// Copyright (c) 2012 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. #include "net/base/net_util.h" #include #include #include #include #include #include #include "build/build_config.h" #if defined(OS_WIN) #include #include #include #include #pragma comment(lib, "iphlpapi.lib") #elif defined(OS_POSIX) #include #include #include #include #if !defined(OS_NACL) #include #if !defined(OS_ANDROID) #include #endif // !defined(OS_NACL) #endif // !defined(OS_ANDROID) #endif // defined(OS_POSIX) #include "base/basictypes.h" #include "base/json/string_escape.h" #include "base/lazy_instance.h" #include "base/logging.h" #include "base/strings/string_number_conversions.h" #include "base/strings/string_piece.h" #include "base/strings/string_split.h" #include "base/strings/string_util.h" #include "base/strings/stringprintf.h" #include "base/strings/utf_string_conversions.h" #include "base/sys_byteorder.h" #include "base/values.h" #include "net/base/dns_util.h" #include "net/base/net_module.h" #include "net/base/registry_controlled_domains/registry_controlled_domain.h" #include "net/grit/net_resources.h" #include "net/http/http_content_disposition.h" #include "url/gurl.h" #include "url/third_party/mozilla/url_parse.h" #include "url/url_canon.h" #include "url/url_canon_ip.h" #if defined(OS_ANDROID) #include "net/android/network_library.h" #endif #if defined(OS_WIN) #include "net/base/winsock_init.h" #endif namespace net { namespace { // The general list of blocked ports. Will be blocked unless a specific // protocol overrides it. (Ex: ftp can use ports 20 and 21) static const int kRestrictedPorts[] = { 1, // tcpmux 7, // echo 9, // discard 11, // systat 13, // daytime 15, // netstat 17, // qotd 19, // chargen 20, // ftp data 21, // ftp access 22, // ssh 23, // telnet 25, // smtp 37, // time 42, // name 43, // nicname 53, // domain 77, // priv-rjs 79, // finger 87, // ttylink 95, // supdup 101, // hostriame 102, // iso-tsap 103, // gppitnp 104, // acr-nema 109, // pop2 110, // pop3 111, // sunrpc 113, // auth 115, // sftp 117, // uucp-path 119, // nntp 123, // NTP 135, // loc-srv /epmap 139, // netbios 143, // imap2 179, // BGP 389, // ldap 465, // smtp+ssl 512, // print / exec 513, // login 514, // shell 515, // printer 526, // tempo 530, // courier 531, // chat 532, // netnews 540, // uucp 556, // remotefs 563, // nntp+ssl 587, // stmp? 601, // ?? 636, // ldap+ssl 993, // ldap+ssl 995, // pop3+ssl 2049, // nfs 3659, // apple-sasl / PasswordServer 4045, // lockd 6000, // X11 6665, // Alternate IRC [Apple addition] 6666, // Alternate IRC [Apple addition] 6667, // Standard IRC [Apple addition] 6668, // Alternate IRC [Apple addition] 6669, // Alternate IRC [Apple addition] 0xFFFF, // Used to block all invalid port numbers (see // third_party/WebKit/Source/platform/weborigin/KURL.cpp, // KURL::port()) }; // FTP overrides the following restricted ports. static const int kAllowedFtpPorts[] = { 21, // ftp data 22, // ssh }; } // namespace static base::LazyInstance >::Leaky g_explicitly_allowed_ports = LAZY_INSTANCE_INITIALIZER; size_t GetCountOfExplicitlyAllowedPorts() { return g_explicitly_allowed_ports.Get().size(); } std::string GetSpecificHeader(const std::string& headers, const std::string& name) { // We want to grab the Value from the "Key: Value" pairs in the headers, // which should look like this (no leading spaces, \n-separated) (we format // them this way in url_request_inet.cc): // HTTP/1.1 200 OK\n // ETag: "6d0b8-947-24f35ec0"\n // Content-Length: 2375\n // Content-Type: text/html; charset=UTF-8\n // Last-Modified: Sun, 03 Sep 2006 04:34:43 GMT\n if (headers.empty()) return std::string(); std::string match('\n' + name + ':'); std::string::const_iterator begin = std::search(headers.begin(), headers.end(), match.begin(), match.end(), base::CaseInsensitiveCompareASCII()); if (begin == headers.end()) return std::string(); begin += match.length(); std::string ret; base::TrimWhitespace(std::string(begin, std::find(begin, headers.end(), '\n')), base::TRIM_ALL, &ret); return ret; } std::string CanonicalizeHost(const std::string& host, url::CanonHostInfo* host_info) { // Try to canonicalize the host. const url::Component raw_host_component(0, static_cast(host.length())); std::string canon_host; url::StdStringCanonOutput canon_host_output(&canon_host); url::CanonicalizeHostVerbose(host.c_str(), raw_host_component, &canon_host_output, host_info); if (host_info->out_host.is_nonempty() && host_info->family != url::CanonHostInfo::BROKEN) { // Success! Assert that there's no extra garbage. canon_host_output.Complete(); DCHECK_EQ(host_info->out_host.len, static_cast(canon_host.length())); } else { // Empty host, or canonicalization failed. We'll return empty. canon_host.clear(); } return canon_host; } std::string GetDirectoryListingHeader(const base::string16& title) { static const base::StringPiece header( NetModule::GetResource(IDR_DIR_HEADER_HTML)); // This can be null in unit tests. DLOG_IF(WARNING, header.empty()) << "Missing resource: directory listing header"; std::string result; if (!header.empty()) result.assign(header.data(), header.size()); result.append("\n"); return result; } inline bool IsHostCharAlphanumeric(char c) { // We can just check lowercase because uppercase characters have already been // normalized. return ((c >= 'a') && (c <= 'z')) || ((c >= '0') && (c <= '9')); } bool IsCanonicalizedHostCompliant(const std::string& host) { if (host.empty()) return false; bool in_component = false; bool most_recent_component_started_alphanumeric = false; for (std::string::const_iterator i(host.begin()); i != host.end(); ++i) { const char c = *i; if (!in_component) { most_recent_component_started_alphanumeric = IsHostCharAlphanumeric(c); if (!most_recent_component_started_alphanumeric && (c != '-') && (c != '_')) { return false; } in_component = true; } else if (c == '.') { in_component = false; } else if (!IsHostCharAlphanumeric(c) && (c != '-') && (c != '_')) { return false; } } return most_recent_component_started_alphanumeric; } base::string16 StripWWW(const base::string16& text) { const base::string16 www(base::ASCIIToUTF16("www.")); return StartsWith(text, www, true) ? text.substr(www.length()) : text; } base::string16 StripWWWFromHost(const GURL& url) { DCHECK(url.is_valid()); return StripWWW(base::ASCIIToUTF16(url.host())); } bool IsPortValid(int port) { return port >= 0 && port <= std::numeric_limits::max(); } bool IsPortAllowedByDefault(int port) { int array_size = arraysize(kRestrictedPorts); for (int i = 0; i < array_size; i++) { if (kRestrictedPorts[i] == port) { return false; } } return IsPortValid(port); } bool IsPortAllowedByFtp(int port) { int array_size = arraysize(kAllowedFtpPorts); for (int i = 0; i < array_size; i++) { if (kAllowedFtpPorts[i] == port) { return true; } } // Port not explicitly allowed by FTP, so return the default restrictions. return IsPortAllowedByDefault(port); } bool IsPortAllowedByOverride(int port) { if (g_explicitly_allowed_ports.Get().empty()) return false; return g_explicitly_allowed_ports.Get().count(port) > 0; } int SetNonBlocking(int