}
@ -1801,18 +1752,13 @@ reexecute:
REEXECUTE();
}
/* Cannot us transfer-encoding and a content-length header together
per the HTTP specification. (RFC 7230 Section 3.3.3) */
if ((parser->extra_flags & (F_TRANSFER_ENCODING >> 8)) &&
/* Cannot use chunked encoding and a content-length header together
per the HTTP specification. */
if ((parser->flags & F_CHUNKED) &&
(parser->flags & F_CONTENTLENGTH)) {
/* Allow it for lenient parsing as long as `Transfer-Encoding` is
* not `chunked`
*/
if (!lenient || (parser->flags & F_CHUNKED)) {
SET_ERRNO(HPE_UNEXPECTED_CONTENT_LENGTH);
goto error;
}
}
UPDATE_STATE(s_headers_done);
@ -1886,31 +1832,8 @@ reexecute:
UPDATE_STATE(NEW_MESSAGE());
CALLBACK_NOTIFY(message_complete);
} else if (parser->flags & F_CHUNKED) {
/* chunked encoding - ignore Content-Length header,
* prepare for a chunk */
/* chunked encoding - ignore Content-Length header */
UPDATE_STATE(s_chunk_size_start);
} else if (parser->extra_flags & (F_TRANSFER_ENCODING >> 8)) {
if (parser->type == HTTP_REQUEST && !lenient) {
/* RFC 7230 3.3.3 */
/* If a Transfer-Encoding header field
* is present in a request and the chunked transfer coding is not
* the final encoding, the message body length cannot be determined
* reliably; the server MUST respond with the 400 (Bad Request)
* status code and then close the connection.
*/
SET_ERRNO(HPE_INVALID_TRANSFER_ENCODING);
RETURN(p - data); /* Error */
} else {
/* RFC 7230 3.3.3 */
/* If a Transfer-Encoding header field is present in a response and
* the chunked transfer coding is not the final encoding, the
* message body length is determined by reading the connection until
* it is closed by the server.
*/
UPDATE_STATE(s_body_identity_eof);
}
} else {
if (parser->content_length == 0) {
/* Content-Length header given but zero: Content-Length: 0\r\n */
@ -2164,12 +2087,6 @@ http_message_needs_eof (const http_parser *parser)
return 0;
}
/* RFC 7230 3.3.3, see `s_headers_almost_done` */
if ((parser->extra_flags & (F_TRANSFER_ENCODING >> 8)) &&
(parser->flags & F_CHUNKED) == 0) {
return 1;
}
if ((parser->flags & F_CHUNKED) || parser->content_length != ULLONG_MAX) {
return 0;
}
@ -2348,14 +2265,14 @@ http_parse_host(const char * buf, struct http_parser_url *u, int found_at) {
switch(new_s) {
case s_http_host:
if (s != s_http_host) {
u->field_data[UF_HOST].off = (uint16_t)(p - buf);
u->field_data[UF_HOST].off = p - buf;
}
u->field_data[UF_HOST].len++;
break;
case s_http_host_v6:
if (s != s_http_host_v6) {
u->field_data[UF_HOST].off = (uint16_t)(p - buf);
u->field_data[UF_HOST].off = p - buf;
}
u->field_data[UF_HOST].len++;
break;
@ -2367,7 +2284,7 @@ http_parse_host(const char * buf, struct http_parser_url *u, int found_at) {
case s_http_host_port:
if (s != s_http_host_port) {
u->field_data[UF_PORT].off = (uint16_t)(p - buf);
u->field_data[UF_PORT].off = p - buf;
u->field_data[UF_PORT].len = 0;
u->field_set |= (1 << UF_PORT);
}
@ -2376,7 +2293,7 @@ http_parse_host(const char * buf, struct http_parser_url *u, int found_at) {
case s_http_userinfo:
if (s != s_http_userinfo) {
u->field_data[UF_USERINFO].off = (uint16_t)(p - buf);
u->field_data[UF_USERINFO].off = p - buf ;
u->field_data[UF_USERINFO].len = 0;
u->field_set |= (1 << UF_USERINFO);
}
@ -2480,7 +2397,7 @@ http_parser_parse_url(const char *buf, size_t buflen, int is_connect,
continue;
}
u->field_data[uf].off = (uint16_t)(p - buf);
u->field_data[uf].off = p - buf;
u->field_data[uf].len = 1;
u->field_set |= (1 << uf);
@ -2561,8 +2478,3 @@ http_parser_version(void) {
HTTP_PARSER_VERSION_MINOR * 0x00100 |
HTTP_PARSER_VERSION_PATCH * 0x00001;
}
void
http_parser_set_max_header_size(uint32_t size) {
max_header_size = size;
}

