GnuTLS 3.5.18

Table of Contents

GnuTLS

This manual is last updated 7 January 2018 for version 3.5.18 of GnuTLS.

Copyright © 2001-2018 Free Software Foundation, Inc.\\ Copyright © 2001-2018 Nikos Mavrogiannopoulos

Permission is granted to copy, distribute and/or modify this document under the terms of the GNU Free Documentation License, Version 1.3 or any later version published by the Free Software Foundation; with no Invariant Sections, no Front-Cover Texts, and no Back-Cover Texts. A copy of the license is included in the section entitled “GNU Free Documentation License”.

1 Preface

This document demonstrates and explains the GnuTLS library API. A brief introduction to the protocols and the technology involved is also included so that an application programmer can better understand the GnuTLS purpose and actual offerings. Even if GnuTLS is a typical library software, it operates over several security and cryptographic protocols which require the programmer to make careful and correct usage of them. Otherwise it is likely to only obtain a false sense of security. The term of security is very broad even if restricted to computer software, and cannot be confined to a single cryptographic library. For that reason, do not consider any program secure just because it uses GnuTLS; there are several ways to compromise a program or a communication line and GnuTLS only helps with some of them.

Although this document tries to be self contained, basic network programming and public key infrastructure (PKI) knowledge is assumed in most of it. A good introduction to networking can be found in [STEVENS], to public key infrastructure in [GUTPKI] and to security engineering in [ANDERSON].

Updated versions of the GnuTLS software and this document will be available from http://www.gnutls.org/.

2 Introduction to GnuTLS

In brief GnuTLS can be described as a library which offers an API to access secure communication protocols. These protocols provide privacy over insecure lines, and were designed to prevent eavesdropping, tampering, or message forgery.

Technically GnuTLS is a portable ANSI C based library which implements the protocols ranging from SSL 3.0 to TLS 1.2 (see Introduction to TLS, for a detailed description of the protocols), accompanied with the required framework for authentication and public key infrastructure. Important features of the GnuTLS library include:

• Support for TLS 1.2, TLS 1.1, TLS 1.0 and SSL 3.0 protocols.
• Support for Datagram TLS 1.0 and 1.2.
• Support for handling and verification of X.509 and OpenPGP certificates.
• Support for password authentication using TLS-SRP.
• Support for keyed authentication using TLS-PSK.
• Support for TPM, PKCS #11 tokens and smart-cards.

The GnuTLS library consists of three independent parts, namely the “TLS protocol part”, the “Certificate part”, and the “Cryptographic back-end” part. The “TLS protocol part” is the actual protocol implementation, and is entirely implemented within the GnuTLS library. The “Certificate part” consists of the certificate parsing, and verification functions and it uses functionality from the libtasn1 library. The “Cryptographic back-end” is provided by the nettle and gmplib libraries.

2.1 Downloading and installing

GnuTLS is available for download at: http://www.gnutls.org/download.html

GnuTLS uses a development cycle where even minor version numbers indicate a stable release and a odd minor version number indicate a development release. For example, GnuTLS 1.6.3 denote a stable release since 6 is even, and GnuTLS 1.7.11 denote a development release since 7 is odd.

GnuTLS depends on nettle and gmplib, and you will need to install it before installing GnuTLS. The nettle library is available from http://www.lysator.liu.se/~nisse/nettle/, while gmplib is available from http://www.gmplib.org/. Don’t forget to verify the cryptographic signature after downloading source code packages.

The package is then extracted, configured and built like many other packages that use Autoconf. For detailed information on configuring and building it, refer to the INSTALL file that is part of the distribution archive. Typically you invoke ./configure and then make check install. There are a number of compile-time parameters, as discussed below.

Several parts of GnuTLS require ASN.1 functionality, which is provided by a library called libtasn1. A copy of libtasn1 is included in GnuTLS. If you want to install it separately (e.g., to make it possibly to use libtasn1 in other programs), you can get it from http://www.gnu.org/software/libtasn1/.

The compression library, libz, the PKCS #11 helper library p11-kit, the TPM library trousers, as well as the IDN library libidn¹ are optional dependencies. Check the README file in the distribution on how to obtain these libraries.

A few configure options may be relevant, summarized below. They disable or enable particular features, to create a smaller library with only the required features. Note however, that although a smaller library is generated, the included programs are not guaranteed to compile if some of these options are given.

--disable-srp-authentication
--disable-psk-authentication
--disable-anon-authentication
--disable-openpgp-authentication
--disable-dhe
--disable-ecdhe
--disable-openssl-compatibility
--disable-dtls-srtp-support
--disable-alpn-support
--disable-heartbeat-support
--disable-libdane
--without-p11-kit
--without-tpm
--without-zlib

For the complete list, refer to the output from configure --help.

2.2 Installing for a software distribution

When installing for a software distribution, it is often desirable to preconfigure GnuTLS with the system-wide paths and files. There two important configuration options, one sets the trust store in system, which are the CA certificates to be used by programs by default (if they don’t override it), and the other sets to DNSSEC root key file used by unbound for DNSSEC verification.

For the latter the following configuration option is available, and if not specified GnuTLS will try to auto-detect the location of that file.

--with-unbound-root-key-file

To set the trust store the following options are available.

--with-default-trust-store-file
--with-default-trust-store-dir
--with-default-trust-store-pkcs11

The first option is used to set a PEM file which contains a list of trusted certificates, while the second will read all certificates in the given path. The recommended option is the last, which allows to use a PKCS #11 trust policy module. That module not only provides the trusted certificates, but allows the categorization of them using purpose, e.g., CAs can be restricted for e-mail usage only, or administrative restrictions of CAs, for examples by restricting a CA to only issue certificates for a given DNS domain using NameConstraints. A publicly available PKCS #11 trust module is p11-kit’s trust module².

2.3 Overview

In this document we present an overview of the supported security protocols in Introduction to TLS, and continue by providing more information on the certificate authentication in Certificate authentication, and shared-key as well anonymous authentication in Shared-key and anonymous authentication. We elaborate on certificate authentication by demonstrating advanced usage of the API in More on certificate authentication. The core of the TLS library is presented in How to use GnuTLS in applications and example applications are listed in GnuTLS application examples. In Other included programs the usage of few included programs that may assist debugging is presented. The last chapter is Internal architecture of GnuTLS that provides a short introduction to GnuTLS’ internal architecture.

3 Introduction to TLS and DTLS

TLS stands for “Transport Layer Security” and is the successor of SSL, the Secure Sockets Layer protocol [SSL3] designed by Netscape. TLS is an Internet protocol, defined by IETF³, described in [RFC5246]. The protocol provides confidentiality, and authentication layers over any reliable transport layer. The description, above, refers to TLS 1.0 but applies to all other TLS versions as the differences between the protocols are not major.

The DTLS protocol, or “Datagram TLS” [RFC4347] is a protocol with identical goals as TLS, but can operate under unreliable transport layers such as UDP. The discussions below apply to this protocol as well, except when noted otherwise.

3.1 TLS Layers

TLS is a layered protocol, and consists of the record protocol, the handshake protocol and the alert protocol. The record protocol is to serve all other protocols and is above the transport layer. The record protocol offers symmetric encryption, data authenticity, and optionally compression. The alert protocol offers some signaling to the other protocols. It can help informing the peer for the cause of failures and other error conditions. See The Alert Protocol, for more information. The alert protocol is above the record protocol.

The handshake protocol is responsible for the security parameters’ negotiation, the initial key exchange and authentication. See The Handshake Protocol, for more information about the handshake protocol. The protocol layering in TLS is shown in Figure 3.1.

Figure 3.1: The TLS protocol layers.

3.2 The Transport Layer

TLS is not limited to any transport layer and can be used above any transport layer, as long as it is a reliable one. DTLS can be used over reliable and unreliable transport layers. GnuTLS supports TCP and UDP layers transparently using the Berkeley sockets API. However, any transport layer can be used by providing callbacks for GnuTLS to access the transport layer (for details see Setting up the transport layer).

3.3 The TLS record protocol

The record protocol is the secure communications provider. Its purpose is to encrypt, authenticate and —optionally— compress packets. The record layer functions can be called at any time after the handshake process is finished, when there is need to receive or send data. In DTLS however, due to re-transmission timers used in the handshake out-of-order handshake data might be received for some time (maximum 60 seconds) after the handshake process is finished.

The functions to access the record protocol are limited to send and receive functions, which might, given the importance of this protocol in TLS, seem awkward. This is because the record protocol’s parameters are all set by the handshake protocol. The record protocol initially starts with NULL parameters, which means no encryption, and no MAC is used. Encryption and authentication begin just after the handshake protocol has finished.

3.3.1 Encryption algorithms used in the record layer

Confidentiality in the record layer is achieved by using symmetric ciphers like AES or CHACHA20. Ciphers are encryption algorithms that use a single, secret, key to encrypt and decrypt data. Early versions of TLS separated between block and stream ciphers and had message authentication plugged in to them by the protocol, though later versions switched to using authenticated-encryption (AEAD) ciphers. The AEAD ciphers are defined to combine encryption and authentication, and as such they are not only more efficient, as the primitives used are designed to interoperate nicely, but they are also known to interoperate in a secure way.

The supported in GnuTLS ciphers and MAC algorithms are shown in Table 3.1 and Table 3.2.

3.3.2 Compression algorithms used in the record layer

The TLS record layer also supports compression. The algorithms implemented in GnuTLS can be found in the table below. The included algorithms perform really good when text, or other compressible data are to be transferred, but offer nothing on already compressed data, such as compressed images, zipped archives etc. These compression algorithms, may be useful in high bandwidth TLS tunnels, and in cases where network usage has to be minimized. It should be noted however that compression increases latency.

The record layer compression in GnuTLS is implemented based on [RFC3749]. The supported algorithms are shown below.

Note that compression enables attacks such as traffic analysis, or even plaintext recovery under certain circumstances. To avoid some of these attacks GnuTLS allows each record to be compressed independently (i.e., stateless compression), by using the "%STATELESS_COMPRESSION" priority string, in order to be used in cases where the attacker controlled data are pt in separate records.

3.3.3 Weaknesses and countermeasures

Some weaknesses that may affect the security of the record layer have been found in TLS 1.0 protocol. These weaknesses can be exploited by active attackers, and exploit the facts that

1. TLS has separate alerts for “decryption_failed” and “bad_record_mac”
2. The decryption failure reason can be detected by timing the response time.
3. The IV for CBC encrypted packets is the last block of the previous encrypted packet.

Those weaknesses were solved in TLS 1.1 [RFC4346] which is implemented in GnuTLS. For this reason we suggest to always negotiate the highest supported TLS version with the peer⁴. For a detailed discussion of the issues see the archives of the TLS Working Group mailing list and [CBCATT].

3.3.4 On record padding

The TLS protocol allows for extra padding of records in CBC ciphers, to prevent statistical analysis based on the length of exchanged messages (see [RFC5246] section 6.2.3.2). GnuTLS appears to be one of few implementations that take advantage of this feature: the user can provide some plaintext data with a range of lengths she wishes to hide, and GnuTLS adds extra padding to make sure the attacker cannot tell the real plaintext length is in a range smaller than the user-provided one. Use gnutls_record_send_range to send length-hidden messages and gnutls_record_can_use_length_hiding to check whether the current session supports length hiding. Using the standard gnutls_record_send will only add minimal padding.

The TLS implementation in the Symbian operating system, frequently used by Nokia and Sony-Ericsson mobile phones, cannot handle non-minimal record padding. What happens when one of these clients handshake with a GnuTLS server is that the client will fail to compute the correct MAC for the record. The client sends a TLS alert (bad_record_mac) and disconnects. Typically this will result in error messages such as ’A TLS fatal alert has been received’, ’Bad record MAC’, or both, on the GnuTLS server side.

If compatibility with such devices is a concern, not sending length-hidden messages solves the problem by using minimal padding.

If you implement an application that has a configuration file, we recommend that you make it possible for users or administrators to specify a GnuTLS protocol priority string, which is used by your application via gnutls_priority_set. To allow the best flexibility, make it possible to have a different priority string for different incoming IP addresses.

3.4 The TLS alert protocol

The alert protocol is there to allow signals to be sent between peers. These signals are mostly used to inform the peer about the cause of a protocol failure. Some of these signals are used internally by the protocol and the application protocol does not have to cope with them (e.g. GNUTLS_A_CLOSE_NOTIFY), and others refer to the application protocol solely (e.g. GNUTLS_A_USER_CANCELLED). An alert signal includes a level indication which may be either fatal or warning. Fatal alerts always terminate the current connection, and prevent future re-negotiations using the current session ID. All alert messages are summarized in the table below.

The alert messages are protected by the record protocol, thus the information that is included does not leak. You must take extreme care for the alert information not to leak to a possible attacker, via public log files etc.

3.5 The TLS handshake protocol

The handshake protocol is responsible for the ciphersuite negotiation, the initial key exchange, and the authentication of the two peers. This is fully controlled by the application layer, thus your program has to set up the required parameters. The main handshake function is gnutls_handshake. In the next paragraphs we elaborate on the handshake protocol, i.e., the ciphersuite negotiation.

3.5.1 TLS ciphersuites

The handshake protocol of TLS negotiates cipher suites of a special form illustrated by the TLS_DHE_RSA_WITH_3DES_CBC_SHA cipher suite name. A typical cipher suite contains these parameters:

• The key exchange algorithm. DHE_RSA in the example.
• The Symmetric encryption algorithm and mode 3DES_CBC in this example.
• The MAC⁵ algorithm used for authentication. MAC_SHA is used in the above example.

The cipher suite negotiated in the handshake protocol will affect the record protocol, by enabling encryption and data authentication. Note that you should not over rely on TLS to negotiate the strongest available cipher suite. Do not enable ciphers and algorithms that you consider weak.

All the supported ciphersuites are listed in ciphersuites.

3.5.2 Authentication

The key exchange algorithms of the TLS protocol offer authentication, which is a prerequisite for a secure connection. The available authentication methods in GnuTLS follow.

• Certificate authentication: Authenticated key exchange using public key infrastructure and certificates (X.509 or OpenPGP).
• SRP authentication: Authenticated key exchange using a password.
• PSK authentication: Authenticated key exchange using a pre-shared key.
• Anonymous authentication: Key exchange without peer authentication.

3.5.3 Client authentication

In the case of ciphersuites that use certificate authentication, the authentication of the client is optional in TLS. A server may request a certificate from the client using the gnutls_certificate_server_set_request function. We elaborate in Certificate credentials.

3.5.4 Resuming sessions

The TLS handshake process performs expensive calculations and a busy server might easily be put under load. To reduce the load, session resumption may be used. This is a feature of the TLS protocol which allows a client to connect to a server after a successful handshake, without the expensive calculations. This is achieved by re-using the previously established keys, meaning the server needs to store the state of established connections (unless session tickets are used – Session tickets).

Session resumption is an integral part of GnuTLS, and Session resumption, ex-resume-client illustrate typical uses of it.

3.6 TLS extensions

A number of extensions to the TLS protocol have been proposed mainly in [TLSEXT]. The extensions supported in GnuTLS are discussed in the subsections that follow.

3.6.1 Maximum fragment length negotiation

This extension allows a TLS implementation to negotiate a smaller value for record packet maximum length. This extension may be useful to clients with constrained capabilities. The functions shown below can be used to control this extension.

size_t gnutls_record_get_max_size (gnutls_session_t session)

ssize_t gnutls_record_set_max_size (gnutls_session_t session, size_t size)

3.6.2 Server name indication

A common problem in HTTPS servers is the fact that the TLS protocol is not aware of the hostname that a client connects to, when the handshake procedure begins. For that reason the TLS server has no way to know which certificate to send.

This extension solves that problem within the TLS protocol, and allows a client to send the HTTP hostname before the handshake begins within the first handshake packet. The functions gnutls_server_name_set and gnutls_server_name_get can be used to enable this extension, or to retrieve the name sent by a client.

int gnutls_server_name_set (gnutls_session_t session, gnutls_server_name_type_t type, const void * name, size_t name_length)

int gnutls_server_name_get (gnutls_session_t session, void * data, size_t * data_length, unsigned int * type, unsigned int indx)

3.6.3 Session tickets

To resume a TLS session, the server normally stores session parameters. This complicates deployment, and can be avoided by delegating the storage to the client. Because session parameters are sensitive they are encrypted and authenticated with a key only known to the server and then sent to the client. The Session Tickets extension is described in RFC 5077 [TLSTKT].

A disadvantage of session tickets is that they eliminate the effects of forward secrecy when a server uses the same key for long time. That is, the secrecy of all sessions on a server using tickets depends on the ticket key being kept secret. For that reason server keys should be rotated and discarded regularly.

Since version 3.1.3 GnuTLS clients transparently support session tickets, unless forward secrecy is explicitly requested (with the PFS priority string).

3.6.4 HeartBeat

This is a TLS extension that allows to ping and receive confirmation from the peer, and is described in [RFC6520]. The extension is disabled by default and gnutls_heartbeat_enable can be used to enable it. A policy may be negotiated to only allow sending heartbeat messages or sending and receiving. The current session policy can be checked with gnutls_heartbeat_allowed. The requests coming from the peer result to GNUTLS_E_HEARTBEAT_PING_RECEIVED being returned from the receive function. Ping requests to peer can be send via gnutls_heartbeat_ping.

int gnutls_heartbeat_allowed (gnutls_session_t session, unsigned int type)

void gnutls_heartbeat_enable (gnutls_session_t session, unsigned int type)

int gnutls_heartbeat_ping (gnutls_session_t session, size_t data_size, unsigned int max_tries, unsigned int flags)

int gnutls_heartbeat_pong (gnutls_session_t session, unsigned int flags)

void gnutls_heartbeat_set_timeouts (gnutls_session_t session, unsigned int retrans_timeout, unsigned int total_timeout)

unsigned int gnutls_heartbeat_get_timeout (gnutls_session_t session)

3.6.5 Safe renegotiation

TLS gives the option to two communicating parties to renegotiate and update their security parameters. One useful example of this feature was for a client to initially connect using anonymous negotiation to a server, and the renegotiate using some authenticated ciphersuite. This occurred to avoid having the client sending its credentials in the clear.

However this renegotiation, as initially designed would not ensure that the party one is renegotiating is the same as the one in the initial negotiation. For example one server could forward all renegotiation traffic to an other server who will see this traffic as an initial negotiation attempt.

This might be seen as a valid design decision, but it seems it was not widely known or understood, thus today some application protocols use the TLS renegotiation feature in a manner that enables a malicious server to insert content of his choice in the beginning of a TLS session.

The most prominent vulnerability was with HTTPS. There servers request a renegotiation to enforce an anonymous user to use a certificate in order to access certain parts of a web site. The attack works by having the attacker simulate a client and connect to a server, with server-only authentication, and send some data intended to cause harm. The server will then require renegotiation from him in order to perform the request. When the proper client attempts to contact the server, the attacker hijacks that connection and forwards traffic to the initial server that requested renegotiation. The attacker will not be able to read the data exchanged between the client and the server. However, the server will (incorrectly) assume that the initial request sent by the attacker was sent by the now authenticated client. The result is a prefix plain-text injection attack.

The above is just one example. Other vulnerabilities exists that do not rely on the TLS renegotiation to change the client’s authenticated status (either TLS or application layer).