fd) { #if defined(OS_WIN) unsigned long no_block = 1; return ioctlsocket(fd, FIONBIO, &no_block); #elif defined(OS_POSIX) int flags = fcntl(fd, F_GETFL, 0); if (-1 == flags) return flags; return fcntl(fd, F_SETFL, flags | O_NONBLOCK); #endif } bool ParseHostAndPort(std::string::const_iterator host_and_port_begin, std::string::const_iterator host_and_port_end, std::string* host, int* port) { if (host_and_port_begin >= host_and_port_end) return false; // When using url, we use char*. const char* auth_begin = &(*host_and_port_begin); int auth_len = host_and_port_end - host_and_port_begin; url::Component auth_component(0, auth_len); url::Component username_component; url::Component password_component; url::Component hostname_component; url::Component port_component; url::ParseAuthority(auth_begin, auth_component, &username_component, &password_component, &hostname_component, &port_component); // There shouldn't be a username/password. if (username_component.is_valid() || password_component.is_valid()) return false; if (!hostname_component.is_nonempty()) return false; // Failed parsing. int parsed_port_number = -1; if (port_component.is_nonempty()) { parsed_port_number = url::ParsePort(auth_begin, port_component); // If parsing failed, port_number will be either PORT_INVALID or // PORT_UNSPECIFIED, both of which are negative. if (parsed_port_number < 0) return false; // Failed parsing the port number. } if (port_component.len == 0) return false; // Reject inputs like "foo:" unsigned char tmp_ipv6_addr[16]; // If the hostname starts with a bracket, it is either an IPv6 literal or // invalid. If it is an IPv6 literal then strip the brackets. if (hostname_component.len > 0 && auth_begin[hostname_component.begin] == '[') { if (auth_begin[hostname_component.end() - 1] == ']' && url::IPv6AddressToNumber( auth_begin, hostname_component, tmp_ipv6_addr)) { // Strip the brackets. hostname_component.begin++; hostname_component.len -= 2; } else { return false; } } // Pass results back to caller. host->assign(auth_begin + hostname_component.begin, hostname_component.len); *port = parsed_port_number; return true; // Success. } bool ParseHostAndPort(const std::string& host_and_port, std::string* host, int* port) { return ParseHostAndPort( host_and_port.begin(), host_and_port.end(), host, port); } std::string GetHostAndPort(const GURL& url) { // For IPv6 literals, GURL::host() already includes the brackets so it is // safe to just append a colon. return base::StringPrintf("%s:%d", url.host().c_str(), url.EffectiveIntPort()); } std::string GetHostAndOptionalPort(const GURL& url) { // For IPv6 literals, GURL::host() already includes the brackets // so it is safe to just append a colon. if (url.has_port()) return base::StringPrintf("%s:%s", url.host().c_str(), url.port().c_str()); return url.host(); } bool IsHostnameNonUnique(const std::string& hostname) { // CanonicalizeHost requires surrounding brackets to parse an IPv6 address. const std::string host_or_ip = hostname.find(':') != std::string::npos ? "[" + hostname + "]" : hostname; url::CanonHostInfo host_info; std::string canonical_name = CanonicalizeHost(host_or_ip, &host_info); // If canonicalization fails, then the input is truly malformed. However, // to avoid mis-reporting bad inputs as "non-unique", treat them as unique. if (canonical_name.empty()) return false; // If |hostname| is an IP address, check to see if it's in an IANA-reserved // range. if (host_info.IsIPAddress()) { IPAddressNumber host_addr; if (!ParseIPLiteralToNumber(hostname.substr(host_info.out_host.begin, host_info.out_host.len), &host_addr)) { return false; } switch (host_info.family) { case url::CanonHostInfo::IPV4: case url::CanonHostInfo::IPV6: return IsIPAddressReserved(host_addr); case url::CanonHostInfo::NEUTRAL: case url::CanonHostInfo::BROKEN: return false; } } // Check for a registry controlled portion of |hostname|, ignoring private // registries, as they already chain to ICANN-administered registries, // and explicitly ignoring unknown registries. // // Note: This means that as new gTLDs are introduced on the Internet, they // will be treated as non-unique until the registry controlled domain list // is updated. However, because gTLDs are expected to provide significant // advance notice to deprecate older versions of this code, this an // acceptable tradeoff. return 0 == registry_controlled_domains::GetRegistryLength( canonical_name, registry_controlled_domains::EXCLUDE_UNKNOWN_REGISTRIES, registry_controlled_domains::EXCLUDE_PRIVATE_REGISTRIES); } SockaddrStorage::SockaddrStorage(const SockaddrStorage& other) : addr_len(other.addr_len), addr(reinterpret_cast(&addr_storage)) { memcpy(addr, other.addr, addr_len); } void SockaddrStorage::operator=(const SockaddrStorage& other) { addr_len = other.addr_len; // addr is already set to &this->addr_storage by default ctor. memcpy(addr, other.addr, addr_len); } // Extracts the address and port portions of a sockaddr. bool GetIPAddressFromSockAddr(const struct sockaddr* sock_addr, socklen_t sock_addr_len, const uint8** address, size_t* address_len, uint16* port) { if (sock_addr->sa_family == AF_INET) { if (sock_addr_len < static_cast(sizeof(struct sockaddr_in))) return false; const struct sockaddr_in* addr = reinterpret_cast(sock_addr); *address = reinterpret_cast(&addr->sin_addr); *address_len = kIPv4AddressSize; if (port) *port = base::NetToHost16(addr->sin_port); return true; } if (sock_addr->sa_family == AF_INET6) { if (sock_addr_len < static_cast(sizeof(struct sockaddr_in6))) return false; const struct sockaddr_in6* addr = reinterpret_cast(sock_addr); *address = reinterpret_cast(&addr->sin6_addr); *address_len = kIPv6AddressSize; if (port) *port = base::NetToHost16(addr->sin6_port); return true; } #if defined(OS_WIN) if (sock_addr->sa_family == AF_BTH) { if (sock_addr_len < static_cast(sizeof(SOCKADDR_BTH))) return false; const SOCKADDR_BTH* addr = reinterpret_cast(sock_addr); *address = reinterpret_cast(&addr->btAddr); *address_len = kBluetoothAddressSize; if (port) *port = static_cast(addr->port); return true; } #endif return false; // Unrecognized |sa_family|. } std::string NetAddressToString(const struct sockaddr* sa, socklen_t sock_addr_len) { const uint8* address; size_t address_len; if (!GetIPAddressFromSockAddr(sa, sock_addr_len, &address, &address_len, NULL)) { NOTREACHED(); return std::string(); } return IPAddressToString(address, address_len); } std::string NetAddressToStringWithPort(const struct sockaddr* sa, socklen_t sock_addr_len) { const uint8* address; size_t address_len; uint16 port; if (!GetIPAddressFromSockAddr(sa, sock_addr_len, &address, &address_len, &port)) { NOTREACHED(); return std::string(); } return IPAddressToStringWithPort(address, address_len, port); } std::string GetHostName() { #if defined(OS_NACL) NOTIMPLEMENTED(); return std::string(); #else // defined(OS_NACL) #if defined(OS_WIN) EnsureWinsockInit(); #endif // Host names are limited to 255 bytes. char buffer[256]; int result = gethostname(buffer, sizeof(buffer)); if (result != 0) { DVLOG(1) << "gethostname() failed