Add support for certificate name checking

net/base/x509_openssl_util.cc

Index: net/base/x509_openssl_util.cc
diff --git a/net/base/x509_openssl_util.cc b/net/base/x509_openssl_util.cc
index 22ab59aab3ac3a6ff8bf19abb1f0f47517972cfb..4d1bc7d9fa9e4bdb4abaccb72c64ea6d34c12a6d 100644
--- a/net/base/x509_openssl_util.cc
+++ b/net/base/x509_openssl_util.cc
@@ -1,4 +1,4 @@
-// Copyright (c) 2006-2008 The Chromium Authors. All rights reserved.
+// Copyright (c) 2010 The Chromium Authors. All rights reserved.
 // Use of this source code is governed by a BSD-style license that can be
 // found in the LICENSE file.
 
@@ -7,6 +7,7 @@
 #include "base/logging.h"
 #include "base/string_number_conversions.h"
 #include "base/string_piece.h"
+#include "base/string_util.h"
 #include "base/time.h"
 
 namespace net {
@@ -108,6 +109,114 @@ bool ParseDate(ASN1_TIME* x509_time, base::Time* time) {
   return true;
 }
 
+// TODO(joth): Investigate if we can upstream this into the OpenSSL library,
+// to avoid duplicating this logic across projects.
+bool VerifyHostname(const std::string& hostname,
+                    const std::vector<std::string>& cert_names) {
+  DCHECK(!hostname.empty());
+
+  // Simple host name validation. A valid domain name must only contain
+  // alpha, digits, hyphens, and dots. An IP address may have digits and dots,
+  // and also square braces and colons for IPv6 addresses.
+  std::string reference_name;
+  reference_name.reserve(hostname.length());
+
+  bool found_alpha = false;
+  bool found_ip6_chars = false;
+  bool found_hyphen = false;
+  int dot_count = 0;
+
+  size_t first_dot_index = std::string::npos;
+  for (std::string::const_iterator it = hostname.begin();
+      it != hostname.end(); ++it) {

[agl, 2010/11/15 17:55:04]

indentation looks to be off by a space here.

[joth, 2010/11/16 14:01:10]

Done.

+    char c = *it;
+    if (IsAsciiAlpha(c)) {
+      found_alpha = true;
+      c = base::ToLowerASCII(c);
+    } else if (c == '.') {
+      ++dot_count;
+      if (first_dot_index == std::string::npos)
+        first_dot_index = reference_name.length();
+    } else if (c == ':') {
+      found_ip6_chars = true;
+    } else if (c == '-') {
+      found_hyphen = true;
+    } else if (!IsAsciiDigit(c)) {
+      LOG(WARNING) << "Invalid char " << c << " in hostname " << hostname;
+      return false;
+    }
+    reference_name.push_back(c);
+  }
+  DCHECK(!reference_name.empty());
+
+  // TODO(joth): Add IP address support. See http://crbug.com/62973.
+  if (found_ip6_chars || !found_alpha) {
+    NOTIMPLEMENTED() << hostname;
+    return false;
+  }
+
+  // |wildcard_domain| is the remainder of |host| after the leading host
+  // component is stripped off, but includes the leading dot e.g.
+  // "www.f.com" -> ".f.com".
+  // If there is no meaningful domain part to |host| (e.g. it is an IP address
+  // or contains no dots) then |wildcard_domain| will be empty.
+  // We required at least 3 components (i.e. 2 dots) as a basic protection
+  // against too-broad wild-carding.
+  base::StringPiece wildcard_domain;
+  if (found_alpha && !found_ip6_chars && dot_count >= 2) {
+    DCHECK(first_dot_index != std::string::npos);
+    wildcard_domain = reference_name;
+    wildcard_domain.remove_prefix(first_dot_index);
+    DCHECK(wildcard_domain.starts_with("."));
+  }
+
+  for (std::vector<std::string>::const_iterator it = cert_names.begin();
+       it != cert_names.end(); ++it) {
+    // Catch badly corrupt cert names up front.
+    if (it->empty() || it->find('\0') != std::string::npos) {
+      LOG(WARNING) << "Bad name in cert: " << *it;
+      continue;
+    }
+    const std::string cert_name_string(StringToLowerASCII(*it));
+    base::StringPiece cert_match(cert_name_string);
+
+    // Remove trailing dot, if any.
+    if (cert_match.ends_with("."))
+      cert_match.remove_suffix(1);
+
+    // The hostname must be at least as long as the cert name it is matching,
+    // as we require the wildcard (if present) to match at least one character.
+    if (cert_match.length() > reference_name.length())
+      continue;
+
+    if (cert_match == reference_name)
+      return true;
+
+    // Next see if this cert name starts with a wildcard, so long as the
+    // hostname we're matching against has a valid 'domain' part to match.
+    // Note the "-10" version of draft-saintandre-tls-server-id-check allows
+    // the wildcard to appear anywhere in the leftmost  label, rather than

[agl, 2010/11/15 17:55:04]

double space here.

[joth, 2010/11/16 14:01:10]

Done.

+    // requiring it to be the only character. See

[agl, 2010/11/15 17:55:04]

'See' what?

[joth, 2010/11/16 14:01:10]

Done.
See http://crbug.com/60719

+    if (wildcard_domain.empty() || !cert_match.starts_with("*"))
+      continue;
+
+    // Erase the * but not the . from the domain, as we need to include the dot
+    // in the comparison.
+    cert_match.remove_prefix(1);
+
+    // Do character by character comparison on the remainder to see
+    // if we have a wildcard match. This intentionally does no special handling
+    // for any other wildcard characters in |domain|; alternatively it could
+    // detect these and skip those candidate cert names.
+    if (cert_match == wildcard_domain)
+      return true;
+  }
+  DVLOG(1) << "Could not find any match for " << hostname
+           << " (canonicalized as " << reference_name
+           << ") in cert names " << JoinString(cert_names, '|');
+  return false;
+}
+
 } // namespace x509_openssl_util
 
 } // namespace net