14
http_parser.gyp
Unescape View File

@ -47,6 +47,10 @@
],
},
'variables': {
'http_max_header_size%': '8192'
},
'targets': [
{
'target_name': 'http_parser',
@ -56,7 +60,10 @@
'defines': [ 'HTTP_PARSER_STRICT=0' ],
'include_dirs': [ '.' ],
},
'defines': [ 'HTTP_PARSER_STRICT=0' ],
'defines': [
'HTTP_MAX_HEADER_SIZE=<(http_max_header_size)',
'HTTP_PARSER_STRICT=0'
],
'sources': [ './http_parser.c', ],
'conditions': [
['OS=="win"', {
@ -79,7 +86,10 @@
'defines': [ 'HTTP_PARSER_STRICT=1' ],
'include_dirs': [ '.' ],
},
'defines': [ 'HTTP_PARSER_STRICT=1' ],
'defines': [
'HTTP_MAX_HEADER_SIZE=<(http_max_header_size)',
'HTTP_PARSER_STRICT=1'
],
'sources': [ './http_parser.c', ],
'conditions': [
['OS=="win"', {

15
http_parser.h
Unescape View File

@ -26,8 +26,8 @@ extern "C" {
/* Also update SONAME in the Makefile whenever you change these. */
#define HTTP_PARSER_VERSION_MAJOR 2
#define HTTP_PARSER_VERSION_MINOR 9
#define HTTP_PARSER_VERSION_PATCH 4
#define HTTP_PARSER_VERSION_MINOR 8
#define HTTP_PARSER_VERSION_PATCH 1
#include <stddef.h>
#if defined(_WIN32) && !defined(__MINGW32__) && \
@ -225,7 +225,6 @@ enum flags
, F_UPGRADE = 1 << 5
, F_SKIPBODY = 1 << 6
, F_CONTENTLENGTH = 1 << 7
, F_TRANSFER_ENCODING = 1 << 8 /* Never set in http_parser.flags */
};
@ -276,9 +275,7 @@ enum flags
XX(INVALID_INTERNAL_STATE, "encountered unexpected internal state")\
XX(STRICT, "strict mode assertion failed") \
XX(PAUSED, "parser is paused") \
XX(UNKNOWN, "an unknown error occurred") \
XX(INVALID_TRANSFER_ENCODING, \
"request has invalid transfer-encoding") \
XX(UNKNOWN, "an unknown error occurred")
/* Define HPE_* values for each errno value above */
@ -299,8 +296,7 @@ struct http_parser {
unsigned int flags : 8; /* F_* values from 'flags' enum; semi-public */
unsigned int state : 7; /* enum state from http_parser.c */
unsigned int header_state : 7; /* enum header_state from http_parser.c */
unsigned int index : 5; /* index into current matcher */
unsigned int extra_flags : 2;
unsigned int index : 7; /* index into current matcher */
unsigned int lenient_http_headers : 1;
uint32_t nread; /* # bytes read in various scenarios */
@ -434,9 +430,6 @@ void http_parser_pause(http_parser *parser, int paused);
/* Checks if this is the final chunk of the body. */
int http_body_is_final(const http_parser *parser);
/* Change the maximum header size provided at compile time. */
void http_parser_set_max_header_size(uint32_t size);
#ifdef __cplusplus
}
#endif