with " << result; buffer[0] = '\0'; } return std::string(buffer); #endif // !defined(OS_NACL) } void GetIdentityFromURL(const GURL& url, base::string16* username, base::string16* password) { UnescapeRule::Type flags = UnescapeRule::SPACES | UnescapeRule::URL_SPECIAL_CHARS; *username = UnescapeAndDecodeUTF8URLComponent(url.username(), flags); *password = UnescapeAndDecodeUTF8URLComponent(url.password(), flags); } std::string GetHostOrSpecFromURL(const GURL& url) { return url.has_host() ? TrimEndingDot(url.host()) : url.spec(); } bool CanStripTrailingSlash(const GURL& url) { // Omit the path only for standard, non-file URLs with nothing but "/" after // the hostname. return url.IsStandard() && !url.SchemeIsFile() && !url.SchemeIsFileSystem() && !url.has_query() && !url.has_ref() && url.path() == "/"; } GURL SimplifyUrlForRequest(const GURL& url) { DCHECK(url.is_valid()); GURL::Replacements replacements; replacements.ClearUsername(); replacements.ClearPassword(); replacements.ClearRef(); return url.ReplaceComponents(replacements); } // Specifies a comma separated list of port numbers that should be accepted // despite bans. If the string is invalid no allowed ports are stored. void SetExplicitlyAllowedPorts(const std::string& allowed_ports) { if (allowed_ports.empty()) return; std::multiset ports; size_t last = 0; size_t size = allowed_ports.size(); // The comma delimiter. const std::string::value_type kComma = ','; // Overflow is still possible for evil user inputs. for (size_t i = 0; i <= size; ++i) { // The string should be composed of only digits and commas. if (i != size && !IsAsciiDigit(allowed_ports[i]) && (allowed_ports[i] != kComma)) return; if (i == size || allowed_ports[i] == kComma) { if (i > last) { int port; base::StringToInt(base::StringPiece(allowed_ports.begin() + last, allowed_ports.begin() + i), &port); ports.insert(port); } last = i + 1; } } g_explicitly_allowed_ports.Get() = ports; } ScopedPortException::ScopedPortException(int port) : port_(port) { g_explicitly_allowed_ports.Get().insert(port); } ScopedPortException::~ScopedPortException() { std::multiset::iterator it = g_explicitly_allowed_ports.Get().find(port_); if (it != g_explicitly_allowed_ports.Get().end()) g_explicitly_allowed_ports.Get().erase(it); else NOTREACHED(); } bool HaveOnlyLoopbackAddresses() { #if defined(OS_ANDROID) return android::HaveOnlyLoopbackAddresses(); #elif defined(OS_NACL) NOTIMPLEMENTED(); return false; #elif defined(OS_POSIX) struct ifaddrs* interface_addr = NULL; int rv = getifaddrs(&interface_addr); if (rv != 0) { DVLOG(1) << "getifaddrs() failed with errno = " << errno; return false; } bool result = true; for (struct ifaddrs* interface = interface_addr; interface != NULL; interface = interface->ifa_next) { if (!(IFF_UP & interface->ifa_flags)) continue; if (IFF_LOOPBACK & interface->ifa_flags) continue; const struct sockaddr* addr = interface->ifa_addr; if (!addr) continue; if (addr->sa_family == AF_INET6) { // Safe cast since this is AF_INET6. const struct sockaddr_in6* addr_in6 = reinterpret_cast(addr); const struct in6_addr* sin6_addr = &addr_in6->sin6_addr; if (IN6_IS_ADDR_LOOPBACK(sin6_addr) || IN6_IS_ADDR_LINKLOCAL(sin6_addr)) continue; } if (addr->sa_family != AF_INET6 && addr->sa_family != AF_INET) continue; result = false; break; } freeifaddrs(interface_addr); return result; #elif defined(OS_WIN) // TODO(wtc): implement with the GetAdaptersAddresses function. NOTIMPLEMENTED(); return false; #else NOTIMPLEMENTED(); return false; #endif // defined(various platforms) } AddressFamily