Fix #77423: parse_url() will deliver a wrong host to userphp-7.3.26PHP-7.3.26

(cherry picked from commit 128fca40376140c60b47a1c3750bb6435866838e)

8 files changed, 63 insertions, 15 deletions

diff --git a/NEWS b/NEWS
index e3c0c8e112..f7418f769e 100644
--- a/NEWS
+++ b/NEWS
@@ -3,6 +3,8 @@ PHP                                                                        NEWS
 07 Jan 2021, PHP 7.3.26
 
 - Standard:
+  . Fixed bug #77423 (FILTER_VALIDATE_URL accepts URLs with invalid userinfo). 
+    (CVE-2020-7071) (cmb)
   . Fixed bug #80457 (stream_get_contents() fails with maxlength=-1 or default).
     (bruno dot premont at restena dot lu)
 
diff --git a/ext/standard/tests/strings/url_t.phpt b/ext/standard/tests/strings/url_t.phpt
index 79ff3bc4a8..f564f59f06 100644
--- a/ext/standard/tests/strings/url_t.phpt
+++ b/ext/standard/tests/strings/url_t.phpt
@@ -575,15 +575,13 @@ $sample_urls = array (
   string(16) "some_page_ref123"
 }
 
---> http://secret@hideout@www.php.net:80/index.php?test=1&test2=char&test3=mixesCI#some_page_ref123: array(7) {
+--> http://secret@hideout@www.php.net:80/index.php?test=1&test2=char&test3=mixesCI#some_page_ref123: array(6) {
   ["scheme"]=>
   string(4) "http"
   ["host"]=>
-  string(11) "www.php.net"
+  string(26) "secret@hideout@www.php.net"
   ["port"]=>
   int(80)
-  ["user"]=>
-  string(14) "secret@hideout"
   ["path"]=>
   string(10) "/index.php"
   ["query"]=>
diff --git a/ext/standard/tests/url/bug77423.phpt b/ext/standard/tests/url/bug77423.phpt
new file mode 100644
index 0000000000..be03fe95e2
--- /dev/null
+++ b/ext/standard/tests/url/bug77423.phpt
@@ -0,0 +1,30 @@
+--TEST--
+Bug #77423 (parse_url() will deliver a wrong host to user)
+--FILE--
+<?php
+$urls = array(
+    "http://php.net\@aliyun.com/aaa.do",
+    "https://example.com\uFF03@bing.com",
+);
+foreach ($urls as $url) {
+    var_dump(filter_var($url, FILTER_VALIDATE_URL));
+    var_dump(parse_url($url));
+}
+?>
+--EXPECT--
+bool(false)
+array(3) {
+  ["scheme"]=>
+  string(4) "http"
+  ["host"]=>
+  string(19) "php.net\@aliyun.com"
+  ["path"]=>
+  string(7) "/aaa.do"
+}
+bool(false)
+array(2) {
+  ["scheme"]=>
+  string(5) "https"
+  ["host"]=>
+  string(26) "example.com\uFF03@bing.com"
+}
diff --git a/ext/standard/tests/url/parse_url_basic_001.phpt b/ext/standard/tests/url/parse_url_basic_001.phpt
index 7ffd001856..57014f0839 100644
--- a/ext/standard/tests/url/parse_url_basic_001.phpt
+++ b/ext/standard/tests/url/parse_url_basic_001.phpt
@@ -506,15 +506,13 @@ echo "Done";
   string(16) "some_page_ref123"
 }
 