114
test.c
Unescape View File

@ -262,6 +262,7 @@ const struct message requests[] =
,.type= HTTP_REQUEST
,.raw= "POST /post_identity_body_world?q=search#hey HTTP/1.1\r\n"
"Accept: */*\r\n"
"Transfer-Encoding: identity\r\n"
"Content-Length: 5\r\n"
"\r\n"
"World"
@ -274,9 +275,10 @@ const struct message requests[] =
,.fragment= "hey"
,.request_path= "/post_identity_body_world"
,.request_url= "/post_identity_body_world?q=search#hey"
,.num_headers= 2
,.num_headers= 3
,.headers=
{ { "Accept", "*/*" }
, { "Transfer-Encoding", "identity" }
, { "Content-Length", "5" }
}
,.body= "World"
@ -1191,61 +1193,6 @@ const struct message requests[] =
,.headers= { { "Host", "example.com" } }
,.body= ""
}
#define POST_MULTI_TE_LAST_CHUNKED 43
, {.name= "post - multi coding transfer-encoding chunked body"
,.type= HTTP_REQUEST
,.raw= "POST / HTTP/1.1\r\n"
"Transfer-Encoding: deflate, chunked\r\n"
"\r\n"
"1e\r\nall your base are belong to us\r\n"
"0\r\n"
"\r\n"
,.should_keep_alive= TRUE
,.message_complete_on_eof= FALSE
,.http_major= 1
,.http_minor= 1
,.method= HTTP_POST
,.query_string= ""
,.fragment= ""
,.request_path= "/"
,.request_url= "/"
,.num_headers= 1
,.headers=
{ { "Transfer-Encoding" , "deflate, chunked" }
}
,.body= "all your base are belong to us"
,.num_chunks_complete= 2
,.chunk_lengths= { 0x1e }
}
#define POST_MULTI_LINE_TE_LAST_CHUNKED 44
, {.name= "post - multi line coding transfer-encoding chunked body"
,.type= HTTP_REQUEST
,.raw= "POST / HTTP/1.1\r\n"
"Transfer-Encoding: deflate,\r\n"
" chunked\r\n"
"\r\n"
"1e\r\nall your base are belong to us\r\n"
"0\r\n"
"\r\n"
,.should_keep_alive= TRUE
,.message_complete_on_eof= FALSE
,.http_major= 1
,.http_minor= 1
,.method= HTTP_POST
,.query_string= ""
,.fragment= ""
,.request_path= "/"
,.request_url= "/"
,.num_headers= 1
,.headers=
{ { "Transfer-Encoding" , "deflate, chunked" }
}
,.body= "all your base are belong to us"
,.num_chunks_complete= 2
,.chunk_lengths= { 0x1e }
}
};
/* * R E S P O N S E S * */
@ -2023,28 +1970,6 @@ const struct message responses[] =
,.num_chunks_complete= 3
,.chunk_lengths= { 2, 2 }
}
#define HTTP_200_MULTI_TE_NOT_LAST_CHUNKED 28
, {.name= "HTTP 200 response with `chunked` being *not last* Transfer-Encoding"
,.type= HTTP_RESPONSE
,.raw= "HTTP/1.1 200 OK\r\n"
"Transfer-Encoding: chunked, identity\r\n"
"\r\n"
"2\r\n"
"OK\r\n"
"0\r\n"
"\r\n"
,.should_keep_alive= FALSE
,.message_complete_on_eof= TRUE
,.http_major= 1
,.http_minor= 1
,.status_code= 200
,.response_status= "OK"
,.num_headers= 1
,.headers= { { "Transfer-Encoding", "chunked, identity" }
}
,.body= "2\r\nOK\r\n0\r\n\r\n"
,.num_chunks_complete= 0
}
};
/* strnlen() is a POSIX.2008 addition. Can't rely on it being available so
@ -3738,7 +3663,7 @@ test_chunked_content_length_error (int req)
parsed = http_parser_execute(&parser, &settings_null, buf, strlen(buf));
assert(parsed == strlen(buf));
buf = "Transfer-Encoding: anything\r\nContent-Length: 1\r\n\r\n";
buf = "Transfer-Encoding: chunked\r\nContent-Length: 1\r\n\r\n";
size_t buflen = strlen(buf);
parsed = http_parser_execute(&parser, &settings_null, buf, buflen);
@ -4221,7 +4146,6 @@ main (void)
printf("http_parser v%u.%u.%u (0x%06lx)\n", major, minor, patch, version);
printf("sizeof(http_parser) = %u\n", (unsigned int)sizeof(http_parser));
assert(sizeof(http_parser) == 4 + 4 + 8 + 2 + 2 + 4 + sizeof(void *));
//// API
test_preserve_data();
@ -4258,13 +4182,6 @@ main (void)
test_invalid_header_field_token_error(HTTP_RESPONSE);
test_invalid_header_field_content_error(HTTP_RESPONSE);
test_simple_type(
"POST / HTTP/1.1\r\n"
"Content-Length:\r\n" // empty
"\r\n",
HPE_INVALID_CONTENT_LENGTH,
HTTP_REQUEST);
test_simple_type(
"POST / HTTP/1.1\r\n"
"Content-Length: 42 \r\n" // Note the surrounding whitespace.
@ -4286,20 +4203,6 @@ main (void)
HPE_INVALID_CONTENT_LENGTH,
HTTP_REQUEST);
test_simple_type(
"POST / HTTP/1.1\r\n"
"Content-Length: 42\r\n"
" Hello world!\r\n",
HPE_INVALID_CONTENT_LENGTH,
HTTP_REQUEST);
test_simple_type(
"POST / HTTP/1.1\r\n"
"Content-Length: 42\r\n"
" \r\n",
HPE_OK,
HTTP_REQUEST);
//// RESPONSES
test_simple_type("HTP/1.1 200 OK\r\n\r\n", HPE_INVALID_VERSION, HTTP_RESPONSE);
@ -4392,9 +4295,6 @@ main (void)
test_simple("GET / HTTP/11.1\r\n\r\n", HPE_INVALID_VERSION);
test_simple("GET / HTTP/1.01\r\n\r\n", HPE_INVALID_VERSION);
test_simple("GET / HTTP/1.0\r\nHello: w\1rld\r\n\r\n", HPE_INVALID_HEADER_TOKEN);
test_simple("GET / HTTP/1.0\r\nHello: woooo\2rld\r\n\r\n", HPE_INVALID_HEADER_TOKEN);
// Extended characters - see nodejs/test/parallel/test-http-headers-obstext.js
test_simple("GET / HTTP/1.1\r\n"
"Test: Düsseldorf\r\n",
@ -4408,12 +4308,6 @@ main (void)
"fooba",
HPE_OK);
// Unknown Transfer-Encoding in request
test_simple("GET / HTTP/1.1\r\n"
"Transfer-Encoding: unknown\r\n"
"\r\n",
HPE_INVALID_TRANSFER_ENCODING);
static const char *all_methods[] = {
"DELETE",
"GET",

Write Preview
Loading…
Cancel
Save