While fixing these application protocols and implementations would be one natural reaction, an extension to TLS has been designed that cryptographically binds together any renegotiated handshakes with the initial negotiation. When the extension is used, the attack is detected and the session can be terminated. The extension is specified in [RFC5746].

GnuTLS supports the safe renegotiation extension. The default behavior is as follows. Clients will attempt to negotiate the safe renegotiation extension when talking to servers. Servers will accept the extension when presented by clients. Clients and servers will permit an initial handshake to complete even when the other side does not support the safe renegotiation extension. Clients and servers will refuse renegotiation attempts when the extension has not been negotiated.

Note that permitting clients to connect to servers when the safe renegotiation extension is not enabled, is open up for attacks. Changing this default behavior would prevent interoperability against the majority of deployed servers out there. We will reconsider this default behavior in the future when more servers have been upgraded. Note that it is easy to configure clients to always require the safe renegotiation extension from servers.

To modify the default behavior, we have introduced some new priority strings (see Priority Strings). The %UNSAFE_RENEGOTIATION priority string permits (re-)handshakes even when the safe renegotiation extension was not negotiated. The default behavior is %PARTIAL_RENEGOTIATION that will prevent renegotiation with clients and servers not supporting the extension. This is secure for servers but leaves clients vulnerable to some attacks, but this is a trade-off between security and compatibility with old servers. The %SAFE_RENEGOTIATION priority string makes clients and servers require the extension for every handshake. The latter is the most secure option for clients, at the cost of not being able to connect to legacy servers. Servers will also deny clients that do not support the extension from connecting.

It is possible to disable use of the extension completely, in both clients and servers, by using the %DISABLE_SAFE_RENEGOTIATION priority string however we strongly recommend you to only do this for debugging and test purposes.

The default values if the flags above are not specified are:

Server:

%PARTIAL_RENEGOTIATION

Client:

%PARTIAL_RENEGOTIATION

For applications we have introduced a new API related to safe renegotiation. The gnutls_safe_renegotiation_status function is used to check if the extension has been negotiated on a session, and can be used both by clients and servers.

3.6.6 OCSP status request

The Online Certificate Status Protocol (OCSP) is a protocol that allows the client to verify the server certificate for revocation without messing with certificate revocation lists. Its drawback is that it requires the client to connect to the server’s CA OCSP server and request the status of the certificate. This extension however, enables a TLS server to include its CA OCSP server response in the handshake. That is an HTTPS server may periodically run ocsptool (see ocsptool Invocation) to obtain its certificate revocation status and serve it to the clients. That way a client avoids an additional connection to the OCSP server.

See OCSP stapling for further information.

Since version 3.1.3 GnuTLS clients transparently support the certificate status request.

3.6.7 SRTP

The TLS protocol was extended in [RFC5764] to provide keying material to the Secure RTP (SRTP) protocol. The SRTP protocol provides an encapsulation of encrypted data that is optimized for voice data. With the SRTP TLS extension two peers can negotiate keys using TLS or DTLS and obtain keying material for use with SRTP. The available SRTP profiles are listed below.

To enable use the following functions.

int gnutls_srtp_set_profile (gnutls_session_t session, gnutls_srtp_profile_t profile)

int gnutls_srtp_set_profile_direct (gnutls_session_t session, const char * profiles, const char ** err_pos)

To obtain the negotiated keys use the function below.

Function: int gnutls_srtp_get_keys (gnutls_session_t session, void * key_material, unsigned int key_material_size, gnutls_datum_t * client_key, gnutls_datum_t * client_salt, gnutls_datum_t * server_key, gnutls_datum_t * server_salt)

session: is a gnutls_session_t type.

key_material: Space to hold the generated key material

key_material_size: The maximum size of the key material

client_key: The master client write key, pointing inside the key material

client_salt: The master client write salt, pointing inside the key material

server_key: The master server write key, pointing inside the key material

server_salt: The master server write salt, pointing inside the key material

This is a helper function to generate the keying material for SRTP. It requires the space of the key material to be pre-allocated (should be at least 2x the maximum key size and salt size). The client_key , client_salt , server_key and server_salt are convenience datums that point inside the key material. They may be NULL .

Returns: On success the size of the key material is returned, otherwise, GNUTLS_E_SHORT_MEMORY_BUFFER if the buffer given is not sufficient, or a negative error code.

Since 3.1.4

Other helper functions are listed below.

int gnutls_srtp_get_selected_profile (gnutls_session_t session, gnutls_srtp_profile_t * profile)

const char * gnutls_srtp_get_profile_name (gnutls_srtp_profile_t profile)

int gnutls_srtp_get_profile_id (const char * name, gnutls_srtp_profile_t * profile)

3.6.8 False Start

The TLS protocol was extended in [draft-ietf-tls-falsestart-01] to allow the client to send data to server in a single round trip. This change however operates on the borderline of the TLS protocol security guarrantees and should be used for the cases where the reduced latency outperforms the risk of an adversary intercepting the transferred data. In GnuTLS applications can use the GNUTLS_ENABLE_FALSE_START as option to gnutls_init to request an early return of the gnutls_handshake function. After that early return the application is expected to transfer any data to be piggybacked on the last handshake message.

After handshake’s early termination, the application is expected to transmit data using gnutls_record_send, and call gnutls_record_recv on any received data as soon, to ensure that handshake completes timely. That is, especially relevant for applications which set an explicit time limit for the handshake process via gnutls_handshake_set_timeout.

Note however, that the API ensures that the early return will not happen if the false start requirements are not satisfied. That is, on ciphersuites which are not whitelisted for false start or on insufficient key sizes, the handshake process will complete properly (i.e., no early return). To verify that false start was used you may use gnutls_session_get_flags and check for the GNUTLS_SFLAGS_FALSE_START flag. For GnuTLS the false start is whitelisted for the following key exchange methods (see [draft-ietf-tls-falsestart-01] for rationale)

• DHE
• ECDHE

but only when the negotiated parameters exceed GNUTLS_SEC_PARAM_HIGH –see Table 6.7, and when under (D)TLS 1.2 or later.

3.6.9 Application Layer Protocol Negotiation (ALPN)

The TLS protocol was extended in RFC7301 to provide the application layer a method of negotiating the application protocol version. This allows for negotiation of the application protocol during the TLS handshake, thus reducing round-trips. The application protocol is described by an opaque string. To enable, use the following functions.

int gnutls_alpn_set_protocols (gnutls_session_t session, const gnutls_datum_t * protocols, unsigned protocols_size, unsigned int flags)

int gnutls_alpn_get_selected_protocol (gnutls_session_t session, gnutls_datum_t * protocol)

Note that these functions are intended to be used with protocols that are registered in the Application Layer Protocol Negotiation IANA registry. While you can use them for other protocols (at the risk of collisions), it is preferable to register them.

3.6.10 Extensions and Supplemental Data

It is possible to transfer supplemental data during the TLS handshake, following [RFC4680]. This is for "custom" protocol modifications for applications which may want to transfer additional data (e.g. additional authentication messages). Such an exchange requires a custom extension to be registered. The provided API for this functionality is low-level and described in TLS Extension Handling.

3.7 How to use TLS in application protocols

This chapter is intended to provide some hints on how to use TLS over simple custom made application protocols. The discussion below mainly refers to the TCP/IP transport layer but may be extended to other ones too.

3.7.1 Separate ports

Traditionally SSL was used in application protocols by assigning a new port number for the secure services. By doing this two separate ports were assigned, one for the non-secure sessions, and one for the secure sessions. This method ensures that if a user requests a secure session then the client will attempt to connect to the secure port and fail otherwise. The only possible attack with this method is to perform a denial of service attack. The most famous example of this method is “HTTP over TLS” or HTTPS protocol [RFC2818].

Despite its wide use, this method has several issues. This approach starts the TLS Handshake procedure just after the client connects on the —so called— secure port. That way the TLS protocol does not know anything about the client, and popular methods like the host advertising in HTTP do not work⁶. There is no way for the client to say “I connected to YYY server” before the Handshake starts, so the server cannot possibly know which certificate to use.

Other than that it requires two separate ports to run a single service, which is unnecessary complication. Due to the fact that there is a limitation on the available privileged ports, this approach was soon deprecated in favor of upward negotiation.

3.7.2 Upward negotiation

Other application protocols⁷ use a different approach to enable the secure layer. They use something often called as the “TLS upgrade” method. This method is quite tricky but it is more flexible. The idea is to extend the application protocol to have a “STARTTLS” request, whose purpose it to start the TLS protocols just after the client requests it. This approach does not require any extra port to be reserved. There is even an extension to HTTP protocol to support this method [RFC2817].

The tricky part, in this method, is that the “STARTTLS” request is sent in the clear, thus is vulnerable to modifications. A typical attack is to modify the messages in a way that the client is fooled and thinks that the server does not have the “STARTTLS” capability. See a typical conversation of a hypothetical protocol:

(client connects to the server)

CLIENT: HELLO I’M MR. XXX

SERVER: NICE TO MEET YOU XXX

CLIENT: PLEASE START TLS

SERVER: OK

*** TLS STARTS

CLIENT: HERE ARE SOME CONFIDENTIAL DATA

And an example of a conversation where someone is acting in between:

(client connects to the server)

CLIENT: HELLO I’M MR. XXX

SERVER: NICE TO MEET YOU XXX

CLIENT: PLEASE START TLS

(here someone inserts this message)

SERVER: SORRY I DON’T HAVE THIS CAPABILITY

CLIENT: HERE ARE SOME CONFIDENTIAL DATA

As you can see above the client was fooled, and was naïve enough to send the confidential data in the clear, despite the server telling the client that it does not support “STARTTLS”.

How do we avoid the above attack? As you may have already noticed this situation is easy to avoid. The client has to ask the user before it connects whether the user requests TLS or not. If the user answered that he certainly wants the secure layer the last conversation should be:

(client connects to the server)

CLIENT: HELLO I’M MR. XXX

SERVER: NICE TO MEET YOU XXX

CLIENT: PLEASE START TLS

(here someone inserts this message)

SERVER: SORRY I DON’T HAVE THIS CAPABILITY

CLIENT: BYE

(the client notifies the user that the secure connection was not possible)

This method, if implemented properly, is far better than the traditional method, and the security properties remain the same, since only denial of service is possible. The benefit is that the server may request additional data before the TLS Handshake protocol starts, in order to send the correct certificate, use the correct password file, or anything else!

3.8 On SSL 2 and older protocols

One of the initial decisions in the GnuTLS development was to implement the known security protocols for the transport layer. Initially TLS 1.0 was implemented since it was the latest at that time, and was considered to be the most advanced in security properties. Later the SSL 3.0 protocol was implemented since it is still the only protocol supported by several servers and there are no serious security vulnerabilities known.

One question that may arise is why we didn’t implement SSL 2.0 in the library. There are several reasons, most important being that it has serious security flaws, unacceptable for a modern security library. Other than that, this protocol is barely used by anyone these days since it has been deprecated since 1996. The security problems in SSL 2.0 include:

• Message integrity compromised. The SSLv2 message authentication uses the MD5 function, and is insecure.
• Man-in-the-middle attack. There is no protection of the handshake in SSLv2, which permits a man-in-the-middle attack.
• Truncation attack. SSLv2 relies on TCP FIN to close the session, so the attacker can forge a TCP FIN, and the peer cannot tell if it was a legitimate end of data or not.
• Weak message integrity for export ciphers. The cryptographic keys in SSLv2 are used for both message authentication and encryption, so if weak encryption schemes are negotiated (say 40-bit keys) the message authentication code uses the same weak key, which isn’t necessary.

Other protocols such as Microsoft’s PCT 1 and PCT 2 were not implemented because they were also abandoned and deprecated by SSL 3.0 and later TLS 1.0.

4 Authentication methods

The initial key exchange of the TLS protocol performs authentication of the peers. In typical scenarios the server is authenticated to the client, and optionally the client to the server.

While many associate TLS with X.509 certificates and public key authentication, the protocol supports various authentication methods, including pre-shared keys, and passwords. In this chapter a description of the existing authentication methods is provided, as well as some guidance on which use-cases each method can be used at.

4.1 Certificate authentication

The most known authentication method of TLS are certificates. The PKIX [PKIX] public key infrastructure is daily used by anyone using a browser today. GnuTLS provides a simple API to use the X.509 certificates [PKIX].

The key exchange algorithms supported by certificate authentication are shown in Table 4.1.

4.1.1 X.509 certificates

The X.509 protocols rely on a hierarchical trust model. In this trust model Certification Authorities (CAs) are used to certify entities. Usually more than one certification authorities exist, and certification authorities may certify other authorities to issue certificates as well, following a hierarchical model.

Figure 4.1: An example of the X.509 hierarchical trust model.

One needs to trust one or more CAs for his secure communications. In that case only the certificates issued by the trusted authorities are acceptable. The framework is illustrated on Figure 4.1.

4.1.1.1 X.509 certificate structure

An X.509 certificate usually contains information about the certificate holder, the signer, a unique serial number, expiration dates and some other fields [PKIX] as shown in Table 4.2.

The certificate’s subject or issuer name is not just a single string. It is a Distinguished name and in the ASN.1 notation is a sequence of several object identifiers with their corresponding values. Some of available OIDs to be used in an X.509 distinguished name are defined in gnutls/x509.h.

The Version field in a certificate has values either 1 or 3 for version 3 certificates. Version 1 certificates do not support the extensions field so it is not possible to distinguish a CA from a person, thus their usage should be avoided.

The validity dates are there to indicate the date that the specific certificate was activated and the date the certificate’s key would be considered invalid.

In GnuTLS the X.509 certificate structures are handled using the gnutls_x509_crt_t type and the corresponding private keys with the gnutls_x509_privkey_t type. All the available functions for X.509 certificate handling have their prototypes in gnutls/x509.h. An example program to demonstrate the X.509 parsing capabilities can be found in ex-x509-info.

4.1.1.2 Importing an X.509 certificate

The certificate structure should be initialized using gnutls_x509_crt_init, and a certificate structure can be imported using gnutls_x509_crt_import.

int gnutls_x509_crt_init (gnutls_x509_crt_t * cert)

int gnutls_x509_crt_import (gnutls_x509_crt_t cert, const gnutls_datum_t * data, gnutls_x509_crt_fmt_t format)

void gnutls_x509_crt_deinit (gnutls_x509_crt_t cert)

In several functions an array of certificates is required. To assist in initialization and import the following two functions are provided.

int gnutls_x509_crt_list_import (gnutls_x509_crt_t * certs, unsigned int * cert_max, const gnutls_datum_t * data, gnutls_x509_crt_fmt_t format, unsigned int flags)

int gnutls_x509_crt_list_import2 (gnutls_x509_crt_t ** certs, unsigned int * size, const gnutls_datum_t * data, gnutls_x509_crt_fmt_t format, unsigned int flags)

In all cases after use a certificate must be deinitialized using gnutls_x509_crt_deinit. Note that although the functions above apply to gnutls_x509_crt_t structure, similar functions exist for the CRL structure gnutls_x509_crl_t.

4.1.1.3 X.509 certificate names

X.509 certificates allow for multiple names and types of names to be specified. CA certificates often rely on X.509 distinguished names (see X.509 distinguished names) for unique identification, while end-user and server certificates rely on the ’subject alternative names’. The subject alternative names provide a typed name, e.g., a DNS name, or an email address, which identifies the owner of the certificate. The following functions provide access to that names.

int gnutls_x509_crt_get_subject_alt_name2 (gnutls_x509_crt_t cert, unsigned int seq, void * san, size_t * san_size, unsigned int * san_type, unsigned int * critical)

int gnutls_x509_crt_set_subject_alt_name (gnutls_x509_crt_t crt, gnutls_x509_subject_alt_name_t type, const void * data, unsigned int data_size, unsigned int flags)

int gnutls_subject_alt_names_init (gnutls_subject_alt_names_t * sans)

int gnutls_subject_alt_names_get (gnutls_subject_alt_names_t sans, unsigned int seq, unsigned int * san_type, gnutls_datum_t * san, gnutls_datum_t * othername_oid)

int gnutls_subject_alt_names_set (gnutls_subject_alt_names_t sans, unsigned int san_type, const gnutls_datum_t * san, const char * othername_oid)

Note however, that server certificates often used the Common Name (CN), part of the certificate DistinguishedName to place a single DNS address. That practice is discouraged (see [RFC6125]), because only a single address can be specified, and the CN field is free-form making matching ambiguous.

4.1.1.4 X.509 distinguished names

The “subject” of an X.509 certificate is not described by a single name, but rather with a distinguished name. This in X.509 terminology is a list of strings each associated an object identifier. To make things simple GnuTLS provides gnutls_x509_crt_get_dn2 which follows the rules in [RFC4514] and returns a single string. Access to each string by individual object identifiers can be accessed using gnutls_x509_crt_get_dn_by_oid.

Function: int gnutls_x509_crt_get_dn2 (gnutls_x509_crt_t cert, gnutls_datum_t * dn)

cert: should contain a gnutls_x509_crt_t type

dn: a pointer to a structure to hold the name

This function will allocate buffer and copy the name of the Certificate. The name will be in the form "C=xxxx,O=yyyy,CN=zzzz" as described in RFC4514. The output string will be ASCII or UTF-8 encoded, depending on the certificate data.

This function does not output a fully RFC4514 compliant string, if that is required see gnutls_x509_crt_get_dn3() .

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since: 3.1.10

int gnutls_x509_crt_get_dn (gnutls_x509_crt_t cert, char * buf, size_t * buf_size)

int gnutls_x509_crt_get_dn_by_oid (gnutls_x509_crt_t cert, const char * oid, unsigned indx, unsigned int raw_flag, void * buf, size_t * buf_size)

int gnutls_x509_crt_get_dn_oid (gnutls_x509_crt_t cert, unsigned indx, void * oid, size_t * oid_size)

Similar functions exist to access the distinguished name of the issuer of the certificate.

int gnutls_x509_crt_get_issuer_dn (gnutls_x509_crt_t cert, char * buf, size_t * buf_size)

int gnutls_x509_crt_get_issuer_dn2 (gnutls_x509_crt_t cert, gnutls_datum_t * dn)

int gnutls_x509_crt_get_issuer_dn_by_oid (gnutls_x509_crt_t cert, const char * oid, unsigned indx, unsigned int raw_flag, void * buf, size_t * buf_size)

int gnutls_x509_crt_get_issuer_dn_oid (gnutls_x509_crt_t cert, unsigned indx, void * oid, size_t * oid_size)

int gnutls_x509_crt_get_issuer (gnutls_x509_crt_t cert, gnutls_x509_dn_t * dn)

The more powerful gnutls_x509_crt_get_subject and gnutls_x509_dn_get_rdn_ava provide efficient but low-level access to the contents of the distinguished name structure.

int gnutls_x509_crt_get_subject (gnutls_x509_crt_t cert, gnutls_x509_dn_t * dn)

int gnutls_x509_crt_get_issuer (gnutls_x509_crt_t cert, gnutls_x509_dn_t * dn)

Function: int gnutls_x509_dn_get_rdn_ava (gnutls_x509_dn_t dn, int irdn, int iava, gnutls_x509_ava_st * ava)

dn: a pointer to DN

irdn: index of RDN

iava: index of AVA.

ava: Pointer to structure which will hold output information.

