urlapi: allow numerical parts in the host name

It can only be an IPv4 address if all parts are all digits and no more than
four parts, otherwise it is a host name. Even slightly wrong IPv4 will now be
passed through as a host name.

Regression from 17a15d8846 shipped in 8.1.0

Extended test 1560 accordingly.

Reported-by: Pavel Kalyugin
Fixes #11129
Closes #11131
This commit is contained in:
Daniel Stenberg 2023-05-18 00:31:17 +02:00
parent 77c9a9845b
commit 92772e6d39
No known key found for this signature in database
GPG key ID: 5CC908FDB71E12C2
2 changed files with 44 additions and 38 deletions

View file

@ -34,6 +34,7 @@
#include "inet_ntop.h"
#include "strdup.h"
#include "idn.h"
#include "curl_memrchr.h"
/* The last 3 #include files should be in this order */
#include "curl_printf.h"
@ -643,8 +644,8 @@ static CURLUcode hostname_check(struct Curl_URL *u, char *hostname,
* Handle partial IPv4 numerical addresses and different bases, like
* '16843009', '0x7f', '0x7f.1' '0177.1.1.1' etc.
*
* If the given input string is syntactically wrong or any part for example is
* too big, this function returns FALSE and doesn't create any output.
* If the given input string is syntactically wrong IPv4 or any part for
* example is too big, this function returns HOST_NAME.
*
* Output the "normalized" version of that input string in plain quad decimal
* integers.
@ -675,7 +676,7 @@ static int ipv4_normalize(struct dynbuf *host)
unsigned long l;
if(!ISDIGIT(*c))
/* most importantly this doesn't allow a leading plus or minus */
return n ? HOST_BAD : HOST_NAME;
return HOST_NAME;
l = strtoul(c, &endp, 0);
parts[n] = l;
@ -684,7 +685,7 @@ static int ipv4_normalize(struct dynbuf *host)
switch(*c) {
case '.':
if(n == 3)
return HOST_BAD;
return HOST_NAME;
n++;
c++;
break;
@ -694,39 +695,40 @@ static int ipv4_normalize(struct dynbuf *host)
break;
default:
return n ? HOST_BAD : HOST_NAME;
return HOST_NAME;
}
/* overflow */
if((l == ULONG_MAX) && (errno == ERANGE))
return HOST_BAD;
return HOST_NAME;
#if SIZEOF_LONG > 4
/* a value larger than 32 bits */
if(l > UINT_MAX)
return HOST_BAD;
return HOST_NAME;
#endif
}
/* this is a valid IPv4 numerical address */
Curl_dyn_reset(host);
switch(n) {
case 0: /* a -- 32 bits */
Curl_dyn_reset(host);
result = Curl_dyn_addf(host, "%u.%u.%u.%u",
parts[0] >> 24, (parts[0] >> 16) & 0xff,
(parts[0] >> 8) & 0xff, parts[0] & 0xff);
break;
case 1: /* a.b -- 8.24 bits */
if((parts[0] > 0xff) || (parts[1] > 0xffffff))
return HOST_BAD;
return HOST_NAME;
Curl_dyn_reset(host);
result = Curl_dyn_addf(host, "%u.%u.%u.%u",
parts[0], (parts[1] >> 16) & 0xff,
(parts[1] >> 8) & 0xff, parts[1] & 0xff);
break;
case 2: /* a.b.c -- 8.8.16 bits */
if((parts[0] > 0xff) || (parts[1] > 0xff) || (parts[2] > 0xffff))
return HOST_BAD;
return HOST_NAME;
Curl_dyn_reset(host);
result = Curl_dyn_addf(host, "%u.%u.%u.%u",
parts[0], parts[1], (parts[2] >> 8) & 0xff,
parts[2] & 0xff);
@ -734,7 +736,8 @@ static int ipv4_normalize(struct dynbuf *host)
case 3: /* a.b.c.d -- 8.8.8.8 bits */
if((parts[0] > 0xff) || (parts[1] > 0xff) || (parts[2] > 0xff) ||
(parts[3] > 0xff))
return HOST_BAD;
return HOST_NAME;
Curl_dyn_reset(host);
result = Curl_dyn_addf(host, "%u.%u.%u.%u",
parts[0], parts[1], parts[2], parts[3]);
break;
@ -796,6 +799,9 @@ static CURLUcode parse_authority(struct Curl_URL *u,
if(result)
goto out;
if(!Curl_dyn_len(host))
return CURLUE_NO_HOST;
switch(ipv4_normalize(host)) {
case HOST_IPV4:
break;