GetAddressFamily(const IPAddressNumber& address) { switch (address.size()) { case kIPv4AddressSize: return ADDRESS_FAMILY_IPV4; case kIPv6AddressSize: return ADDRESS_FAMILY_IPV6; default: return ADDRESS_FAMILY_UNSPECIFIED; } } int ConvertAddressFamily(AddressFamily address_family) { switch (address_family) { case ADDRESS_FAMILY_UNSPECIFIED: return AF_UNSPEC; case ADDRESS_FAMILY_IPV4: return AF_INET; case ADDRESS_FAMILY_IPV6: return AF_INET6; } NOTREACHED(); return AF_UNSPEC; } const uint16* GetPortFieldFromSockaddr(const struct sockaddr* address, socklen_t address_len) { if (address->sa_family == AF_INET) { DCHECK_LE(sizeof(sockaddr_in), static_cast(address_len)); const struct sockaddr_in* sockaddr = reinterpret_cast(address); return &sockaddr->sin_port; } else if (address->sa_family == AF_INET6) { DCHECK_LE(sizeof(sockaddr_in6), static_cast(address_len)); const struct sockaddr_in6* sockaddr = reinterpret_cast(address); return &sockaddr->sin6_port; } else { NOTREACHED(); return NULL; } } int GetPortFromSockaddr(const struct sockaddr* address, socklen_t address_len) { const uint16* port_field = GetPortFieldFromSockaddr(address, address_len); if (!port_field) return -1; return base::NetToHost16(*port_field); } bool IsLocalhost(const std::string& host) { if (host == "localhost" || host == "localhost.localdomain" || host == "localhost6" || host == "localhost6.localdomain6" || IsLocalhostTLD(host)) return true; IPAddressNumber ip_number; if (ParseIPLiteralToNumber(host, &ip_number)) { size_t size = ip_number.size(); switch (size) { case kIPv4AddressSize: { IPAddressNumber localhost_prefix; localhost_prefix.push_back(127); for (int i = 0; i < 3; ++i) { localhost_prefix.push_back(0); } return IPNumberMatchesPrefix(ip_number, localhost_prefix, 8); } case kIPv6AddressSize: { struct in6_addr sin6_addr; memcpy(&sin6_addr, &ip_number[0], kIPv6AddressSize); return !!IN6_IS_ADDR_LOOPBACK(&sin6_addr); } default: NOTREACHED(); } } return false; } bool IsLocalhostTLD(const std::string& host) { const char kLocalhostTLD[] = ".localhost"; const size_t kLocalhostTLDLength = arraysize(kLocalhostTLD) - 1; if (host.empty()) return false; size_t host_len = host.size(); if (*host.rbegin() == '.') --host_len; if (host_len < kLocalhostTLDLength) return false; const char* host_suffix = host.data() + host_len - kLocalhostTLDLength; return base::strncasecmp(host_suffix, kLocalhostTLD, kLocalhostTLDLength) == 0; } bool HasGoogleHost(const GURL& url) { static const char* kGoogleHostSuffixes[] = { ".google.com", ".youtube.com", ".gmail.com", ".doubleclick.net", ".gstatic.com", ".googlevideo.com", ".googleusercontent.com", ".googlesyndication.com", ".google-analytics.com", ".googleadservices.com", ".googleapis.com", ".ytimg.com", }; const std::string& host = url.host(); for (const char* suffix : kGoogleHostSuffixes) { if (EndsWith(host, suffix, false)) return true; } return false; } NetworkInterface::NetworkInterface() : type(NetworkChangeNotifier::CONNECTION_UNKNOWN), prefix_length(0) { } NetworkInterface::NetworkInterface(const std::string& name, const std::string& friendly_name, uint32 interface_index, NetworkChangeNotifier::ConnectionType type, const IPAddressNumber& address, uint32 prefix_length, int ip_address_attributes) : name(name), friendly_name(friendly_name), interface_index(interface_index), type(type), address(address), prefix_length(prefix_length), ip_address_attributes(ip_address_attributes) { } NetworkInterface::~NetworkInterface() { } ScopedWifiOptions::~ScopedWifiOptions() { } } // namespace net