Get pointers to data within the DN. The format of the ava structure is shown below.

struct gnutls_x509_ava_st { gnutls_datum_t oid; gnutls_datum_t value; unsigned long value_tag; };

The X.509 distinguished name is a sequence of sequences of strings and this is what the irdn and iava indexes model.

Note that ava will contain pointers into the dn structure which in turns points to the original certificate. Thus you should not modify any data or deallocate any of those.

This is a low-level function that requires the caller to do the value conversions when necessary (e.g. from UCS-2).

Returns: Returns 0 on success, or an error code.

4.1.1.5 X.509 extensions

X.509 version 3 certificates include a list of extensions that can be used to obtain additional information on the subject or the issuer of the certificate. Those may be e-mail addresses, flags that indicate whether the belongs to a CA etc. All the supported X.509 version 3 extensions are shown in Table 4.3.

The certificate extensions access is split into two parts. The first requires to retrieve the extension, and the second is the parsing part.

To enumerate and retrieve the DER-encoded extension data available in a certificate the following two functions are available.

int gnutls_x509_crt_get_extension_info (gnutls_x509_crt_t cert, unsigned indx, void * oid, size_t * oid_size, unsigned int * critical)

int gnutls_x509_crt_get_extension_data2 (gnutls_x509_crt_t cert, unsigned indx, gnutls_datum_t * data)

int gnutls_x509_crt_get_extension_by_oid2 (gnutls_x509_crt_t cert, const char * oid, unsigned indx, gnutls_datum_t * output, unsigned int * critical)

After a supported DER-encoded extension is retrieved it can be parsed using the APIs in x509-ext.h. Complex extensions may require initializing an intermediate structure that holds the parsed extension data. Examples of simple parsing functions are shown below.

int gnutls_x509_ext_import_basic_constraints (const gnutls_datum_t * ext, unsigned int * ca, int * pathlen)

int gnutls_x509_ext_export_basic_constraints (unsigned int ca, int pathlen, gnutls_datum_t * ext)

int gnutls_x509_ext_import_key_usage (const gnutls_datum_t * ext, unsigned int * key_usage)

int gnutls_x509_ext_export_key_usage (unsigned int usage, gnutls_datum_t * ext)

More complex extensions, such as Name Constraints, require an intermediate structure, in that case gnutls_x509_name_constraints_t to be initialized in order to store the parsed extension data.

int gnutls_x509_ext_import_name_constraints (const gnutls_datum_t * ext, gnutls_x509_name_constraints_t nc, unsigned int flags)

int gnutls_x509_ext_export_name_constraints (gnutls_x509_name_constraints_t nc, gnutls_datum_t * ext)

After the name constraints are extracted in the structure, the following functions can be used to access them.

int gnutls_x509_name_constraints_get_permitted (gnutls_x509_name_constraints_t nc, unsigned idx, unsigned * type, gnutls_datum_t * name)

int gnutls_x509_name_constraints_get_excluded (gnutls_x509_name_constraints_t nc, unsigned idx, unsigned * type, gnutls_datum_t * name)

int gnutls_x509_name_constraints_add_permitted (gnutls_x509_name_constraints_t nc, gnutls_x509_subject_alt_name_t type, const gnutls_datum_t * name)

int gnutls_x509_name_constraints_add_excluded (gnutls_x509_name_constraints_t nc, gnutls_x509_subject_alt_name_t type, const gnutls_datum_t * name)

unsigned gnutls_x509_name_constraints_check (gnutls_x509_name_constraints_t nc, gnutls_x509_subject_alt_name_t type, const gnutls_datum_t * name)

unsigned gnutls_x509_name_constraints_check_crt (gnutls_x509_name_constraints_t nc, gnutls_x509_subject_alt_name_t type, gnutls_x509_crt_t cert)

Other utility functions are listed below.

int gnutls_x509_name_constraints_init (gnutls_x509_name_constraints_t * nc)

void gnutls_x509_name_constraints_deinit (gnutls_x509_name_constraints_t nc)

Similar functions exist for all of the other supported extensions, listed in Table 4.3.

Note, that there are also direct APIs to access extensions that may be simpler to use for non-complex extensions. They are available in x509.h and some examples are listed below.

int gnutls_x509_crt_get_basic_constraints (gnutls_x509_crt_t cert, unsigned int * critical, unsigned int * ca, int * pathlen)

int gnutls_x509_crt_set_basic_constraints (gnutls_x509_crt_t crt, unsigned int ca, int pathLenConstraint)

int gnutls_x509_crt_get_key_usage (gnutls_x509_crt_t cert, unsigned int * key_usage, unsigned int * critical)

int gnutls_x509_crt_set_key_usage (gnutls_x509_crt_t crt, unsigned int usage)

4.1.1.6 Accessing public and private keys

Each X.509 certificate contains a public key that corresponds to a private key. To get a unique identifier of the public key the gnutls_x509_crt_get_key_id function is provided. To export the public key or its parameters you may need to convert the X.509 structure to a gnutls_pubkey_t. See Abstract public keys for more information.

Function: int gnutls_x509_crt_get_key_id (gnutls_x509_crt_t crt, unsigned int flags, unsigned char * output_data, size_t * output_data_size)

crt: Holds the certificate

flags: should be one of the flags from gnutls_keyid_flags_t

output_data: will contain the key ID

output_data_size: holds the size of output_data (and will be replaced by the actual size of parameters)

This function will return a unique ID that depends on the public key parameters. This ID can be used in checking whether a certificate corresponds to the given private key.

If the buffer provided is not long enough to hold the output, then *output_data_size is updated and GNUTLS_E_SHORT_MEMORY_BUFFER will be returned. The output will normally be a SHA-1 hash output, which is 20 bytes.

Returns: In case of failure a negative error code will be returned, and 0 on success.

The private key parameters may be directly accessed by using one of the following functions.

int gnutls_x509_privkey_get_pk_algorithm2 (gnutls_x509_privkey_t key, unsigned int * bits)

int gnutls_x509_privkey_export_rsa_raw2 (gnutls_x509_privkey_t key, gnutls_datum_t * m, gnutls_datum_t * e, gnutls_datum_t * d, gnutls_datum_t * p, gnutls_datum_t * q, gnutls_datum_t * u, gnutls_datum_t * e1, gnutls_datum_t * e2)

int gnutls_x509_privkey_export_ecc_raw (gnutls_x509_privkey_t key, gnutls_ecc_curve_t * curve, gnutls_datum_t * x, gnutls_datum_t * y, gnutls_datum_t * k)

int gnutls_x509_privkey_export_dsa_raw (gnutls_x509_privkey_t key, gnutls_datum_t * p, gnutls_datum_t * q, gnutls_datum_t * g, gnutls_datum_t * y, gnutls_datum_t * x)

int gnutls_x509_privkey_get_key_id (gnutls_x509_privkey_t key, unsigned int flags, unsigned char * output_data, size_t * output_data_size)

4.1.1.7 Verifying X.509 certificate paths

Verifying certificate paths is important in X.509 authentication. For this purpose the following functions are provided.

Function: int gnutls_x509_trust_list_add_cas (gnutls_x509_trust_list_t list, const gnutls_x509_crt_t * clist, unsigned clist_size, unsigned int flags)

list: The list

clist: A list of CAs

clist_size: The length of the CA list

flags: flags from gnutls_trust_list_flags_t

This function will add the given certificate authorities to the trusted list. The list of CAs must not be deinitialized during this structure’s lifetime.

If the flag GNUTLS_TL_NO_DUPLICATES is specified, then the provided clist entries that are duplicates will not be added to the list and will be deinitialized.

Returns: The number of added elements is returned.

Since: 3.0.0

Function: int gnutls_x509_trust_list_add_named_crt (gnutls_x509_trust_list_t list, gnutls_x509_crt_t cert, const void * name, size_t name_size, unsigned int flags)

list: The list

cert: A certificate

name: An identifier for the certificate

name_size: The size of the identifier

flags: should be 0.

This function will add the given certificate to the trusted list and associate it with a name. The certificate will not be be used for verification with gnutls_x509_trust_list_verify_crt() but with gnutls_x509_trust_list_verify_named_crt() or gnutls_x509_trust_list_verify_crt2() - the latter only since GnuTLS 3.4.0 and if a hostname is provided.

In principle this function can be used to set individual "server" certificates that are trusted by the user for that specific server but for no other purposes.

The certificate must not be deinitialized during the lifetime of the trusted list.

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since: 3.0.0

Function: int gnutls_x509_trust_list_add_crls (gnutls_x509_trust_list_t list, const gnutls_x509_crl_t * crl_list, unsigned crl_size, unsigned int flags, unsigned int verification_flags)

list: The list

crl_list: A list of CRLs

crl_size: The length of the CRL list

flags: flags from gnutls_trust_list_flags_t

verification_flags: gnutls_certificate_verify_flags if flags specifies GNUTLS_TL_VERIFY_CRL

This function will add the given certificate revocation lists to the trusted list. The list of CRLs must not be deinitialized during this structure’s lifetime.

This function must be called after gnutls_x509_trust_list_add_cas() to allow verifying the CRLs for validity. If the flag GNUTLS_TL_NO_DUPLICATES is given, then any provided CRLs that are a duplicate, will be deinitialized and not added to the list (that assumes that gnutls_x509_trust_list_deinit() will be called with all=1).

If GNUTLS_TL_VERIFY_CRL is given the CRLs will be verified before being added.

Returns: The number of added elements is returned.

Since: 3.0

Function: int gnutls_x509_trust_list_verify_crt (gnutls_x509_trust_list_t list, gnutls_x509_crt_t * cert_list, unsigned int cert_list_size, unsigned int flags, unsigned int * voutput, gnutls_verify_output_function func)

list: The list

cert_list: is the certificate list to be verified

cert_list_size: is the certificate list size

flags: Flags that may be used to change the verification algorithm. Use OR of the gnutls_certificate_verify_flags enumerations.

voutput: will hold the certificate verification output.

func: If non-null will be called on each chain element verification with the output.

This function will try to verify the given certificate and return its status. The voutput parameter will hold an OR’ed sequence of gnutls_certificate_status_t flags.

The details of the verification are the same as in gnutls_x509_trust_list_verify_crt2() .

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since: 3.0

Function: int gnutls_x509_trust_list_verify_crt2 (gnutls_x509_trust_list_t list, gnutls_x509_crt_t * cert_list, unsigned int cert_list_size, gnutls_typed_vdata_st * data, unsigned int elements, unsigned int flags, unsigned int * voutput, gnutls_verify_output_function func)

list: The list

cert_list: is the certificate list to be verified

cert_list_size: is the certificate list size

data: an array of typed data

elements: the number of data elements

flags: Flags that may be used to change the verification algorithm. Use OR of the gnutls_certificate_verify_flags enumerations.

voutput: will hold the certificate verification output.

func: If non-null will be called on each chain element verification with the output.

This function will attempt to verify the given certificate chain and return its status. The voutput parameter will hold an OR’ed sequence of gnutls_certificate_status_t flags.

When a certificate chain of cert_list_size with more than one certificates is provided, the verification status will apply to the first certificate in the chain that failed verification. The verification process starts from the end of the chain (from CA to end certificate). The first certificate in the chain must be the end-certificate while the rest of the members may be sorted or not.

Additionally a certificate verification profile can be specified from the ones in gnutls_certificate_verification_profiles_t by ORing the result of GNUTLS_PROFILE_TO_VFLAGS() to the verification flags.

Additional verification parameters are possible via the data types; the acceptable types are GNUTLS_DT_DNS_HOSTNAME and GNUTLS_DT_KEY_PURPOSE_OID . The former accepts as data a null-terminated hostname, and the latter a null-terminated object identifier (e.g., GNUTLS_KP_TLS_WWW_SERVER ). If a DNS hostname is provided then this function will compare the hostname in the end certificate against the given. If names do not match the GNUTLS_CERT_UNEXPECTED_OWNER status flag will be set. In addition it will consider certificates provided with gnutls_x509_trust_list_add_named_crt() .

If a key purpose OID is provided and the end-certificate contains the extended key usage PKIX extension, it will be required to match the provided OID or be marked for any purpose, otherwise verification will fail with GNUTLS_CERT_PURPOSE_MISMATCH status.

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value. Note that verification failure will not result to an error code, only voutput will be updated.

Since: 3.3.8

Function: int gnutls_x509_trust_list_verify_named_crt (gnutls_x509_trust_list_t list, gnutls_x509_crt_t cert, const void * name, size_t name_size, unsigned int flags, unsigned int * voutput, gnutls_verify_output_function func)

list: The list

cert: is the certificate to be verified

name: is the certificate’s name

name_size: is the certificate’s name size

flags: Flags that may be used to change the verification algorithm. Use OR of the gnutls_certificate_verify_flags enumerations.

voutput: will hold the certificate verification output.

func: If non-null will be called on each chain element verification with the output.

This function will try to find a certificate that is associated with the provided name –see gnutls_x509_trust_list_add_named_crt() . If a match is found the certificate is considered valid. In addition to that this function will also check CRLs. The voutput parameter will hold an OR’ed sequence of gnutls_certificate_status_t flags.

Additionally a certificate verification profile can be specified from the ones in gnutls_certificate_verification_profiles_t by ORing the result of GNUTLS_PROFILE_TO_VFLAGS() to the verification flags.

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since: 3.0.0

Function: int gnutls_x509_trust_list_add_trust_file (gnutls_x509_trust_list_t list, const char * ca_file, const char * crl_file, gnutls_x509_crt_fmt_t type, unsigned int tl_flags, unsigned int tl_vflags)

list: The list

ca_file: A file containing a list of CAs (optional)

crl_file: A file containing a list of CRLs (optional)

type: The format of the certificates

tl_flags: flags from gnutls_trust_list_flags_t

tl_vflags: gnutls_certificate_verify_flags if flags specifies GNUTLS_TL_VERIFY_CRL

This function will add the given certificate authorities to the trusted list. PKCS 11 URLs are also accepted, instead of files, by this function. A PKCS 11 URL implies a trust database (a specially marked module in p11-kit); the URL "pkcs11:" implies all trust databases in the system. Only a single URL specifying trust databases can be set; they cannot be stacked with multiple calls.

Returns: The number of added elements is returned.

Since: 3.1

Function: int gnutls_x509_trust_list_add_trust_mem (gnutls_x509_trust_list_t list, const gnutls_datum_t * cas, const gnutls_datum_t * crls, gnutls_x509_crt_fmt_t type, unsigned int tl_flags, unsigned int tl_vflags)

list: The list

cas: A buffer containing a list of CAs (optional)

crls: A buffer containing a list of CRLs (optional)

type: The format of the certificates

tl_flags: flags from gnutls_trust_list_flags_t

tl_vflags: gnutls_certificate_verify_flags if flags specifies GNUTLS_TL_VERIFY_CRL

This function will add the given certificate authorities to the trusted list.

Returns: The number of added elements is returned.

Since: 3.1

Function: int gnutls_x509_trust_list_add_system_trust (gnutls_x509_trust_list_t list, unsigned int tl_flags, unsigned int tl_vflags)

list: The structure of the list

tl_flags: GNUTLS_TL_*

tl_vflags: gnutls_certificate_verify_flags if flags specifies GNUTLS_TL_VERIFY_CRL

This function adds the system’s default trusted certificate authorities to the trusted list. Note that on unsupported systems this function returns GNUTLS_E_UNIMPLEMENTED_FEATURE .

This function implies the flag GNUTLS_TL_NO_DUPLICATES .

Returns: The number of added elements or a negative error code on error.

Since: 3.1

The verification function will verify a given certificate chain against a list of certificate authorities and certificate revocation lists, and output a bit-wise OR of elements of the gnutls_certificate_status_t enumeration shown in Figure 4.2. The GNUTLS_CERT_INVALID flag is always set on a verification error and more detailed flags will also be set when appropriate.

An example of certificate verification is shown in ex-verify2. It is also possible to have a set of certificates that are trusted for a particular server but not to authorize other certificates. This purpose is served by the functions gnutls_x509_trust_list_add_named_crt and gnutls_x509_trust_list_verify_named_crt.

4.1.1.8 Verifying a certificate in the context of TLS session

When operating in the context of a TLS session, the trusted certificate authority list may also be set using:

int gnutls_certificate_set_x509_trust_file (gnutls_certificate_credentials_t cred, const char * cafile, gnutls_x509_crt_fmt_t type)

int gnutls_certificate_set_x509_trust_dir (gnutls_certificate_credentials_t cred, const char * ca_dir, gnutls_x509_crt_fmt_t type)

int gnutls_certificate_set_x509_crl_file (gnutls_certificate_credentials_t res, const char * crlfile, gnutls_x509_crt_fmt_t type)

int gnutls_certificate_set_x509_system_trust (gnutls_certificate_credentials_t cred)

These functions allow the specification of the trusted certificate authorities, either via a file, a directory or use the system-specified certificate authorities. Unless the authorities are application specific, it is generally recommended to use the system trust storage (see gnutls_certificate_set_x509_system_trust).

Unlike the previous section it is not required to setup a trusted list, and there are two approaches to verify the peer’s certificate and identity. The recommended in GnuTLS 3.5.0 and later is via the gnutls_session_set_verify_cert, but for older GnuTLS versions you may use an explicit callback set via gnutls_certificate_set_verify_function and then utilize gnutls_certificate_verify_peers3 for verification. The reported verification status is identical to the verification functions described in the previous section.

Note that in certain cases it is required to check the marked purpose of the end certificate (e.g. GNUTLS_KP_TLS_WWW_SERVER); in these cases the more advanced gnutls_session_set_verify_cert2 and gnutls_certificate_verify_peers should be used instead.

There is also the possibility to pass some input to the verification functions in the form of flags. For gnutls_x509_trust_list_verify_crt2 the flags are passed directly, but for gnutls_certificate_verify_peers3, the flags are set using gnutls_certificate_set_verify_flags. All the available flags are part of the enumeration gnutls_certificate_verify_flags shown in Figure 4.3.

4.1.1.9 Verifying a certificate using PKCS #11

Some systems provide a system wide trusted certificate storage accessible using the PKCS #11 API. That is, the trusted certificates are queried and accessed using the PKCS #11 API, and trusted certificate properties, such as purpose, are marked using attached extensions. One example is the p11-kit trust module⁸.

These special PKCS #11 modules can be used for GnuTLS certificate verification if marked as trust policy modules, i.e., with trust-policy: yes in the p11-kit module file. The way to use them is by specifying to the file verification function (e.g., gnutls_certificate_set_x509_trust_file), a pkcs11 URL, or simply pkcs11: to use all the marked with trust policy modules.

The trust modules of p11-kit assign a purpose to trusted authorities using the extended key usage object identifiers. The common purposes are shown in Table 4.4. Note that typically according to [RFC5280] the extended key usage object identifiers apply to end certificates. Their application to CA certificates is an extension used by the trust modules.

With such modules, it is recommended to use the verification functions gnutls_x509_trust_list_verify_crt2, or gnutls_certificate_verify_peers, which allow to explicitly specify the key purpose. The other verification functions which do not allow setting a purpose, would operate as if GNUTLS_KP_TLS_WWW_SERVER was requested from the trusted authorities.