View file

@ -474,6 +474,13 @@ static const struct testcase get_parts_list[] ={
};
static const struct urltestcase get_url_list[] = {
{"https://1.0x1000000", "https://1.0x1000000/", 0, 0, CURLUE_OK},
{"https://0x7f.1", "https://127.0.0.1/", 0, 0, CURLUE_OK},
{"https://1.2.3.256.com", "https://1.2.3.256.com/", 0, 0, CURLUE_OK},
{"https://10.com", "https://10.com/", 0, 0, CURLUE_OK},
{"https://1.2.com", "https://1.2.com/", 0, 0, CURLUE_OK},
{"https://1.2.3.com", "https://1.2.3.com/", 0, 0, CURLUE_OK},
{"https://1.2.com.99", "https://1.2.com.99/", 0, 0, CURLUE_OK},
{"https://[fe80::0000:20c:29ff:fe9c:409b]:80/moo",
"https://[fe80::20c:29ff:fe9c:409b]:80/moo",
0, 0, CURLUE_OK},
@ -522,22 +529,24 @@ static const struct urltestcase get_url_list[] = {
/* IPv4 trickeries */
{"https://16843009", "https://1.1.1.1/", 0, 0, CURLUE_OK},
{"https://0x7f.1", "https://127.0.0.1/", 0, 0, CURLUE_OK},
{"https://0177.1", "https://127.0.0.1/", 0, 0, CURLUE_OK},
{"https://0111.02.0x3", "https://73.2.0.3/", 0, 0, CURLUE_OK},
{"https://0111.02.0x3.", "https://0111.02.0x3./", 0, 0, CURLUE_OK},
{"https://0111.02.030", "https://73.2.0.24/", 0, 0, CURLUE_OK},
{"https://0111.02.030.", "https://0111.02.030./", 0, 0, CURLUE_OK},
{"https://0xff.0xff.0377.255", "https://255.255.255.255/", 0, 0, CURLUE_OK},
{"https://1.0xffffff", "https://1.255.255.255/", 0, 0, CURLUE_OK},
/* IPv4 numerical overflows or syntax errors will not normalize */
{"https://a127.0.0.1", "https://a127.0.0.1/", 0, 0, CURLUE_OK},
{"https://\xff.127.0.0.1", "https://%FF.127.0.0.1/", 0, CURLU_URLENCODE,
CURLUE_OK},
{"https://127.-0.0.1", "https://127.-0.0.1/", 0, 0, CURLUE_BAD_HOSTNAME},
{"https://127.-0.0.1", "https://127.-0.0.1/", 0, 0, CURLUE_OK},
{"https://127.0. 1", "https://127.0.0.1/", 0, 0, CURLUE_MALFORMED_INPUT},
{"https://1.0x1000000", "https://1.0x1000000/", 0, 0, CURLUE_BAD_HOSTNAME},
{"https://1.2.3.256", "https://1.2.3.256/", 0, 0, CURLUE_BAD_HOSTNAME},
{"https://1.2.3.4.5", "https://1.2.3.4.5/", 0, 0, CURLUE_BAD_HOSTNAME},
{"https://1.2.0x100.3", "https://1.2.0x100.3/", 0, 0, CURLUE_BAD_HOSTNAME},
{"https://4294967296", "https://4294967296/", 0, 0, CURLUE_BAD_HOSTNAME},
{"https://1.2.3.256", "https://1.2.3.256/", 0, 0, CURLUE_OK},
{"https://1.2.3.256.", "https://1.2.3.256./", 0, 0, CURLUE_OK},
{"https://1.2.3.4.5", "https://1.2.3.4.5/", 0, 0, CURLUE_OK},
{"https://1.2.0x100.3", "https://1.2.0x100.3/", 0, 0, CURLUE_OK},
{"https://4294967296", "https://4294967296/", 0, 0, CURLUE_OK},
{"https://123host", "https://123host/", 0, 0, CURLUE_OK},
/* 40 bytes scheme is the max allowed */
{"AAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAAA://hostname/path",
@ -599,20 +608,11 @@ static const struct urltestcase get_url_list[] = {
0, 0, CURLUE_OK},
/* here the password has the semicolon */
{"http://user:pass;word@host/file",
"http://user:pass;word@host/file",
0, 0, CURLUE_OK},
{"file:///file.txt#moo",
"file:///file.txt#moo",
0, 0, CURLUE_OK},
{"file:////file.txt",
"file:////file.txt",
0, 0, CURLUE_OK},
{"file:///file.txt",
"file:///file.txt",
0, 0, CURLUE_OK},
{"file:./",
"file://",
0, 0, CURLUE_BAD_SCHEME},
"http://user:pass;word@host/file", 0, 0, CURLUE_OK},
{"file:///file.txt#moo", "file:///file.txt#moo", 0, 0, CURLUE_OK},
{"file:////file.txt", "file:////file.txt", 0, 0, CURLUE_OK},
{"file:///file.txt", "file:///file.txt", 0, 0, CURLUE_OK},
{"file:./", "file://", 0, 0, CURLUE_OK},
{"http://example.com/hello/../here",
"http://example.com/hello/../here",
CURLU_PATH_AS_IS, 0, CURLUE_OK},
@ -1124,7 +1124,7 @@ static int get_url(void)
}
curl_free(url);
}
else if(rc != get_url_list[i].ucode) {
if(rc != get_url_list[i].ucode) {
fprintf(stderr, "Get URL\nin: %s\nreturned %d (expected %d)\n",
get_url_list[i].in, (int)rc, get_url_list[i].ucode);
error++;
@ -1515,6 +1515,9 @@ int test(char *URL)
{
(void)URL; /* not used */
if(get_url())
return 3;
if(huge())
return 9;
@ -1533,9 +1536,6 @@ int test(char *URL)
if(set_parts())
return 2;
if(get_url())
return 3;
if(get_parts())
return 4;