---> http://secret@hideout@www.php.net:80/index.php?test=1&test2=char&test3=mixesCI#some_page_ref123: array(7) {
+--> http://secret@hideout@www.php.net:80/index.php?test=1&test2=char&test3=mixesCI#some_page_ref123: array(6) {
   ["scheme"]=>
   string(4) "http"
   ["host"]=>
-  string(11) "www.php.net"
+  string(26) "secret@hideout@www.php.net"
   ["port"]=>
   int(80)
-  ["user"]=>
-  string(14) "secret@hideout"
   ["path"]=>
   string(10) "/index.php"
   ["query"]=>
diff --git a/ext/standard/tests/url/parse_url_basic_003.phpt b/ext/standard/tests/url/parse_url_basic_003.phpt
index 8b0e7eb875..e7cd24c969 100644
--- a/ext/standard/tests/url/parse_url_basic_003.phpt
+++ b/ext/standard/tests/url/parse_url_basic_003.phpt
@@ -68,7 +68,7 @@ echo "Done";
 --> http://secret:@www.php.net/index.php?test=1&test2=char&test3=mixesCI#some_page_ref123   : string(11) "www.php.net"
 --> http://:hideout@www.php.net:80/index.php?test=1&test2=char&test3=mixesCI#some_page_ref123   : string(11) "www.php.net"
 --> http://secret:hideout@www.php.net/index.php?test=1&test2=char&test3=mixesCI#some_page_ref123   : string(11) "www.php.net"
---> http://secret@hideout@www.php.net:80/index.php?test=1&test2=char&test3=mixesCI#some_page_ref123   : string(11) "www.php.net"
+--> http://secret@hideout@www.php.net:80/index.php?test=1&test2=char&test3=mixesCI#some_page_ref123   : string(26) "secret@hideout@www.php.net"
 --> http://secret:hid:out@www.php.net:80/index.php?test=1&test2=char&test3=mixesCI#some_page_ref123   : string(11) "www.php.net"
 --> nntp://news.php.net   : string(12) "news.php.net"
 --> ftp://ftp.gnu.org/gnu/glic/glibc.tar.gz   : string(11) "ftp.gnu.org"
diff --git a/ext/standard/tests/url/parse_url_basic_005.phpt b/ext/standard/tests/url/parse_url_basic_005.phpt
index a5ca381a69..8617d37059 100644
--- a/ext/standard/tests/url/parse_url_basic_005.phpt
+++ b/ext/standard/tests/url/parse_url_basic_005.phpt
@@ -68,7 +68,7 @@ echo "Done";
 --> http://secret:@www.php.net/index.php?test=1&test2=char&test3=mixesCI#some_page_ref123   : string(6) "secret"
 --> http://:hideout@www.php.net:80/index.php?test=1&test2=char&test3=mixesCI#some_page_ref123   : string(0) ""
 --> http://secret:hideout@www.php.net/index.php?test=1&test2=char&test3=mixesCI#some_page_ref123   : string(6) "secret"
---> http://secret@hideout@www.php.net:80/index.php?test=1&test2=char&test3=mixesCI#some_page_ref123   : string(14) "secret@hideout"
+--> http://secret@hideout@www.php.net:80/index.php?test=1&test2=char&test3=mixesCI#some_page_ref123   : NULL
 --> http://secret:hid:out@www.php.net:80/index.php?test=1&test2=char&test3=mixesCI#some_page_ref123   : string(6) "secret"
 --> nntp://news.php.net   : NULL
 --> ftp://ftp.gnu.org/gnu/glic/glibc.tar.gz   : NULL
diff --git a/ext/standard/tests/url/parse_url_unterminated.phpt b/ext/standard/tests/url/parse_url_unterminated.phpt
index 8f46eb3dbe..1377cb7735 100644
--- a/ext/standard/tests/url/parse_url_unterminated.phpt
+++ b/ext/standard/tests/url/parse_url_unterminated.phpt
@@ -508,15 +508,13 @@ echo "Done";
   string(16) "some_page_ref123"
 }
 
---> http://secret@hideout@www.php.net:80/index.php?test=1&test2=char&test3=mixesCI#some_page_ref123: array(7) {
+--> http://secret@hideout@www.php.net:80/index.php?test=1&test2=char&test3=mixesCI#some_page_ref123: array(6) {
   ["scheme"]=>
   string(4) "http"
   ["host"]=>
-  string(11) "www.php.net"
+  string(26) "secret@hideout@www.php.net"
   ["port"]=>
   int(80)
-  ["user"]=>
-  string(14) "secret@hideout"
   ["path"]=>
   string(10) "/index.php"
   ["query"]=>
diff --git a/ext/standard/url.c b/ext/standard/url.c
index 98dc0f786d..157d61466e 100644
--- a/ext/standard/url.c
+++ b/ext/standard/url.c
@@ -102,6 +102,24 @@ static const char *binary_strcspn(const char *s, const char *e, const char *char
 	return e;
 }
 
+static int is_userinfo_valid(const char *str, size_t len)
+{
+	char *valid = "-._~!$&'()*+,;=:";
+	char *p = str;
+	while (p - str < len) {
+		if (isalpha(*p) || isdigit(*p) || strchr(valid, *p)) {
+			p++;
+		} else if (*p == '%' && p - str <= len - 3 && isdigit(*(p+1)) && isxdigit(*(p+2))) {
+			p += 3;
+		} else {
+			return 0;
+		}
+	}
+	return 1;
+}
+
+/* {{{ php_url_parse
+ */
 PHPAPI php_url *php_url_parse_ex(char const *str, size_t length)
 {
 	zend_bool has_port;
@@ -242,13 +260,17 @@ parse_host:
 			ret->pass = zend_string_init(pp, (p-pp), 0);
 			php_replace_controlchars_ex(ZSTR_VAL(ret->pass), ZSTR_LEN(ret->pass));
 		} else {
-			ret->user = zend_string_init(s, (p-s), 0);
+            if (!is_userinfo_valid(s, p-s)) {
+                goto check_port;
+            }
+            ret->user = zend_string_init(s, (p-s), 0);
 			php_replace_controlchars_ex(ZSTR_VAL(ret->user), ZSTR_LEN(ret->user));
 		}
 
 		s = p + 1;
 	}
 
+check_port:
 	/* check for port */
 	if (s < ue && *s == '[' && *(e-1) == ']') {
 		/* Short circuit portscan,