4.1.2 OpenPGP certificates

Previous versions of GnuTLS supported limited OpenPGP key authentication. That functionality has been deprecated and although is still made available, it is no longer recommended to use. The reason is that, supporting alternative authentication methods, when X.509 and PKIX were new and not well established, seemed like a good idea, in today’s internet X.509 is unquestionably the main container for certificates. As such supporting more options with no clear use-cases, is a distraction that consumes considerable resources for improving and testing. For that we have decided to mark this functionality as deprecated, and will be dropped in one of the next GnuTLS major releases.

4.1.3 Advanced certificate verification

The verification of X.509 certificates in the HTTPS and other Internet protocols is typically done by loading a trusted list of commercial Certificate Authorities (see gnutls_certificate_set_x509_system_trust), and using them as trusted anchors. However, there are several examples (eg. the Diginotar incident) where one of these authorities was compromised. This risk can be mitigated by using in addition to CA certificate verification, other verification methods. In this section we list the available in GnuTLS methods.

4.1.3.1 Verifying a certificate using trust on first use authentication

It is possible to use a trust on first use (TOFU) authentication method in GnuTLS. That is the concept used by the SSH programs, where the public key of the peer is not verified, or verified in an out-of-bound way, but subsequent connections to the same peer require the public key to remain the same. Such a system in combination with the typical CA verification of a certificate, and OCSP revocation checks, can help to provide multiple factor verification, where a single point of failure is not enough to compromise the system. For example a server compromise may be detected using OCSP, and a CA compromise can be detected using the trust on first use method. Such a hybrid system with X.509 and trust on first use authentication is shown in Simple client example with SSH-style certificate verification.

See Certificate verification on how to use the available functionality.

4.1.3.2 Verifying a certificate using DANE (DNSSEC)

The DANE protocol is a protocol that can be used to verify TLS certificates using the DNS (or better DNSSEC) protocols. The DNS security extensions (DNSSEC) provide an alternative public key infrastructure to the commercial CAs that are typically used to sign TLS certificates. The DANE protocol takes advantage of the DNSSEC infrastructure to verify TLS certificates. This can be in addition to the verification by CA infrastructure or may even replace it where DNSSEC is fully deployed. Note however, that DNSSEC deployment is fairly new and it would be better to use it as an additional verification method rather than the only one.

The DANE functionality is provided by the libgnutls-dane library that is shipped with GnuTLS and the function prototypes are in gnutls/dane.h. See Certificate verification for information on how to use the library.

Note however, that the DANE RFC mandates the verification methods one should use in addition to the validation via DNSSEC TLSA entries. GnuTLS doesn’t follow that RFC requirement, and the term DANE verification in this manual refers to the TLSA entry verification. In GnuTLS any other verification methods can be used (e.g., PKIX or TOFU) on top of DANE.

4.1.4 Digital signatures

In this section we will provide some information about digital signatures, how they work, and give the rationale for disabling some of the algorithms used.

Digital signatures work by using somebody’s secret key to sign some arbitrary data. Then anybody else could use the public key of that person to verify the signature. Since the data may be arbitrary it is not suitable input to a cryptographic digital signature algorithm. For this reason and also for performance cryptographic hash algorithms are used to preprocess the input to the signature algorithm. This works as long as it is difficult enough to generate two different messages with the same hash algorithm output. In that case the same signature could be used as a proof for both messages. Nobody wants to sign an innocent message of donating 1 euro to Greenpeace and find out that they donated 1.000.000 euros to Bad Inc.

For a hash algorithm to be called cryptographic the following three requirements must hold:

1. Preimage resistance. That means the algorithm must be one way and given the output of the hash function H(x), it is impossible to calculate x.
2. 2nd preimage resistance. That means that given a pair x,y with y=H(x) it is impossible to calculate an x' such that y=H(x').
3. Collision resistance. That means that it is impossible to calculate random x and x' such H(x')=H(x).

The last two requirements in the list are the most important in digital signatures. These protect against somebody who would like to generate two messages with the same hash output. When an algorithm is considered broken usually it means that the Collision resistance of the algorithm is less than brute force. Using the birthday paradox the brute force attack takes 2^{((hash size) / 2)} operations. Today colliding certificates using the MD5 hash algorithm have been generated as shown in [WEGER].

There has been cryptographic results for the SHA-1 hash algorithms as well, although they are not yet critical. Before 2004, MD5 had a presumed collision strength of 2^{64}, but it has been showed to have a collision strength well under 2^{50}. As of November 2005, it is believed that SHA-1’s collision strength is around 2^{63}. We consider this sufficiently hard so that we still support SHA-1. We anticipate that SHA-256/386/512 will be used in publicly-distributed certificates in the future. When 2^{63} can be considered too weak compared to the computer power available sometime in the future, SHA-1 will be disabled as well. The collision attacks on SHA-1 may also get better, given the new interest in tools for creating them.

4.1.4.1 Trading security for interoperability

If you connect to a server and use GnuTLS’ functions to verify the certificate chain, and get a GNUTLS_CERT_INSECURE_ALGORITHM validation error (see Verifying X.509 certificate paths), it means that somewhere in the certificate chain there is a certificate signed using RSA-MD2 or RSA-MD5. These two digital signature algorithms are considered broken, so GnuTLS fails verifying the certificate. In some situations, it may be useful to be able to verify the certificate chain anyway, assuming an attacker did not utilize the fact that these signatures algorithms are broken. This section will give help on how to achieve that.

It is important to know that you do not have to enable any of the flags discussed here to be able to use trusted root CA certificates self-signed using RSA-MD2 or RSA-MD5. The certificates in the trusted list are considered trusted irrespective of the signature.

If you are using gnutls_certificate_verify_peers3 to verify the certificate chain, you can call gnutls_certificate_set_verify_flags with the flags:

• GNUTLS_VERIFY_ALLOW_SIGN_RSA_MD2
• GNUTLS_VERIFY_ALLOW_SIGN_RSA_MD5

as in the following example:

  gnutls_certificate_set_verify_flags (x509cred,
                                       GNUTLS_VERIFY_ALLOW_SIGN_RSA_MD5);

This will signal the verifier algorithm to enable RSA-MD5 when verifying the certificates.

If you are using gnutls_x509_crt_verify or gnutls_x509_crt_list_verify, you can pass the GNUTLS_VERIFY_ALLOW_SIGN_RSA_MD5 parameter directly in the flags parameter.

If you are using these flags, it may also be a good idea to warn the user when verification failure occur for this reason. The simplest is to not use the flags by default, and only fall back to using them after warning the user. If you wish to inspect the certificate chain yourself, you can use gnutls_certificate_get_peers to extract the raw server’s certificate chain, gnutls_x509_crt_list_import to parse each of the certificates, and then gnutls_x509_crt_get_signature_algorithm to find out the signing algorithm used for each certificate. If any of the intermediary certificates are using GNUTLS_SIGN_RSA_MD2 or GNUTLS_SIGN_RSA_MD5, you could present a warning.

4.2 More on certificate authentication

Certificates are not the only structures involved in a public key infrastructure. Several other structures that are used for certificate requests, encrypted private keys, revocation lists, GnuTLS abstract key structures, etc., are discussed in this chapter.

4.2.1 PKCS #10 certificate requests

A certificate request is a structure, which contain information about an applicant of a certificate service. It typically contains a public key, a distinguished name and secondary data such as a challenge password. GnuTLS supports the requests defined in PKCS #10 [RFC2986]. Other formats of certificate requests are not currently supported by GnuTLS.

A certificate request can be generated by associating it with a private key, setting the subject’s information and finally self signing it. The last step ensures that the requester is in possession of the private key.

int gnutls_x509_crq_set_version (gnutls_x509_crq_t crq, unsigned int version)

int gnutls_x509_crq_set_dn (gnutls_x509_crq_t crq, const char * dn, const char ** err)

int gnutls_x509_crq_set_dn_by_oid (gnutls_x509_crq_t crq, const char * oid, unsigned int raw_flag, const void * data, unsigned int sizeof_data)

int gnutls_x509_crq_set_key_usage (gnutls_x509_crq_t crq, unsigned int usage)

int gnutls_x509_crq_set_key_purpose_oid (gnutls_x509_crq_t crq, const void * oid, unsigned int critical)

int gnutls_x509_crq_set_basic_constraints (gnutls_x509_crq_t crq, unsigned int ca, int pathLenConstraint)

The gnutls_x509_crq_set_key and gnutls_x509_crq_sign2 functions associate the request with a private key and sign it. If a request is to be signed with a key residing in a PKCS #11 token it is recommended to use the signing functions shown in Abstract key types.

Function: int gnutls_x509_crq_set_key (gnutls_x509_crq_t crq, gnutls_x509_privkey_t key)

crq: should contain a gnutls_x509_crq_t type

key: holds a private key

This function will set the public parameters from the given private key to the request.

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Function: int gnutls_x509_crq_sign2 (gnutls_x509_crq_t crq, gnutls_x509_privkey_t key, gnutls_digest_algorithm_t dig, unsigned int flags)

crq: should contain a gnutls_x509_crq_t type

key: holds a private key

dig: The message digest to use, i.e., GNUTLS_DIG_SHA1

flags: must be 0

This function will sign the certificate request with a private key. This must be the same key as the one used in gnutls_x509_crt_set_key() since a certificate request is self signed.

This must be the last step in a certificate request generation since all the previously set parameters are now signed.

Returns: GNUTLS_E_SUCCESS on success, otherwise a negative error code. GNUTLS_E_ASN1_VALUE_NOT_FOUND is returned if you didn’t set all information in the certificate request (e.g., the version using gnutls_x509_crq_set_version() ).

The following example is about generating a certificate request, and a private key. A certificate request can be later be processed by a CA which should return a signed certificate.

/* This example code is placed in the public domain. */

#ifdef HAVE_CONFIG_H
#include <config.h>
#endif

#include <stdio.h>
#include <stdlib.h>
#include <string.h>
#include <gnutls/gnutls.h>
#include <gnutls/x509.h>
#include <gnutls/abstract.h>
#include <time.h>

/* This example will generate a private key and a certificate
 * request.
 */

int main(void)
{
        gnutls_x509_crq_t crq;
        gnutls_x509_privkey_t key;
        unsigned char buffer[10 * 1024];
        size_t buffer_size = sizeof(buffer);
        unsigned int bits;

        gnutls_global_init();

        /* Initialize an empty certificate request, and
         * an empty private key.
         */
        gnutls_x509_crq_init(&crq);

        gnutls_x509_privkey_init(&key);

        /* Generate an RSA key of moderate security.
         */
        bits =
            gnutls_sec_param_to_pk_bits(GNUTLS_PK_RSA,
                                        GNUTLS_SEC_PARAM_MEDIUM);
        gnutls_x509_privkey_generate(key, GNUTLS_PK_RSA, bits, 0);

        /* Add stuff to the distinguished name
         */
        gnutls_x509_crq_set_dn_by_oid(crq, GNUTLS_OID_X520_COUNTRY_NAME,
                                      0, "GR", 2);

        gnutls_x509_crq_set_dn_by_oid(crq, GNUTLS_OID_X520_COMMON_NAME,
                                      0, "Nikos", strlen("Nikos"));

        /* Set the request version.
         */
        gnutls_x509_crq_set_version(crq, 1);

        /* Set a challenge password.
         */
        gnutls_x509_crq_set_challenge_password(crq,
                                               "something to remember here");

        /* Associate the request with the private key
         */
        gnutls_x509_crq_set_key(crq, key);

        /* Self sign the certificate request.
         */
        gnutls_x509_crq_sign2(crq, key, GNUTLS_DIG_SHA1, 0);

        /* Export the PEM encoded certificate request, and
         * display it.
         */
        gnutls_x509_crq_export(crq, GNUTLS_X509_FMT_PEM, buffer,
                               &buffer_size);

        printf("Certificate Request: \n%s", buffer);


        /* Export the PEM encoded private key, and
         * display it.
         */
        buffer_size = sizeof(buffer);
        gnutls_x509_privkey_export(key, GNUTLS_X509_FMT_PEM, buffer,
                                   &buffer_size);

        printf("\n\nPrivate key: \n%s", buffer);

        gnutls_x509_crq_deinit(crq);
        gnutls_x509_privkey_deinit(key);

        return 0;

}

4.2.2 PKIX certificate revocation lists

A certificate revocation list (CRL) is a structure issued by an authority periodically containing a list of revoked certificates serial numbers. The CRL structure is signed with the issuing authorities’ keys. A typical CRL contains the fields as shown in Table 4.5. Certificate revocation lists are used to complement the expiration date of a certificate, in order to account for other reasons of revocation, such as compromised keys, etc.

Each CRL is valid for limited amount of time and is required to provide, except for the current issuing time, also the issuing time of the next update.

The basic CRL structure functions follow.

int gnutls_x509_crl_init (gnutls_x509_crl_t * crl)

int gnutls_x509_crl_import (gnutls_x509_crl_t crl, const gnutls_datum_t * data, gnutls_x509_crt_fmt_t format)

int gnutls_x509_crl_export (gnutls_x509_crl_t crl, gnutls_x509_crt_fmt_t format, void * output_data, size_t * output_data_size)

int gnutls_x509_crl_export (gnutls_x509_crl_t crl, gnutls_x509_crt_fmt_t format, void * output_data, size_t * output_data_size)

Reading a CRL

The most important function that extracts the certificate revocation information from a CRL is gnutls_x509_crl_get_crt_serial. Other functions that return other fields of the CRL structure are also provided.

Function: int gnutls_x509_crl_get_crt_serial (gnutls_x509_crl_t crl, unsigned indx, unsigned char * serial, size_t * serial_size, time_t * t)

crl: should contain a gnutls_x509_crl_t type

indx: the index of the certificate to extract (starting from 0)

serial: where the serial number will be copied

serial_size: initially holds the size of serial

t: if non null, will hold the time this certificate was revoked

This function will retrieve the serial number of the specified, by the index, revoked certificate.

Note that this function will have performance issues in large sequences of revoked certificates. In that case use gnutls_x509_crl_iter_crt_serial() .

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

int gnutls_x509_crl_get_version (gnutls_x509_crl_t crl)

int gnutls_x509_crl_get_issuer_dn (const gnutls_x509_crl_t crl, char * buf, size_t * sizeof_buf)

int gnutls_x509_crl_get_issuer_dn2 (gnutls_x509_crl_t crl, gnutls_datum_t * dn)

time_t gnutls_x509_crl_get_this_update (gnutls_x509_crl_t crl)

time_t gnutls_x509_crl_get_next_update (gnutls_x509_crl_t crl)

int gnutls_x509_crl_get_crt_count (gnutls_x509_crl_t crl)

Generation of a CRL

The following functions can be used to generate a CRL.

int gnutls_x509_crl_set_version (gnutls_x509_crl_t crl, unsigned int version)

int gnutls_x509_crl_set_crt_serial (gnutls_x509_crl_t crl, const void * serial, size_t serial_size, time_t revocation_time)

int gnutls_x509_crl_set_crt (gnutls_x509_crl_t crl, gnutls_x509_crt_t crt, time_t revocation_time)

int gnutls_x509_crl_set_next_update (gnutls_x509_crl_t crl, time_t exp_time)

int gnutls_x509_crl_set_this_update (gnutls_x509_crl_t crl, time_t act_time)

The gnutls_x509_crl_sign2 and gnutls_x509_crl_privkey_sign functions sign the revocation list with a private key. The latter function can be used to sign with a key residing in a PKCS #11 token.

Function: int gnutls_x509_crl_sign2 (gnutls_x509_crl_t crl, gnutls_x509_crt_t issuer, gnutls_x509_privkey_t issuer_key, gnutls_digest_algorithm_t dig, unsigned int flags)

crl: should contain a gnutls_x509_crl_t type

issuer: is the certificate of the certificate issuer

issuer_key: holds the issuer’s private key

dig: The message digest to use. GNUTLS_DIG_SHA256 is the safe choice unless you know what you’re doing.

flags: must be 0

This function will sign the CRL with the issuer’s private key, and will copy the issuer’s information into the CRL.

This must be the last step in a certificate CRL since all the previously set parameters are now signed.

A known limitation of this function is, that a newly-signed CRL will not be fully functional (e.g., for signature verification), until it is exported an re-imported.

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Function: int gnutls_x509_crl_privkey_sign (gnutls_x509_crl_t crl, gnutls_x509_crt_t issuer, gnutls_privkey_t issuer_key, gnutls_digest_algorithm_t dig, unsigned int flags)

crl: should contain a gnutls_x509_crl_t type

issuer: is the certificate of the certificate issuer

issuer_key: holds the issuer’s private key

dig: The message digest to use. GNUTLS_DIG_SHA256 is the safe choice unless you know what you’re doing.

flags: must be 0

This function will sign the CRL with the issuer’s private key, and will copy the issuer’s information into the CRL.

This must be the last step in a certificate CRL since all the previously set parameters are now signed.

A known limitation of this function is, that a newly-signed CRL will not be fully functional (e.g., for signature verification), until it is exported an re-imported.

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since 2.12.0

Few extensions on the CRL structure are supported, including the CRL number extension and the authority key identifier.

int gnutls_x509_crl_set_number (gnutls_x509_crl_t crl, const void * nr, size_t nr_size)

int gnutls_x509_crl_set_authority_key_id (gnutls_x509_crl_t crl, const void * id, size_t id_size)

4.2.3 OCSP certificate status checking

Certificates may be revoked before their expiration time has been reached. There are several reasons for revoking certificates, but a typical situation is when the private key associated with a certificate has been compromised. Traditionally, Certificate Revocation Lists (CRLs) have been used by application to implement revocation checking, however, several problems with CRLs have been identified [RIVESTCRL].

The Online Certificate Status Protocol, or OCSP [RFC2560], is a widely implemented protocol which performs certificate revocation status checking. An application that wish to verify the identity of a peer will verify the certificate against a set of trusted certificates and then check whether the certificate is listed in a CRL and/or perform an OCSP check for the certificate.

Applications are typically expected to contact the OCSP server in order to request the certificate validity status. The OCSP server replies with an OCSP response. This section describes this online communication (which can be avoided when using OCSP stapled responses, for that, see OCSP stapling).

Before performing the OCSP query, the application will need to figure out the address of the OCSP server. The OCSP server address can be provided by the local user in manual configuration or may be stored in the certificate that is being checked. When stored in a certificate the OCSP server is in the extension field called the Authority Information Access (AIA). The following function extracts this information from a certificate.

int gnutls_x509_crt_get_authority_info_access (gnutls_x509_crt_t crt, unsigned int seq, int what, gnutls_datum_t * data, unsigned int * critical)

There are several functions in GnuTLS for creating and manipulating OCSP requests and responses. The general idea is that a client application creates an OCSP request object, stores some information about the certificate to check in the request, and then exports the request in DER format. The request will then need to be sent to the OCSP responder, which needs to be done by the application (GnuTLS does not send and receive OCSP packets). Normally an OCSP response is received that the application will need to import into an OCSP response object. The digital signature in the OCSP response needs to be verified against a set of trust anchors before the information in the response can be trusted.

The ASN.1 structure of OCSP requests are briefly as follows. It is useful to review the structures to get an understanding of which fields are modified by GnuTLS functions.

OCSPRequest     ::=     SEQUENCE {
    tbsRequest                  TBSRequest,
    optionalSignature   [0]     EXPLICIT Signature OPTIONAL }

TBSRequest      ::=     SEQUENCE {
    version             [0]     EXPLICIT Version DEFAULT v1,
    requestorName       [1]     EXPLICIT GeneralName OPTIONAL,
    requestList                 SEQUENCE OF Request,
    requestExtensions   [2]     EXPLICIT Extensions OPTIONAL }

Request         ::=     SEQUENCE {
    reqCert                     CertID,
    singleRequestExtensions     [0] EXPLICIT Extensions OPTIONAL }

CertID          ::=     SEQUENCE {
    hashAlgorithm       AlgorithmIdentifier,
    issuerNameHash      OCTET STRING, -- Hash of Issuer's DN
    issuerKeyHash       OCTET STRING, -- Hash of Issuers public key
    serialNumber        CertificateSerialNumber }

The basic functions to initialize, import, export and deallocate OCSP requests are the following.

int gnutls_ocsp_req_init (gnutls_ocsp_req_t * req)

void gnutls_ocsp_req_deinit (gnutls_ocsp_req_t req)

int gnutls_ocsp_req_import (gnutls_ocsp_req_t req, const gnutls_datum_t * data)

int gnutls_ocsp_req_export (gnutls_ocsp_req_t req, gnutls_datum_t * data)

int gnutls_ocsp_req_print (gnutls_ocsp_req_t req, gnutls_ocsp_print_formats_t format, gnutls_datum_t * out)

To generate an OCSP request the issuer name hash, issuer key hash, and the checked certificate’s serial number are required. There are two interfaces available for setting those in an OCSP request. The is a low-level function when you have the issuer name hash, issuer key hash, and certificate serial number in binary form. The second is more useful if you have the certificate (and its issuer) in a gnutls_x509_crt_t type. There is also a function to extract this information from existing an OCSP request.

int gnutls_ocsp_req_add_cert_id (gnutls_ocsp_req_t req, gnutls_digest_algorithm_t digest, const gnutls_datum_t * issuer_name_hash, const gnutls_datum_t * issuer_key_hash, const gnutls_datum_t * serial_number)

int gnutls_ocsp_req_add_cert (gnutls_ocsp_req_t req, gnutls_digest_algorithm_t digest, gnutls_x509_crt_t issuer, gnutls_x509_crt_t cert)

int gnutls_ocsp_req_get_cert_id (gnutls_ocsp_req_t req, unsigned indx, gnutls_digest_algorithm_t * digest, gnutls_datum_t * issuer_name_hash, gnutls_datum_t * issuer_key_hash, gnutls_datum_t * serial_number)

Each OCSP request may contain a number of extensions. Extensions are identified by an Object Identifier (OID) and an opaque data buffer whose syntax and semantics is implied by the OID. You can extract or set those extensions using the following functions.

int gnutls_ocsp_req_get_extension (gnutls_ocsp_req_t req, unsigned indx, gnutls_datum_t * oid, unsigned int * critical, gnutls_datum_t * data)

int gnutls_ocsp_req_set_extension (gnutls_ocsp_req_t req, const char * oid, unsigned int critical, const gnutls_datum_t * data)

A common OCSP Request extension is the nonce extension (OID 1.3.6.1.5.5.7.48.1.2), which is used to avoid replay attacks of earlier recorded OCSP responses. The nonce extension carries a value that is intended to be sufficiently random and unique so that an attacker will not be able to give a stale response for the same nonce.

int gnutls_ocsp_req_get_nonce (gnutls_ocsp_req_t req, unsigned int * critical, gnutls_datum_t * nonce)

int gnutls_ocsp_req_set_nonce (gnutls_ocsp_req_t req, unsigned int critical, const gnutls_datum_t * nonce)

int gnutls_ocsp_req_randomize_nonce (gnutls_ocsp_req_t req)

The OCSP response structures is a complex structure. A simplified overview of it is in Table 4.6. Note that a response may contain information on multiple certificates.

We provide basic functions for initialization, importing, exporting and deallocating OCSP responses.

int gnutls_ocsp_resp_init (gnutls_ocsp_resp_t * resp)

void gnutls_ocsp_resp_deinit (gnutls_ocsp_resp_t resp)

int gnutls_ocsp_resp_import (gnutls_ocsp_resp_t resp, const gnutls_datum_t * data)

int gnutls_ocsp_resp_export (gnutls_ocsp_resp_t resp, gnutls_datum_t * data)

int gnutls_ocsp_resp_print (gnutls_ocsp_resp_t resp, gnutls_ocsp_print_formats_t format, gnutls_datum_t * out)

The utility function that extracts the revocation as well as other information from a response is shown below.

Function: int gnutls_ocsp_resp_get_single (gnutls_ocsp_resp_t resp, unsigned indx, gnutls_digest_algorithm_t * digest, gnutls_datum_t * issuer_name_hash, gnutls_datum_t * issuer_key_hash, gnutls_datum_t * serial_number, unsigned int * cert_status, time_t * this_update, time_t * next_update, time_t * revocation_time, unsigned int * revocation_reason)

resp: should contain a gnutls_ocsp_resp_t type

indx: Specifies response number to get. Use (0) to get the first one.

digest: output variable with gnutls_digest_algorithm_t hash algorithm

issuer_name_hash: output buffer with hash of issuer’s DN

issuer_key_hash: output buffer with hash of issuer’s public key

serial_number: output buffer with serial number of certificate to check

cert_status: a certificate status, a gnutls_ocsp_cert_status_t enum.

this_update: time at which the status is known to be correct.

next_update: when newer information will be available, or (time_t)-1 if unspecified

revocation_time: when cert_status is GNUTLS_OCSP_CERT_REVOKED , holds time of revocation.

revocation_reason: revocation reason, a gnutls_x509_crl_reason_t enum.

This function will return the certificate information of the indx ’ed response in the Basic OCSP Response resp . The information returned corresponds to the OCSP SingleResponse structure except the final singleExtensions.

Each of the pointers to output variables may be NULL to indicate that the caller is not interested in that value.

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error code is returned. If you have reached the last CertID available GNUTLS_E_REQUESTED_DATA_NOT_AVAILABLE will be returned.

The possible revocation reasons available in an OCSP response are shown below.

Note, that the OCSP response needs to be verified against some set of trust anchors before it can be relied upon. It is also important to check whether the received OCSP response corresponds to the certificate being checked.

int gnutls_ocsp_resp_verify (gnutls_ocsp_resp_t resp, gnutls_x509_trust_list_t trustlist, unsigned int * verify, unsigned int flags)

int gnutls_ocsp_resp_verify_direct (gnutls_ocsp_resp_t resp, gnutls_x509_crt_t issuer, unsigned int * verify, unsigned int flags)

int gnutls_ocsp_resp_check_crt (gnutls_ocsp_resp_t resp, unsigned int indx, gnutls_x509_crt_t crt)

4.2.4 OCSP stapling

To avoid applications contacting the OCSP server directly, TLS servers can provide a "stapled" OCSP response in the TLS handshake. That way the client application needs to do nothing more. GnuTLS will automatically consider the stapled OCSP response during the TLS certificate verification (see gnutls_certificate_verify_peers2). The stapled response can be obtained using gnutls_ocsp_status_request_get.

In addition, since GnuTLS 3.5.1 the client will consider the [RFC7633] OCSP-Must-staple certificate extension, and will consider it while checking for stapled OCSP responses. If the extension is present and no OCSP staple is found the certificate verification will fail and the status code GNUTLS_CERT_MISSING_OCSP_STATUS will returned from the verification function.

GnuTLS servers can provide this response to their clients using the following functions.

void gnutls_certificate_set_ocsp_status_request_function (gnutls_certificate_credentials_t sc, gnutls_status_request_ocsp_func ocsp_func, void * ptr)

int gnutls_certificate_set_ocsp_status_request_file (gnutls_certificate_credentials_t sc, const char * response_file, unsigned idx)

int gnutls_ocsp_status_request_is_checked (gnutls_session_t session, unsigned int flags)

The simplest approach is for a server to provide the OCSP server’s response using the gnutls_certificate_set_ocsp_status_request_file. The response may be updated periodically using the following command (see also ocsptool Invocation).

ocsptool --ask --load-cert server_cert.pem --load-issuer the_issuer.pem
         --load-signer the_issuer.pem --outfile ocsp.response

4.2.5 Managing encrypted keys

Transferring or storing private keys in plain may not be a good idea, since any compromise is irreparable. Storing the keys in hardware security modules (see Smart cards and HSMs) could solve the storage problem but it is not always practical or efficient enough. This section describes ways to store and transfer encrypted private keys.

There are methods for key encryption, namely the PKCS #8, PKCS #12 and OpenSSL’s custom encrypted private key formats. The PKCS #8 and the OpenSSL’s method allow encryption of the private key, while the PKCS #12 method allows, in addition, the bundling of accompanying data into the structure. That is typically the corresponding certificate, as well as a trusted CA certificate.

High level functionality

Generic and higher level private key import functions are available, that import plain or encrypted keys and will auto-detect the encrypted key format.

Function: int gnutls_privkey_import_x509_raw (gnutls_privkey_t pkey, const gnutls_datum_t * data, gnutls_x509_crt_fmt_t format, const char * password, unsigned int flags)

pkey: The private key

data: The private key data to be imported

format: The format of the private key

password: A password (optional)

flags: an ORed sequence of gnutls_pkcs_encrypt_flags_t

This function will import the given private key to the abstract gnutls_privkey_t type.

The supported formats are basic unencrypted key, PKCS8, PKCS12, and the openssl format.

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since: 3.1.0

Function: int gnutls_x509_privkey_import2 (gnutls_x509_privkey_t key, const gnutls_datum_t * data, gnutls_x509_crt_fmt_t format, const char * password, unsigned int flags)

key: The data to store the parsed key

data: The DER or PEM encoded key.

format: One of DER or PEM

password: A password (optional)

flags: an ORed sequence of gnutls_pkcs_encrypt_flags_t

This function will import the given DER or PEM encoded key, to the native gnutls_x509_privkey_t format, irrespective of the input format. The input format is auto-detected.

The supported formats are basic unencrypted key, PKCS8, PKCS12, and the openssl format.

If the provided key is encrypted but no password was given, then GNUTLS_E_DECRYPTION_FAILED is returned. Since GnuTLS 3.4.0 this function will utilize the PIN callbacks if any.

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Any keys imported using those functions can be imported to a certificate credentials structure using gnutls_certificate_set_key, or alternatively they can be directly imported using gnutls_certificate_set_x509_key_file2.

PKCS #8 structures

PKCS #8 keys can be imported and exported as normal private keys using the functions below. An addition to the normal import functions, are a password and a flags argument. The flags can be any element of the gnutls_pkcs_encrypt_flags_t enumeration. Note however, that GnuTLS only supports the PKCS #5 PBES2 encryption scheme. Keys encrypted with the obsolete PBES1 scheme cannot be decrypted.

int gnutls_x509_privkey_import_pkcs8 (gnutls_x509_privkey_t key, const gnutls_datum_t * data, gnutls_x509_crt_fmt_t format, const char * password, unsigned int flags)

int gnutls_x509_privkey_export_pkcs8 (gnutls_x509_privkey_t key, gnutls_x509_crt_fmt_t format, const char * password, unsigned int flags, void * output_data, size_t * output_data_size)

int gnutls_x509_privkey_export2_pkcs8 (gnutls_x509_privkey_t key, gnutls_x509_crt_fmt_t format, const char * password, unsigned int flags, gnutls_datum_t * out)

PKCS #12 structures

A PKCS #12 structure [PKCS12] usually contains a user’s private keys and certificates. It is commonly used in browsers to export and import the user’s identities. A file containing such a key can be directly imported to a certificate credentials structure by using gnutls_certificate_set_x509_simple_pkcs12_file.

In GnuTLS the PKCS #12 structures are handled using the gnutls_pkcs12_t type. This is an abstract type that may hold several gnutls_pkcs12_bag_t types. The bag types are the holders of the actual data, which may be certificates, private keys or encrypted data. A bag of type encrypted should be decrypted in order for its data to be accessed.

To reduce the complexity in parsing the structures the simple helper function gnutls_pkcs12_simple_parse is provided. For more advanced uses, manual parsing of the structure is required using the functions below.

int gnutls_pkcs12_get_bag (gnutls_pkcs12_t pkcs12, int indx, gnutls_pkcs12_bag_t bag)

int gnutls_pkcs12_verify_mac (gnutls_pkcs12_t pkcs12, const char * pass)

int gnutls_pkcs12_bag_decrypt (gnutls_pkcs12_bag_t bag, const char * pass)

int gnutls_pkcs12_bag_get_count (gnutls_pkcs12_bag_t bag)

Function: int gnutls_pkcs12_simple_parse (gnutls_pkcs12_t p12, const char * password, gnutls_x509_privkey_t * key, gnutls_x509_crt_t ** chain, unsigned int * chain_len, gnutls_x509_crt_t ** extra_certs, unsigned int * extra_certs_len, gnutls_x509_crl_t * crl, unsigned int flags)

p12: A pkcs12 type

password: optional password used to decrypt the structure, bags and keys.

key: a structure to store the parsed private key.

chain: the corresponding to key certificate chain (may be NULL )

chain_len: will be updated with the number of additional (may be NULL )

extra_certs: optional pointer to receive an array of additional certificates found in the PKCS12 structure (may be NULL ).

extra_certs_len: will be updated with the number of additional certs (may be NULL ).

crl: an optional structure to store the parsed CRL (may be NULL ).

flags: should be zero or one of GNUTLS_PKCS12_SP_*

This function parses a PKCS12 structure in pkcs12 and extracts the private key, the corresponding certificate chain, any additional certificates and a CRL.

The extra_certs and extra_certs_len parameters are optional and both may be set to NULL . If either is non-NULL , then both must be set. The value for extra_certs is allocated using gnutls_malloc() .

Encrypted PKCS12 bags and PKCS8 private keys are supported, but only with password based security and the same password for all operations.

Note that a PKCS12 structure may contain many keys and/or certificates, and there is no way to identify which key/certificate pair you want. For this reason this function is useful for PKCS12 files that contain only one key/certificate pair and/or one CRL.

If the provided structure has encrypted fields but no password is provided then this function returns GNUTLS_E_DECRYPTION_FAILED .

Note that normally the chain constructed does not include self signed certificates, to comply with TLS’ requirements. If, however, the flag GNUTLS_PKCS12_SP_INCLUDE_SELF_SIGNED is specified then self signed certificates will be included in the chain.

Prior to using this function the PKCS 12 structure integrity must be verified using gnutls_pkcs12_verify_mac() .

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since: 3.1.0

int gnutls_pkcs12_bag_get_data (gnutls_pkcs12_bag_t bag, unsigned indx, gnutls_datum_t * data)

int gnutls_pkcs12_bag_get_key_id (gnutls_pkcs12_bag_t bag, unsigned indx, gnutls_datum_t * id)

int gnutls_pkcs12_bag_get_friendly_name (gnutls_pkcs12_bag_t bag, unsigned indx, char ** name)

The functions below are used to generate a PKCS #12 structure. An example of their usage is shown at PKCS12 structure generation example.

int gnutls_pkcs12_set_bag (gnutls_pkcs12_t pkcs12, gnutls_pkcs12_bag_t bag)

int gnutls_pkcs12_bag_encrypt (gnutls_pkcs12_bag_t bag, const char * pass, unsigned int flags)

int gnutls_pkcs12_generate_mac (gnutls_pkcs12_t pkcs12, const char * pass)

int gnutls_pkcs12_bag_set_data (gnutls_pkcs12_bag_t bag, gnutls_pkcs12_bag_type_t type, const gnutls_datum_t * data)

int gnutls_pkcs12_bag_set_crl (gnutls_pkcs12_bag_t bag, gnutls_x509_crl_t crl)

int gnutls_pkcs12_bag_set_crt (gnutls_pkcs12_bag_t bag, gnutls_x509_crt_t crt)

int gnutls_pkcs12_bag_set_key_id (gnutls_pkcs12_bag_t bag, unsigned indx, const gnutls_datum_t * id)

int gnutls_pkcs12_bag_set_friendly_name (gnutls_pkcs12_bag_t bag, unsigned indx, const char * name)

OpenSSL encrypted keys

Unfortunately the structures discussed in the previous sections are not the only structures that may hold an encrypted private key. For example the OpenSSL library offers a custom key encryption method. Those structures are also supported in GnuTLS with gnutls_x509_privkey_import_openssl.

Function: int gnutls_x509_privkey_import_openssl (gnutls_x509_privkey_t key, const gnutls_datum_t * data, const char * password)

key: The data to store the parsed key

data: The DER or PEM encoded key.

password: the password to decrypt the key (if it is encrypted).

This function will convert the given PEM encrypted to the native gnutls_x509_privkey_t format. The output will be stored in key .

The password should be in ASCII. If the password is not provided or wrong then GNUTLS_E_DECRYPTION_FAILED will be returned.

If the Certificate is PEM encoded it should have a header of "PRIVATE KEY" and the "DEK-Info" header.

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

4.2.6 Invoking certtool

Tool to parse and generate X.509 certificates, requests and private keys. It can be used interactively or non interactively by specifying the template command line option.

The tool accepts files or URLs supported by GnuTLS. In case PIN is required for the URL access you can provide it using the environment variables GNUTLS_PIN and GNUTLS_SO_PIN.

This section was generated by AutoGen, using the agtexi-cmd template and the option descriptions for the certtool program. This software is released under the GNU General Public License, version 3 or later.

certtool help/usage (--help)

This is the automatically generated usage text for certtool.

The text printed is the same whether selected with the help option (--help) or the more-help option (--more-help). more-help will print the usage text by passing it through a pager program. more-help is disabled on platforms without a working fork(2) function. The PAGER environment variable is used to select the program, defaulting to more. Both will exit with a status code of 0.

certtool - GnuTLS certificate tool
Usage:  certtool [ -<flag> [<val>] | --<name>[{=| }<val>] ]...

   -d, --debug=num            Enable debugging
                                - it must be in the range:
                                  0 to 9999
   -V, --verbose              More verbose output
                                - may appear multiple times
       --infile=file          Input file
                                - file must pre-exist
       --outfile=str          Output file
   -s, --generate-self-signed  Generate a self-signed certificate
   -c, --generate-certificate  Generate a signed certificate
       --generate-proxy       Generates a proxy certificate
       --generate-crl         Generate a CRL
   -u, --update-certificate   Update a signed certificate
   -p, --generate-privkey     Generate a private key
       --provable             Generate a private key or parameters from a seed using a provable method
       --verify-provable-privkey  Verify a private key generated from a seed using a provable method
       --seed=str             When generating a private key use the given hex-encoded seed
   -q, --generate-request     Generate a PKCS #10 certificate request
                                - prohibits the option 'infile'
   -e, --verify-chain         Verify a PEM encoded certificate chain
       --verify               Verify a PEM encoded certificate chain using a trusted list
       --verify-crl           Verify a CRL using a trusted list
                                - requires the option 'load-ca-certificate'
       --verify-hostname=str  Specify a hostname to be used for certificate chain verification
       --verify-email=str     Specify a email to be used for certificate chain verification
                                - prohibits the option 'verify-hostname'
       --verify-purpose=str   Specify a purpose OID to be used for certificate chain verification
       --verify-allow-broken  Allow broken algorithms, such as MD5 for verification
       --generate-dh-params   Generate PKCS #3 encoded Diffie-Hellman parameters
       --get-dh-params        Get the included PKCS #3 encoded Diffie-Hellman parameters
       --dh-info              Print information PKCS #3 encoded Diffie-Hellman parameters
       --load-privkey=str     Loads a private key file
       --load-pubkey=str      Loads a public key file
       --load-request=str     Loads a certificate request file
       --load-certificate=str Loads a certificate file
       --load-ca-privkey=str  Loads the certificate authority's private key file
       --load-ca-certificate=str Loads the certificate authority's certificate file
       --load-crl=str         Loads the provided CRL
       --load-data=str        Loads auxiliary data
       --password=str         Password to use
       --null-password        Enforce a NULL password
       --empty-password       Enforce an empty password
       --hex-numbers          Print big number in an easier format to parse
       --cprint               In certain operations it prints the information in C-friendly format
   -i, --certificate-info     Print information on the given certificate
       --fingerprint          Print the fingerprint of the given certificate
       --key-id               Print the key ID of the given certificate
       --certificate-pubkey   Print certificate's public key
       --pgp-certificate-info  Print information on the given OpenPGP certificate
       --pgp-ring-info        Print information on the given OpenPGP keyring structure
   -l, --crl-info             Print information on the given CRL structure
       --crq-info             Print information on the given certificate request
       --no-crq-extensions    Do not use extensions in certificate requests
       --p12-info             Print information on a PKCS #12 structure
       --p12-name=str         The PKCS #12 friendly name to use
       --p7-generate          Generate a PKCS #7 structure
       --p7-sign              Signs using a PKCS #7 structure
       --p7-detached-sign     Signs using a detached PKCS #7 structure
       --p7-include-cert      The signer's certificate will be included in the cert list.
                                - disabled as '--no-p7-include-cert'
                                - enabled by default
       --p7-time              Will include a timestamp in the PKCS #7 structure
                                - disabled as '--no-p7-time'
       --p7-show-data         Will show the embedded data in the PKCS #7 structure
                                - disabled as '--no-p7-show-data'
       --p7-info              Print information on a PKCS #7 structure
       --p7-verify            Verify the provided PKCS #7 structure
       --p8-info              Print information on a PKCS #8 structure
       --smime-to-p7          Convert S/MIME to PKCS #7 structure
   -k, --key-info             Print information on a private key
       --pgp-key-info         Print information on an OpenPGP private key
       --pubkey-info          Print information on a public key
       --v1                   Generate an X.509 version 1 certificate (with no extensions)
       --to-p12               Generate a PKCS #12 structure
       --to-p8                Generate a PKCS #8 structure
   -8, --pkcs8                Use PKCS #8 format for private keys
       --rsa                  Generate RSA key
       --dsa                  Generate DSA key
       --ecc                  Generate ECC (ECDSA) key
       --ecdsa                an alias for the 'ecc' option
       --hash=str             Hash algorithm to use for signing
       --inder                Use DER format for input certificates, private keys, and DH parameters
                                - disabled as '--no-inder'
       --inraw                an alias for the 'inder' option
       --outder               Use DER format for output certificates, private keys, and DH parameters
                                - disabled as '--no-outder'
       --outraw               an alias for the 'outder' option
       --bits=num             Specify the number of bits for key generate
       --curve=str            Specify the curve used for EC key generation
       --sec-param=str        Specify the security level [low, legacy, medium, high, ultra]
       --disable-quick-random  No effect
       --template=str         Template file to use for non-interactive operation
       --stdout-info          Print information to stdout instead of stderr
       --ask-pass             Enable interaction for entering password when in batch mode.
       --pkcs-cipher=str      Cipher to use for PKCS #8 and #12 operations
       --provider=str         Specify the PKCS #11 provider library
   -v, --version[=arg]        output version information and exit
   -h, --help                 display extended usage information and exit
   -!, --more-help            extended usage information passed thru pager

Options are specified by doubled hyphens and their name or by a single
hyphen and the flag character.

Tool to parse and generate X.509 certificates, requests and private keys.
It can be used interactively or non interactively by specifying the
template command line option.

The tool accepts files or URLs supported by GnuTLS.  In case PIN is
required for the URL access you can provide it using the environment
variables GNUTLS_PIN and GNUTLS_SO_PIN.

debug option (-d)

This is the “enable debugging” option. This option takes a number argument. Specifies the debug level.

generate-crl option

This is the “generate a crl” option. This option generates a CRL. When combined with –load-crl it would use the loaded CRL as base for the generated (i.e., all revoked certificates in the base will be copied to the new CRL).

provable option

This is the “generate a private key or parameters from a seed using a provable method” option. This will use the FIPS-186-4 algorithms (i.e., Shawe-Taylor) for provable key generation. When specified the private keys or parameters will be generated from a seed, and can be later validated with –verify-provable-privkey to be correctly generated from the seed. You may specify –seed or allow GnuTLS to generate one (recommended). This option can be combined with –generate-privkey or –generate-dh-params.

That option applies to RSA and DSA keys. On the DSA keys the PQG parameters are generated using the seed, and on RSA the two primes.

verify-provable-privkey option

This is the “verify a private key generated from a seed using a provable method” option. This will use the FIPS-186-4 algorithms for provable key generation. You may specify –seed or use the seed stored in the private key structure.

generate-request option (-q)

This is the “generate a pkcs #10 certificate request” option.

This option has some usage constraints. It:

• must not appear in combination with any of the following options: infile.

Will generate a PKCS #10 certificate request. To specify a private key use –load-privkey.

verify-chain option (-e)

This is the “verify a pem encoded certificate chain” option. The last certificate in the chain must be a self signed one. It can be combined with –verify-purpose or –verify-hostname.

verify option

This is the “verify a pem encoded certificate chain using a trusted list” option. The trusted certificate list can be loaded with –load-ca-certificate. If no certificate list is provided, then the system’s certificate list is used. Note that during verification multiple paths may be explored. On a successful verification the successful path will be the last one. It can be combined with –verify-purpose or –verify-hostname.

verify-crl option

This is the “verify a crl using a trusted list” option.

This option has some usage constraints. It:

• must appear in combination with the following options: load-ca-certificate.

The trusted certificate list must be loaded with –load-ca-certificate.

verify-hostname option

This is the “specify a hostname to be used for certificate chain verification” option. This option takes a string argument. This is to be combined with one of the verify certificate options.

verify-email option

This is the “specify a email to be used for certificate chain verification” option. This option takes a string argument.

This option has some usage constraints. It:

• must not appear in combination with any of the following options: verify-hostname.

This is to be combined with one of the verify certificate options.

verify-purpose option

This is the “specify a purpose oid to be used for certificate chain verification” option. This option takes a string argument. This object identifier restricts the purpose of the certificates to be verified. Example purposes are 1.3.6.1.5.5.7.3.1 (TLS WWW), 1.3.6.1.5.5.7.3.4 (EMAIL) etc. Note that a CA certificate without a purpose set (extended key usage) is valid for any purpose.

verify-allow-broken option

This is the “allow broken algorithms, such as md5 for verification” option. This can be combined with –p7-verify, –verify or –verify-chain.

generate-dh-params option

This is the “generate pkcs #3 encoded diffie-hellman parameters” option. The will generate random parameters to be used with Diffie-Hellman key exchange. The output parameters will be in PKCS #3 format. Note that it is recommended to use the –get-dh-params option instead.

get-dh-params option

This is the “get the included pkcs #3 encoded diffie-hellman parameters” option. Returns stored DH parameters in GnuTLS. Those parameters returned are defined in RFC7919, and can be considered standard parameters for a TLS key exchange.

load-privkey option

This is the “loads a private key file” option. This option takes a string argument. This can be either a file or a PKCS #11 URL

load-pubkey option

This is the “loads a public key file” option. This option takes a string argument. This can be either a file or a PKCS #11 URL

load-request option

This is the “loads a certificate request file” option. This option takes a string argument. This option can be used with a file

load-certificate option

This is the “loads a certificate file” option. This option takes a string argument. This option can be used with a file

load-ca-privkey option

This is the “loads the certificate authority’s private key file” option. This option takes a string argument. This can be either a file or a PKCS #11 URL

load-ca-certificate option

This is the “loads the certificate authority’s certificate file” option. This option takes a string argument. This option can be used with a file

load-crl option

This is the “loads the provided crl” option. This option takes a string argument. This option can be used with a file

load-data option

This is the “loads auxiliary data” option. This option takes a string argument. This option can be used with a file

password option

This is the “password to use” option. This option takes a string argument. You can use this option to specify the password in the command line instead of reading it from the tty. Note, that the command line arguments are available for view in others in the system. Specifying password as ” is the same as specifying no password.

null-password option

This is the “enforce a null password” option. This option enforces a NULL password. This is different than the empty or no password in schemas like PKCS #8.

empty-password option

This is the “enforce an empty password” option. This option enforces an empty password. This is different than the NULL or no password in schemas like PKCS #8.

cprint option

This is the “in certain operations it prints the information in c-friendly format” option. In certain operations it prints the information in C-friendly format, suitable for including into C programs.

fingerprint option

This is the “print the fingerprint of the given certificate” option. This is a simple hash of the DER encoding of the certificate. It can be combined with the –hash parameter. However, it is recommended for identification to use the key-id which depends only on the certificate’s key.

key-id option

This is the “print the key id of the given certificate” option. This is a hash of the public key of the given certificate. It identifies the key uniquely, remains the same on a certificate renewal and depends only on signed fields of the certificate.

p12-info option

This is the “print information on a pkcs #12 structure” option. This option will dump the contents and print the metadata of the provided PKCS #12 structure.

p12-name option

This is the “the pkcs #12 friendly name to use” option. This option takes a string argument. The name to be used for the primary certificate and private key in a PKCS #12 file.

p7-generate option

This is the “generate a pkcs #7 structure” option. This option generates a PKCS #7 certificate container structure. To add certificates in the structure use –load-certificate and –load-crl.

p7-sign option

This is the “signs using a pkcs #7 structure” option. This option generates a PKCS #7 structure containing a signature for the provided data from infile. The data are stored within the structure. The signer certificate has to be specified using –load-certificate and –load-privkey. The input to –load-certificate can be a list of certificates. In case of a list, the first certificate is used for signing and the other certificates are included in the structure.

p7-detached-sign option

This is the “signs using a detached pkcs #7 structure” option. This option generates a PKCS #7 structure containing a signature for the provided data from infile. The signer certificate has to be specified using –load-certificate and –load-privkey. The input to –load-certificate can be a list of certificates. In case of a list, the first certificate is used for signing and the other certificates are included in the structure.

p7-include-cert option

This is the “the signer’s certificate will be included in the cert list.” option.

This option has some usage constraints. It:

• can be disabled with –no-p7-include-cert.
• It is enabled by default.

This options works with –p7-sign or –p7-detached-sign and will include or exclude the signer’s certificate into the generated signature.

p7-time option

This is the “will include a timestamp in the pkcs #7 structure” option.

This option has some usage constraints. It:

• can be disabled with –no-p7-time.

This option will include a timestamp in the generated signature

p7-show-data option

This is the “will show the embedded data in the pkcs #7 structure” option.

This option has some usage constraints. It:

• can be disabled with –no-p7-show-data.

This option can be combined with –p7-verify or –p7-info and will display the embedded signed data in the PKCS #7 structure.

p7-verify option

This is the “verify the provided pkcs #7 structure” option. This option verifies the signed PKCS #7 structure. The certificate list to use for verification can be specified with –load-ca-certificate. When no certificate list is provided, then the system’s certificate list is used. Alternatively a direct signer can be provided using –load-certificate. A key purpose can be enforced with the –verify-purpose option, and the –load-data option will utilize detached data.

p8-info option

This is the “print information on a pkcs #8 structure” option. This option will print information about encrypted PKCS #8 structures. That option does not require the decryption of the structure.

pubkey-info option

This is the “print information on a public key” option. The option combined with –load-request, –load-pubkey, –load-privkey and –load-certificate will extract the public key of the object in question.

to-p12 option

This is the “generate a pkcs #12 structure” option. It requires a certificate, a private key and possibly a CA certificate to be specified.

rsa option

This is the “generate rsa key” option. When combined with –generate-privkey generates an RSA private key.

dsa option

This is the “generate dsa key” option. When combined with –generate-privkey generates a DSA private key.

ecc option

This is the “generate ecc (ecdsa) key” option. When combined with –generate-privkey generates an elliptic curve private key to be used with ECDSA.

ecdsa option

This is an alias for the ecc option, see the ecc option documentation.

hash option

This is the “hash algorithm to use for signing” option. This option takes a string argument. Available hash functions are SHA1, RMD160, SHA256, SHA384, SHA512, SHA3-224, SHA3-256, SHA3-384, SHA3-512.

inder option

This is the “use der format for input certificates, private keys, and dh parameters ” option.

This option has some usage constraints. It:

• can be disabled with –no-inder.

The input files will be assumed to be in DER or RAW format. Unlike options that in PEM input would allow multiple input data (e.g. multiple certificates), when reading in DER format a single data structure is read.

inraw option

This is an alias for the inder option, see the inder option documentation.

outder option

This is the “use der format for output certificates, private keys, and dh parameters” option.

This option has some usage constraints. It:

• can be disabled with –no-outder.

The output will be in DER or RAW format.

outraw option

This is an alias for the outder option, see the outder option documentation.

curve option

This is the “specify the curve used for ec key generation” option. This option takes a string argument. Supported values are secp192r1, secp224r1, secp256r1, secp384r1 and secp521r1.

sec-param option

This is the “specify the security level [low, legacy, medium, high, ultra]” option. This option takes a string argument Security parameter. This is alternative to the bits option.

ask-pass option

This is the “enable interaction for entering password when in batch mode.” option. This option will enable interaction to enter password when in batch mode. That is useful when the template option has been specified.

pkcs-cipher option

This is the “cipher to use for pkcs #8 and #12 operations” option. This option takes a string argument Cipher. Cipher may be one of 3des, 3des-pkcs12, aes-128, aes-192, aes-256, rc2-40, arcfour.

provider option

This is the “specify the pkcs #11 provider library” option. This option takes a string argument. This will override the default options in /etc/gnutls/pkcs11.conf

certtool exit status

One of the following exit values will be returned:

‘0 (EXIT_SUCCESS)’

Successful program execution.

‘1 (EXIT_FAILURE)’

The operation failed or the command syntax was not valid.

certtool See Also

p11tool (1)

certtool Examples

Generating private keys

To create an RSA private key, run:

$ certtool --generate-privkey --outfile key.pem --rsa

To create a DSA or elliptic curves (ECDSA) private key use the above command combined with ’dsa’ or ’ecc’ options.

Generating certificate requests

To create a certificate request (needed when the certificate is issued by another party), run:

certtool --generate-request --load-privkey key.pem \
   --outfile request.pem

If the private key is stored in a smart card you can generate a request by specifying the private key object URL.

$ ./certtool --generate-request --load-privkey "pkcs11:..." \
  --load-pubkey "pkcs11:..." --outfile request.pem

Generating a self-signed certificate

To create a self signed certificate, use the command:

$ certtool --generate-privkey --outfile ca-key.pem
$ certtool --generate-self-signed --load-privkey ca-key.pem \
   --outfile ca-cert.pem

Note that a self-signed certificate usually belongs to a certificate authority, that signs other certificates.

Generating a certificate

To generate a certificate using the previous request, use the command:

$ certtool --generate-certificate --load-request request.pem \
   --outfile cert.pem --load-ca-certificate ca-cert.pem \
   --load-ca-privkey ca-key.pem

To generate a certificate using the private key only, use the command:

$ certtool --generate-certificate --load-privkey key.pem \
   --outfile cert.pem --load-ca-certificate ca-cert.pem \
   --load-ca-privkey ca-key.pem

Certificate information

To view the certificate information, use:

$ certtool --certificate-info --infile cert.pem

PKCS #12 structure generation

To generate a PKCS #12 structure using the previous key and certificate, use the command:

$ certtool --load-certificate cert.pem --load-privkey key.pem \
   --to-p12 --outder --outfile key.p12

Some tools (reportedly web browsers) have problems with that file because it does not contain the CA certificate for the certificate. To work around that problem in the tool, you can use the –load-ca-certificate parameter as follows:

$ certtool --load-ca-certificate ca.pem \
  --load-certificate cert.pem --load-privkey key.pem \
  --to-p12 --outder --outfile key.p12

Diffie-Hellman parameter generation

To generate parameters for Diffie-Hellman key exchange, use the command:

$ certtool --generate-dh-params --outfile dh.pem --sec-param medium

Proxy certificate generation

Proxy certificate can be used to delegate your credential to a temporary, typically short-lived, certificate. To create one from the previously created certificate, first create a temporary key and then generate a proxy certificate for it, using the commands:

$ certtool --generate-privkey > proxy-key.pem
$ certtool --generate-proxy --load-ca-privkey key.pem \
  --load-privkey proxy-key.pem --load-certificate cert.pem \
  --outfile proxy-cert.pem

Certificate revocation list generation

To create an empty Certificate Revocation List (CRL) do:

$ certtool --generate-crl --load-ca-privkey x509-ca-key.pem \
           --load-ca-certificate x509-ca.pem

To create a CRL that contains some revoked certificates, place the certificates in a file and use --load-certificate as follows:

$ certtool --generate-crl --load-ca-privkey x509-ca-key.pem \
  --load-ca-certificate x509-ca.pem --load-certificate revoked-certs.pem

To verify a Certificate Revocation List (CRL) do:

$ certtool --verify-crl --load-ca-certificate x509-ca.pem < crl.pem

certtool Files

Certtool’s template file format

A template file can be used to avoid the interactive questions of certtool. Initially create a file named ’cert.cfg’ that contains the information about the certificate. The template can be used as below:

$ certtool --generate-certificate --load-privkey key.pem  \
   --template cert.cfg --outfile cert.pem \
   --load-ca-certificate ca-cert.pem --load-ca-privkey ca-key.pem

An example certtool template file that can be used to generate a certificate request or a self signed certificate follows.

# X.509 Certificate options
#
# DN options

# The organization of the subject.
organization = "Koko inc."

# The organizational unit of the subject.
unit = "sleeping dept."

# The locality of the subject.
# locality =

# The state of the certificate owner.
state = "Attiki"

# The country of the subject. Two letter code.
country = GR

# The common name of the certificate owner.
cn = "Cindy Lauper"

# A user id of the certificate owner.
#uid = "clauper"

# Set domain components
#dc = "name"
#dc = "domain"

# If the supported DN OIDs are not adequate you can set
# any OID here.
# For example set the X.520 Title and the X.520 Pseudonym
# by using OID and string pairs.
#dn_oid = "2.5.4.12 Dr."
#dn_oid = "2.5.4.65 jackal"

# This is deprecated and should not be used in new
# certificates.
# pkcs9_email = "none@none.org"

# An alternative way to set the certificate's distinguished name directly
# is with the "dn" option. The attribute names allowed are:
# C (country), street, O (organization), OU (unit), title, CN (common name),
# L (locality), ST (state), placeOfBirth, gender, countryOfCitizenship, 
# countryOfResidence, serialNumber, telephoneNumber, surName, initials, 
# generationQualifier, givenName, pseudonym, dnQualifier, postalCode, name, 
# businessCategory, DC, UID, jurisdictionOfIncorporationLocalityName, 
# jurisdictionOfIncorporationStateOrProvinceName,
# jurisdictionOfIncorporationCountryName, XmppAddr, and numeric OIDs.

#dn = "cn = Nikos,st = New\, Something,C=GR,surName=Mavrogiannopoulos,2.5.4.9=Arkadias"

# The serial number of the certificate
# Comment the field for a time-based serial number.
serial = 007

# In how many days, counting from today, this certificate will expire.
# Use -1 if there is no expiration date.
expiration_days = 700

# Alternatively you may set concrete dates and time. The GNU date string 
# formats are accepted. See:
# http://www.gnu.org/software/tar/manual/html_node/Date-input-formats.html

#activation_date = "2004-02-29 16:21:42"
#expiration_date = "2025-02-29 16:24:41"

# X.509 v3 extensions

# A dnsname in case of a WWW server.
#dns_name = "www.none.org"
#dns_name = "www.morethanone.org"

# An othername defined by an OID and a hex encoded string
#other_name = "1.3.6.1.5.2.2 302ca00d1b0b56414e5245494e2e4f5247a11b3019a006020400000002a10f300d1b047269636b1b0561646d696e"
#other_name_utf8 = "1.2.4.5.6 A UTF8 string"
#other_name_octet = "1.2.4.5.6 A string that will be encoded as ASN.1 octet string"

# Allows writing an XmppAddr Identifier
#xmpp_name = juliet@im.example.com

# Names used in PKINIT
#krb5_principal = user@REALM.COM
#krb5_principal = HTTP/user@REALM.COM

# A subject alternative name URI
#uri = "http://www.example.com"

# An IP address in case of a server.
#ip_address = "192.168.1.1"

# An email in case of a person
email = "none@none.org"

# TLS feature (rfc7633) extension. That can is used to indicate mandatory TLS
# extension features to be provided by the server. In practice this is used
# to require the Status Request (extid: 5) extension from the server. That is,
# to require the server holding this certificate to provide a stapled OCSP response.
# You can have multiple lines for multiple TLS features.

# To ask for OCSP status request use:
#tls_feature = 5

# Challenge password used in certificate requests
challenge_password = 123456

# Password when encrypting a private key
#password = secret

# An URL that has CRLs (certificate revocation lists)
# available. Needed in CA certificates.
#crl_dist_points = "http://www.getcrl.crl/getcrl/"

# Whether this is a CA certificate or not
#ca

# Subject Unique ID (in hex)
#subject_unique_id = 00153224

# Issuer Unique ID (in hex)
#issuer_unique_id = 00153225

#### Key usage

# The following key usage flags are used by CAs and end certificates

# Whether this certificate will be used to sign data (needed
# in TLS DHE ciphersuites). This is the digitalSignature flag
# in RFC5280 terminology.
signing_key

# Whether this certificate will be used to encrypt data (needed
# in TLS RSA ciphersuites). Note that it is preferred to use different
# keys for encryption and signing. This is the keyEncipherment flag
# in RFC5280 terminology.
encryption_key

# Whether this key will be used to sign other certificates. The
# keyCertSign flag in RFC5280 terminology.
#cert_signing_key

# Whether this key will be used to sign CRLs. The
# cRLSign flag in RFC5280 terminology.
#crl_signing_key

# The keyAgreement flag of RFC5280. It's purpose is loosely
# defined. Not use it unless required by a protocol.
#key_agreement

# The dataEncipherment flag of RFC5280. It's purpose is loosely
# defined. Not use it unless required by a protocol.
#data_encipherment

# The nonRepudiation flag of RFC5280. It's purpose is loosely
# defined. Not use it unless required by a protocol.
#non_repudiation

#### Extended key usage (key purposes)

# The following extensions are used in an end certificate
# to clarify its purpose. Some CAs also use it to indicate
# the types of certificates they are purposed to sign.


# Whether this certificate will be used for a TLS client;
# this sets the id-kp-serverAuth (1.3.6.1.5.5.7.3.1) of 
# extended key usage.
#tls_www_client

# Whether this certificate will be used for a TLS server;
# This sets the id-kp-clientAuth (1.3.6.1.5.5.7.3.2) of 
# extended key usage.
#tls_www_server

# Whether this key will be used to sign code. This sets the
# id-kp-codeSigning (1.3.6.1.5.5.7.3.3) of extended key usage
# extension.
#code_signing_key

# Whether this key will be used to sign OCSP data. This sets the
# id-kp-OCSPSigning (1.3.6.1.5.5.7.3.9) of extended key usage extension.
#ocsp_signing_key

# Whether this key will be used for time stamping. This sets the
# id-kp-timeStamping (1.3.6.1.5.5.7.3.8) of extended key usage extension.
#time_stamping_key

# Whether this key will be used for email protection. This sets the
# id-kp-emailProtection (1.3.6.1.5.5.7.3.4) of extended key usage extension.
#email_protection_key

# Whether this key will be used for IPsec IKE operations (1.3.6.1.5.5.7.3.17).
#ipsec_ike_key

## adding custom key purpose OIDs

# for microsoft smart card logon
# key_purpose_oid = 1.3.6.1.4.1.311.20.2.2

# for email protection
# key_purpose_oid = 1.3.6.1.5.5.7.3.4

# for any purpose (must not be used in intermediate CA certificates)
# key_purpose_oid = 2.5.29.37.0

### end of key purpose OIDs

### Adding arbitrary extensions
# This requires to provide the extension OIDs, as well as the extension data in
# hex format. The following two options are available since GnuTLS 3.5.3.
#add_extension = "1.2.3.4 0x0AAB01ACFE"

# As above but encode the data as an octet string
#add_extension = "1.2.3.4 octet_string(0x0AAB01ACFE)"

# For portability critical extensions shouldn't be set to certificates.
#add_critical_extension = "5.6.7.8 0x1AAB01ACFE"

# When generating a certificate from a certificate
# request, then honor the extensions stored in the request
# and store them in the real certificate.
#honor_crq_extensions

# Alternatively only specific extensions can be copied.
#honor_crq_ext = 2.5.29.17
#honor_crq_ext = 2.5.29.15

# Path length contraint. Sets the maximum number of
# certificates that can be used to certify this certificate.
# (i.e. the certificate chain length)
#path_len = -1
#path_len = 2

# OCSP URI
# ocsp_uri = http://my.ocsp.server/ocsp

# CA issuers URI
# ca_issuers_uri = http://my.ca.issuer

# Certificate policies
#policy1 = 1.3.6.1.4.1.5484.1.10.99.1.0
#policy1_txt = "This is a long policy to summarize"
#policy1_url = http://www.example.com/a-policy-to-read

#policy2 = 1.3.6.1.4.1.5484.1.10.99.1.1
#policy2_txt = "This is a short policy"
#policy2_url = http://www.example.com/another-policy-to-read

# Name constraints

# DNS
#nc_permit_dns = example.com
#nc_exclude_dns = test.example.com

# EMAIL
#nc_permit_email = "nmav@ex.net"

# Exclude subdomains of example.com
#nc_exclude_email = .example.com

# Exclude all e-mail addresses of example.com
#nc_exclude_email = example.com

# IP
#nc_permit_ip = 192.168.0.0/16
#nc_exclude_ip = 192.168.5.0/24
#nc_permit_ip = fc0a:eef2:e7e7:a56e::/64


# Options for proxy certificates
#proxy_policy_language = 1.3.6.1.5.5.7.21.1


# Options for generating a CRL

# The number of days the next CRL update will be due.
# next CRL update will be in 43 days
#crl_next_update = 43

# this is the 5th CRL by this CA
# Comment the field for a time-based number.
#crl_number = 5

# Specify the update dates more precisely.
#crl_this_update_date = "2004-02-29 16:21:42"
#crl_next_update_date = "2025-02-29 16:24:41"

# The date that the certificates will be made seen as
# being revoked.
#crl_revocation_date = "2025-02-29 16:24:41"

4.2.7 Invoking ocsptool

Ocsptool is a program that can parse and print information about OCSP requests/responses, generate requests and verify responses.

This section was generated by AutoGen, using the agtexi-cmd template and the option descriptions for the ocsptool program. This software is released under the GNU General Public License, version 3 or later.

ocsptool help/usage (--help)

This is the automatically generated usage text for ocsptool.

The text printed is the same whether selected with the help option (--help) or the more-help option (--more-help). more-help will print the usage text by passing it through a pager program. more-help is disabled on platforms without a working fork(2) function. The PAGER environment variable is used to select the program, defaulting to more. Both will exit with a status code of 0.

ocsptool - GnuTLS OCSP tool
Usage:  ocsptool [ -<flag> [<val>] | --<name>[{=| }<val>] ]...

   -d, --debug=num            Enable debugging
                                - it must be in the range:
                                  0 to 9999
   -V, --verbose              More verbose output
                                - may appear multiple times
       --infile=file          Input file
                                - file must pre-exist
       --outfile=str          Output file
       --ask[=arg]            Ask an OCSP/HTTP server on a certificate validity
                                - requires these options:
                                load-cert
                                load-issuer
   -e, --verify-response      Verify response
   -i, --request-info         Print information on a OCSP request
   -j, --response-info        Print information on a OCSP response
   -q, --generate-request     Generate an OCSP request
       --nonce                Use (or not) a nonce to OCSP request
                                - disabled as '--no-nonce'
       --load-issuer=file     Read issuer certificate from file
                                - file must pre-exist
       --load-cert=file       Read certificate to check from file
                                - file must pre-exist
       --load-trust=file      Read OCSP trust anchors from file
                                - prohibits the option 'load-signer'
                                - file must pre-exist
       --load-signer=file     Read OCSP response signer from file
                                - prohibits the option 'load-trust'
                                - file must pre-exist
       --inder                Use DER format for input certificates and private keys
                                - disabled as '--no-inder'
   -Q, --load-request=file    Read DER encoded OCSP request from file
                                - file must pre-exist
   -S, --load-response=file   Read DER encoded OCSP response from file
                                - file must pre-exist
       --ignore-errors        Ignore any verification errors
   -v, --version[=arg]        output version information and exit
   -h, --help                 display extended usage information and exit
   -!, --more-help            extended usage information passed thru pager

Options are specified by doubled hyphens and their name or by a single
hyphen and the flag character.

Ocsptool is a program that can parse and print information about OCSP
requests/responses, generate requests and verify responses.

debug option (-d)

This is the “enable debugging” option. This option takes a number argument. Specifies the debug level.

ask option

This is the “ask an ocsp/http server on a certificate validity” option. This option takes an optional string argument server name|url.

This option has some usage constraints. It:

• must appear in combination with the following options: load-cert, load-issuer.

Connects to the specified HTTP OCSP server and queries on the validity of the loaded certificate.

ocsptool exit status

One of the following exit values will be returned:

‘0 (EXIT_SUCCESS)’

Successful program execution.

‘1 (EXIT_FAILURE)’

The operation failed or the command syntax was not valid.

ocsptool See Also

certtool (1)

ocsptool Examples

Print information about an OCSP request

To parse an OCSP request and print information about the content, the -i or --request-info parameter may be used as follows. The -Q parameter specify the name of the file containing the OCSP request, and it should contain the OCSP request in binary DER format.

$ ocsptool -i -Q ocsp-request.der

The input file may also be sent to standard input like this:

$ cat ocsp-request.der | ocsptool --request-info

Print information about an OCSP response

Similar to parsing OCSP requests, OCSP responses can be parsed using the -j or --response-info as follows.

$ ocsptool -j -Q ocsp-response.der
$ cat ocsp-response.der | ocsptool --response-info

Generate an OCSP request

The -q or --generate-request parameters are used to generate an OCSP request. By default the OCSP request is written to standard output in binary DER format, but can be stored in a file using --outfile. To generate an OCSP request the issuer of the certificate to check needs to be specified with --load-issuer and the certificate to check with --load-cert. By default PEM format is used for these files, although --inder can be used to specify that the input files are in DER format.

$ ocsptool -q --load-issuer issuer.pem --load-cert client.pem \
           --outfile ocsp-request.der

When generating OCSP requests, the tool will add an OCSP extension containing a nonce. This behaviour can be disabled by specifying --no-nonce.

Verify signature in OCSP response

To verify the signature in an OCSP response the -e or --verify-response parameter is used. The tool will read an OCSP response in DER format from standard input, or from the file specified by --load-response. The OCSP response is verified against a set of trust anchors, which are specified using --load-trust. The trust anchors are concatenated certificates in PEM format. The certificate that signed the OCSP response needs to be in the set of trust anchors, or the issuer of the signer certificate needs to be in the set of trust anchors and the OCSP Extended Key Usage bit has to be asserted in the signer certificate.

$ ocsptool -e --load-trust issuer.pem \
           --load-response ocsp-response.der

The tool will print status of verification.

Verify signature in OCSP response against given certificate

It is possible to override the normal trust logic if you know that a certain certificate is supposed to have signed the OCSP response, and you want to use it to check the signature. This is achieved using --load-signer instead of --load-trust. This will load one certificate and it will be used to verify the signature in the OCSP response. It will not check the Extended Key Usage bit.

$ ocsptool -e --load-signer ocsp-signer.pem \
           --load-response ocsp-response.der

This approach is normally only relevant in two situations. The first is when the OCSP response does not contain a copy of the signer certificate, so the --load-trust code would fail. The second is if you want to avoid the indirect mode where the OCSP response signer certificate is signed by a trust anchor.

Real-world example

Here is an example of how to generate an OCSP request for a certificate and to verify the response. For illustration we’ll use the blog.josefsson.org host, which (as of writing) uses a certificate from CACert. First we’ll use gnutls-cli to get a copy of the server certificate chain. The server is not required to send this information, but this particular one is configured to do so.

$ echo | gnutls-cli -p 443 blog.josefsson.org --print-cert > chain.pem

Use a text editor on chain.pem to create three files for each separate certificates, called cert.pem for the first certificate for the domain itself, secondly issuer.pem for the intermediate certificate and root.pem for the final root certificate.

The domain certificate normally contains a pointer to where the OCSP responder is located, in the Authority Information Access Information extension. For example, from certtool -i < cert.pem there is this information:

Authority Information Access Information (not critical):
Access Method: 1.3.6.1.5.5.7.48.1 (id-ad-ocsp)
Access Location URI: http://ocsp.CAcert.org/

This means the CA support OCSP queries over HTTP. We are now ready to create a OCSP request for the certificate.

$ ocsptool --ask ocsp.CAcert.org --load-issuer issuer.pem \
           --load-cert cert.pem --outfile ocsp-response.der

The request is sent via HTTP to the OCSP server address specified. If the address is omitted ocsptool will use the address stored in the certificate.

4.2.8 Invoking danetool

Tool to generate and check DNS resource records for the DANE protocol.

This section was generated by AutoGen, using the agtexi-cmd template and the option descriptions for the danetool program. This software is released under the GNU General Public License, version 3 or later.

danetool help/usage (--help)

This is the automatically generated usage text for danetool.

The text printed is the same whether selected with the help option (--help) or the more-help option (--more-help). more-help will print the usage text by passing it through a pager program. more-help is disabled on platforms without a working fork(2) function. The PAGER environment variable is used to select the program, defaulting to more. Both will exit with a status code of 0.

danetool - GnuTLS DANE tool
Usage:  danetool [ -<flag> [<val>] | --<name>[{=| }<val>] ]...

   -d, --debug=num            Enable debugging
                                - it must be in the range:
                                  0 to 9999
   -V, --verbose              More verbose output
                                - may appear multiple times
       --infile=file          Input file
                                - file must pre-exist
       --outfile=str          Output file
       --load-pubkey=str      Loads a public key file
       --load-certificate=str Loads a certificate file
       --dlv=str              Sets a DLV file
       --hash=str             Hash algorithm to use for signing
       --check=str            Check a host's DANE TLSA entry
       --check-ee             Check only the end-entity's certificate
       --check-ca             Check only the CA's certificate
       --tlsa-rr              Print the DANE RR data on a certificate or public key
                                - requires the option 'host'
       --host=str             Specify the hostname to be used in the DANE RR
       --proto=str            The protocol set for DANE data (tcp, udp etc.)
       --port=str             The port or service to connect to, for DANE data
       --app-proto=str        an alias for the 'starttls-proto' option
       --starttls-proto=str   The application protocol to be used to obtain the server's certificate
(https, ftp, smtp, imap, ldap, xmpp, lmtp, pop3, nntp, sieve, postgres)
       --ca                   Whether the provided certificate or public key is a Certificate
Authority
       --x509                 Use the hash of the X.509 certificate, rather than the public key
       --local                an alias for the 'domain' option
                                - enabled by default
       --domain               The provided certificate or public key is issued by the local domain
                                - disabled as '--no-domain'
                                - enabled by default
       --local-dns            Use the local DNS server for DNSSEC resolving
                                - disabled as '--no-local-dns'
       --insecure             Do not verify any DNSSEC signature
       --inder                Use DER format for input certificates and private keys
                                - disabled as '--no-inder'
       --inraw                an alias for the 'inder' option
       --print-raw            Print the received DANE data in raw format
                                - disabled as '--no-print-raw'
       --quiet                Suppress several informational messages
   -v, --version[=arg]        output version information and exit
   -h, --help                 display extended usage information and exit
   -!, --more-help            extended usage information passed thru pager

Options are specified by doubled hyphens and their name or by a single
hyphen and the flag character.

Tool to generate and check DNS resource records for the DANE protocol.

debug option (-d)

This is the “enable debugging” option. This option takes a number argument. Specifies the debug level.

load-pubkey option

This is the “loads a public key file” option. This option takes a string argument. This can be either a file or a PKCS #11 URL

load-certificate option

This is the “loads a certificate file” option. This option takes a string argument. This can be either a file or a PKCS #11 URL

dlv option

This is the “sets a dlv file” option. This option takes a string argument. This sets a DLV file to be used for DNSSEC verification.

hash option

This is the “hash algorithm to use for signing” option. This option takes a string argument. Available hash functions are SHA1, RMD160, SHA256, SHA384, SHA512.

check option

This is the “check a host’s dane tlsa entry” option. This option takes a string argument. Obtains the DANE TLSA entry from the given hostname and prints information. Note that the actual certificate of the host can be provided using –load-certificate, otherwise danetool will connect to the server to obtain it. The exit code on verification success will be zero.

check-ee option

This is the “check only the end-entity’s certificate” option. Checks the end-entity’s certificate only. Trust anchors or CAs are not considered.

check-ca option

This is the “check only the ca’s certificate” option. Checks the trust anchor’s and CA’s certificate only. End-entities are not considered.

tlsa-rr option

This is the “print the dane rr data on a certificate or public key” option.

This option has some usage constraints. It:

• must appear in combination with the following options: host.

This command prints the DANE RR data needed to enable DANE on a DNS server.

host option

This is the “specify the hostname to be used in the dane rr” option. This option takes a string argument Hostname. This command sets the hostname for the DANE RR.

proto option

This is the “the protocol set for dane data (tcp, udp etc.)” option. This option takes a string argument Protocol. This command specifies the protocol for the service set in the DANE data.

app-proto option

This is an alias for the starttls-proto option, see the starttls-proto option documentation.

starttls-proto option

This is the “the application protocol to be used to obtain the server’s certificate (https, ftp, smtp, imap, ldap, xmpp, lmtp, pop3, nntp, sieve, postgres)” option. This option takes a string argument. When the server’s certificate isn’t provided danetool will connect to the server to obtain the certificate. In that case it is required to know the protocol to talk with the server prior to initiating the TLS handshake.

ca option

This is the “whether the provided certificate or public key is a certificate authority” option. Marks the DANE RR as a CA certificate if specified.

x509 option

This is the “use the hash of the x.509 certificate, rather than the public key” option. This option forces the generated record to contain the hash of the full X.509 certificate. By default only the hash of the public key is used.

local option

This is an alias for the domain option, see the domain option documentation.

domain option

This is the “the provided certificate or public key is issued by the local domain” option.

This option has some usage constraints. It:

• can be disabled with –no-domain.
• It is enabled by default.

DANE distinguishes certificates and public keys offered via the DNSSEC to trusted and local entities. This flag indicates that this is a domain-issued certificate, meaning that there could be no CA involved.

local-dns option

This is the “use the local dns server for dnssec resolving” option.

This option has some usage constraints. It:

• can be disabled with –no-local-dns.

This option will use the local DNS server for DNSSEC. This is disabled by default due to many servers not allowing DNSSEC.

insecure option

This is the “do not verify any dnssec signature” option. Ignores any DNSSEC signature verification results.

inder option

This is the “use der format for input certificates and private keys” option.

This option has some usage constraints. It:

• can be disabled with –no-inder.

The input files will be assumed to be in DER or RAW format. Unlike options that in PEM input would allow multiple input data (e.g. multiple certificates), when reading in DER format a single data structure is read.

inraw option

This is an alias for the inder option, see the inder option documentation.

print-raw option

This is the “print the received dane data in raw format” option.

This option has some usage constraints. It:

• can be disabled with –no-print-raw.

This option will print the received DANE data.

quiet option

This is the “suppress several informational messages” option. In that case on the exit code can be used as an indication of verification success

danetool exit status

One of the following exit values will be returned:

‘0 (EXIT_SUCCESS)’

Successful program execution.

‘1 (EXIT_FAILURE)’

The operation failed or the command syntax was not valid.

danetool See Also

certtool (1)

danetool Examples

DANE TLSA RR generation

To create a DANE TLSA resource record for a certificate (or public key) that was issued localy and may or may not be signed by a CA use the following command.

$ danetool --tlsa-rr --host www.example.com --load-certificate cert.pem

To create a DANE TLSA resource record for a CA signed certificate, which will be marked as such use the following command.

$ danetool --tlsa-rr --host www.example.com --load-certificate cert.pem \
  --no-domain

The former is useful to add in your DNS entry even if your certificate is signed by a CA. That way even users who do not trust your CA will be able to verify your certificate using DANE.

In order to create a record for the CA signer of your certificate use the following.

$ danetool --tlsa-rr --host www.example.com --load-certificate cert.pem \
  --ca --no-domain

To read a server’s DANE TLSA entry, use:

$ danetool --check www.example.com --proto tcp --port 443

To verify an HTTPS server’s DANE TLSA entry, use:

$ danetool --check www.example.com --proto tcp --port 443 --load-certificate chain.pem

To verify an SMTP server’s DANE TLSA entry, use:

$ danetool --check www.example.com --proto tcp --starttls-proto=smtp --load-certificate chain.pem

4.3 Shared-key and anonymous authentication

In addition to certificate authentication, the TLS protocol may be used with password, shared-key and anonymous authentication methods. The rest of this chapter discusses details of these methods.

4.3.1 SRP authentication

4.3.1.1 Authentication using SRP

GnuTLS supports authentication via the Secure Remote Password or SRP protocol (see [RFC2945,TOMSRP] for a description). The SRP key exchange is an extension to the TLS protocol, and it provides an authenticated with a password key exchange. The peers can be identified using a single password, or there can be combinations where the client is authenticated using SRP and the server using a certificate.

The advantage of SRP authentication, over other proposed secure password authentication schemes, is that SRP is not susceptible to off-line dictionary attacks. Moreover, SRP does not require the server to hold the user’s password. This kind of protection is similar to the one used traditionally in the UNIX /etc/passwd file, where the contents of this file did not cause harm to the system security if they were revealed. The SRP needs instead of the plain password something called a verifier, which is calculated using the user’s password, and if stolen cannot be used to impersonate the user.

Typical conventions in SRP are a password file, called tpasswd that holds the SRP verifiers (encoded passwords) and another file, tpasswd.conf, which holds the allowed SRP parameters. The included in GnuTLS helper follow those conventions. The srptool program, discussed in the next section is a tool to manipulate the SRP parameters.

The implementation in GnuTLS is based on [TLSSRP]. The supported key exchange methods are shown below.

SRP:

Authentication using the SRP protocol.

SRP_DSS:

Client authentication using the SRP protocol. Server is authenticated using a certificate with DSA parameters.

SRP_RSA:

Client authentication using the SRP protocol. Server is authenticated using a certificate with RSA parameters.

Function: int gnutls_srp_verifier (const char * username, const char * password, const gnutls_datum_t * salt, const gnutls_datum_t * generator, const gnutls_datum_t * prime, gnutls_datum_t * res)

username: is the user’s name

password: is the user’s password

salt: should be some randomly generated bytes

generator: is the generator of the group

prime: is the group’s prime

res: where the verifier will be stored.

This function will create an SRP verifier, as specified in RFC2945. The prime and generator should be one of the static parameters defined in gnutls/gnutls.h or may be generated.

The verifier will be allocated with gnutls_malloc () and will be stored in res using binary format.

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, or an error code.

int gnutls_srp_base64_encode2 (const gnutls_datum_t * data, gnutls_datum_t * result)

int gnutls_srp_base64_decode2 (const gnutls_datum_t * b64_data, gnutls_datum_t * result)

4.3.1.2 Invoking srptool

Simple program that emulates the programs in the Stanford SRP (Secure Remote Password) libraries using GnuTLS. It is intended for use in places where you don’t expect SRP authentication to be the used for system users.

In brief, to use SRP you need to create two files. These are the password file that holds the users and the verifiers associated with them and the configuration file to hold the group parameters (called tpasswd.conf).

This section was generated by AutoGen, using the agtexi-cmd template and the option descriptions for the srptool program. This software is released under the GNU General Public License, version 3 or later.

srptool help/usage (--help)

This is the automatically generated usage text for srptool.

The text printed is the same whether selected with the help option (--help) or the more-help option (--more-help). more-help will print the usage text by passing it through a pager program. more-help is disabled on platforms without a working fork(2) function. The PAGER environment variable is used to select the program, defaulting to more. Both will exit with a status code of 0.

srptool - GnuTLS SRP tool
Usage:  srptool [ -<flag> [<val>] | --<name>[{=| }<val>] ]...

   -d, --debug=num            Enable debugging
                                - it must be in the range:
                                  0 to 9999
   -i, --index=num            specify the index of the group parameters in tpasswd.conf to use
   -u, --username=str         specify a username
   -p, --passwd=str           specify a password file
   -s, --salt=num             specify salt size
       --verify               just verify the password.
   -v, --passwd-conf=str      specify a password conf file.
       --create-conf=str      Generate a password configuration file.
   -v, --version[=arg]        output version information and exit
   -h, --help                 display extended usage information and exit
   -!, --more-help            extended usage information passed thru pager

Options are specified by doubled hyphens and their name or by a single
hyphen and the flag character.

Simple program that emulates the programs in the Stanford SRP (Secure
Remote Password) libraries using GnuTLS.  It is intended for use in places
where you don't expect SRP authentication to be the used for system users.

In brief, to use SRP you need to create two files.  These are the password
file that holds the users and the verifiers associated with them and the
configuration file to hold the group parameters (called tpasswd.conf).

debug option (-d)

This is the “enable debugging” option. This option takes a number argument. Specifies the debug level.

verify option

This is the “just verify the password.” option. Verifies the password provided against the password file.

passwd-conf option (-v)

This is the “specify a password conf file.” option. This option takes a string argument. Specify a filename or a PKCS #11 URL to read the CAs from.

create-conf option

This is the “generate a password configuration file.” option. This option takes a string argument. This generates a password configuration file (tpasswd.conf) containing the required for TLS parameters.

srptool exit status

One of the following exit values will be returned:

‘0 (EXIT_SUCCESS)’

Successful program execution.

‘1 (EXIT_FAILURE)’

The operation failed or the command syntax was not valid.

srptool See Also

gnutls-cli-debug (1), gnutls-serv (1), srptool (1), psktool (1), certtool (1)

srptool Examples

To create tpasswd.conf which holds the g and n values for SRP protocol (generator and a large prime), run:

$ srptool --create-conf /etc/tpasswd.conf

This command will create /etc/tpasswd and will add user ’test’ (you will also be prompted for a password). Verifiers are stored by default in the way libsrp expects.

$ srptool --passwd /etc/tpasswd --passwd-conf /etc/tpasswd.conf -u test

This command will check against a password. If the password matches the one in /etc/tpasswd you will get an ok.

$ srptool --passwd /etc/tpasswd --passwd\-conf /etc/tpasswd.conf --verify -u test

4.3.2 PSK authentication

4.3.2.1 Authentication using PSK

Authentication using Pre-shared keys is a method to authenticate using usernames and binary keys. This protocol avoids making use of public key infrastructure and expensive calculations, thus it is suitable for constraint clients.

The implementation in GnuTLS is based on [TLSPSK]. The supported PSK key exchange methods are:

PSK:

Authentication using the PSK protocol.

DHE-PSK:

Authentication using the PSK protocol and Diffie-Hellman key exchange. This method offers perfect forward secrecy.

ECDHE-PSK:

Authentication using the PSK protocol and Elliptic curve Diffie-Hellman key exchange. This method offers perfect forward secrecy.

RSA-PSK:

Authentication using the PSK protocol for the client and an RSA certificate for the server.

Helper functions to generate and maintain PSK keys are also included in GnuTLS.

int gnutls_key_generate (gnutls_datum_t * key, unsigned int key_size)

int gnutls_hex_encode (const gnutls_datum_t * data, char * result, size_t * result_size)

int gnutls_hex_decode (const gnutls_datum_t * hex_data, void * result, size_t * result_size)

4.3.2.2 Invoking psktool

Program that generates random keys for use with TLS-PSK. The keys are stored in hexadecimal format in a key file.

This section was generated by AutoGen, using the agtexi-cmd template and the option descriptions for the psktool program. This software is released under the GNU General Public License, version 3 or later.

psktool help/usage (--help)

This is the automatically generated usage text for psktool.

The text printed is the same whether selected with the help option (--help) or the more-help option (--more-help). more-help will print the usage text by passing it through a pager program. more-help is disabled on platforms without a working fork(2) function. The PAGER environment variable is used to select the program, defaulting to more. Both will exit with a status code of 0.

psktool - GnuTLS PSK tool
Usage:  psktool [ -<flag> [<val>] | --<name>[{=| }<val>] ]...

   -d, --debug=num            Enable debugging
                                - it must be in the range:
                                  0 to 9999
   -s, --keysize=num          specify the key size in bytes
                                - it must be in the range:
                                  0 to 512
   -u, --username=str         specify a username
   -p, --passwd=str           specify a password file
   -v, --version[=arg]        output version information and exit
   -h, --help                 display extended usage information and exit
   -!, --more-help            extended usage information passed thru pager

Options are specified by doubled hyphens and their name or by a single
hyphen and the flag character.

Program that generates random keys for use with TLS-PSK.  The keys are
stored in hexadecimal format in a key file.

debug option (-d)

This is the “enable debugging” option. This option takes a number argument. Specifies the debug level.

psktool exit status

One of the following exit values will be returned:

‘0 (EXIT_SUCCESS)’

Successful program execution.

‘1 (EXIT_FAILURE)’

The operation failed or the command syntax was not valid.

psktool See Also

gnutls-cli-debug (1), gnutls-serv (1), srptool (1), certtool (1)

psktool Examples

To add a user ’psk_identity’ in passwd.psk for use with GnuTLS run:

$ ./psktool -u psk_identity -p passwd.psk
Generating a random key for user 'psk_identity'
Key stored to passwd.psk
$ cat psks.txt
psk_identity:88f3824b3e5659f52d00e959bacab954b6540344
$

This command will create passwd.psk if it does not exist and will add user ’psk_identity’ (you will also be prompted for a password).

4.3.3 Anonymous authentication

The anonymous key exchange offers encryption without any indication of the peer’s identity. This kind of authentication is vulnerable to a man in the middle attack, but can be used even if there is no prior communication or shared trusted parties with the peer. It is useful to establish a session over which certificate authentication will occur in order to hide the indentities of the participants from passive eavesdroppers.

Unless in the above case, it is not recommended to use anonymous authentication. In the cases where there is no prior communication with the peers, an alternative with better properties, such as key continuity, is trust on first use (see Verifying a certificate using trust on first use authentication).

The available key exchange algorithms for anonymous authentication are shown below, but note that few public servers support them, and they have to be explicitly enabled.

ANON_DH:

This algorithm exchanges Diffie-Hellman parameters.

ANON_ECDH:

This algorithm exchanges elliptic curve Diffie-Hellman parameters. It is more efficient than ANON_DH on equivalent security levels.

4.4 Selecting an appropriate authentication method

This section provides some guidance on how to use the available authentication methods in GnuTLS in various scenarios.

4.4.1 Two peers with an out-of-band channel

Let’s consider two peers who need to communicate over an untrusted channel (the Internet), but have an out-of-band channel available. The latter channel is considered safe from eavesdropping and message modification and thus can be used for an initial bootstrapping of the protocol. The options available are:

• Pre-shared keys (see PSK authentication). The server and a client communicate a shared randomly generated key over the trusted channel and use it to negotiate further sessions over the untrusted channel.
• Passwords (see SRP authentication). The client communicates to the server its username and password of choice and uses it to negotiate further sessions over the untrusted channel.
• Public keys (see Certificate authentication). The client and the server exchange their public keys (or fingerprints of them) over the trusted channel. On future sessions over the untrusted channel they verify the key being the same (similar to Verifying a certificate using trust on first use authentication).

Provided that the out-of-band channel is trusted all of the above provide a similar level of protection. An out-of-band channel may be the initial bootstrapping of a user’s PC in a corporate environment, in-person communication, communication over an alternative network (e.g. the phone network), etc.

4.4.2 Two peers without an out-of-band channel

When an out-of-band channel is not available a peer cannot be reliably authenticated. What can be done, however, is to allow some form of registration of users connecting for the first time and ensure that their keys remain the same after that initial connection. This is termed key continuity or trust on first use (TOFU).

The available option is to use public key authentication (see Certificate authentication). The client and the server store each other’s public keys (or fingerprints of them) and associate them with their identity. On future sessions over the untrusted channel they verify the keys being the same (see Verifying a certificate using trust on first use authentication).

To mitigate the uncertainty of the information exchanged in the first connection other channels over the Internet may be used, e.g., DNSSEC (see Verifying a certificate using DANE).

4.4.3 Two peers and a trusted third party

When a trusted third party is available (or a certificate authority) the most suitable option is to use certificate authentication (see Certificate authentication). The client and the server obtain certificates that associate their identity and public keys using a digital signature by the trusted party and use them to on the subsequent communications with each other. Each party verifies the peer’s certificate using the trusted third party’s signature. The parameters of the third party’s signature are present in its certificate which must be available to all communicating parties.

While the above is the typical authentication method for servers in the Internet by using the commercial CAs, the users that act as clients in the protocol rarely possess such certificates. In that case a hybrid method can be used where the server is authenticated by the client using the commercial CAs and the client is authenticated based on some information the client provided over the initial server-authenticated channel. The available options are:

• Passwords (see SRP authentication). The client communicates to the server its username and password of choice on the initial server-authenticated connection and uses it to negotiate further sessions. This is possible because the SRP protocol allows for the server to be authenticated using a certificate and the client using the password.
• Public keys (see Certificate authentication). The client sends its public key to the server (or a fingerprint of it) over the initial server-authenticated connection. On future sessions the client verifies the server using the third party certificate and the server verifies that the client’s public key remained the same (see Verifying a certificate using trust on first use authentication).

5 Abstract key types and Hardware security modules

In several cases storing the long term cryptographic keys in a hard disk or even in memory poses a significant risk. Once the system they are stored is compromised the keys must be replaced as the secrecy of future sessions is no longer guaranteed. Moreover, past sessions that were not protected by a perfect forward secrecy offering ciphersuite are also to be assumed compromised.

If such threats need to be addressed, then it may be wise storing the keys in a security module such as a smart card, an HSM or the TPM chip. Those modules ensure the protection of the cryptographic keys by only allowing operations on them and preventing their extraction. The purpose of the abstract key API is to provide an API that will allow the handle of keys in memory and files, as well as keys stored in such modules.

In GnuTLS the approach is to handle all keys transparently by the high level API, e.g., the API that loads a key or certificate from a file. The high-level API will accept URIs in addition to files that specify keys on an HSM or in TPM, and a callback function will be used to obtain any required keys. The URI format is defined in [PKCS11URI].

More information on the API is provided in the next sections. Examples of a URI of a certificate stored in an HSM, as well as a key stored in the TPM chip are shown below. To discover the URIs of the objects the p11tool (see p11tool Invocation).

pkcs11:token=Nikos;serial=307521161601031;model=PKCS%2315; \
manufacturer=EnterSafe;object=test1;type=cert

5.1 Abstract key types

Since there are many forms of a public or private keys supported by GnuTLS such as X.509, OpenPGP, PKCS #11 or TPM it is desirable to allow common operations on them. For these reasons the abstract gnutls_privkey_t and gnutls_pubkey_t were introduced in gnutls/abstract.h header. Those types are initialized using a specific type of key and then can be used to perform operations in an abstract way. For example in order to sign an X.509 certificate with a key that resides in a token the following steps can be used.

#inlude <gnutls/abstract.h>

void sign_cert( gnutls_x509_crt_t to_be_signed)
{
gnutls_x509_crt_t ca_cert;
gnutls_privkey_t abs_key;

  /* initialize the abstract key */
  gnutls_privkey_init(&abs_key);

  /* keys stored in tokens are identified by URLs */
  gnutls_privkey_import_url(abs_key, key_url);

  gnutls_x509_crt_init(&ca_cert);
  gnutls_x509_crt_import_url(&ca_cert, cert_url);

  /* sign the certificate to be signed */
  gnutls_x509_crt_privkey_sign(to_be_signed, ca_cert, abs_key, 
                               GNUTLS_DIG_SHA256, 0);
}

5.1.1 Public keys

An abstract gnutls_pubkey_t can be initialized using the functions below. It can be imported through an existing structure like gnutls_x509_crt_t, or through an ASN.1 encoding of the X.509 SubjectPublicKeyInfo sequence.

int gnutls_pubkey_import_x509 (gnutls_pubkey_t key, gnutls_x509_crt_t crt, unsigned int flags)

int gnutls_pubkey_import_pkcs11 (gnutls_pubkey_t key, gnutls_pkcs11_obj_t obj, unsigned int flags)

int gnutls_pubkey_import_url (gnutls_pubkey_t key, const char * url, unsigned int flags)

int gnutls_pubkey_import_privkey (gnutls_pubkey_t key, gnutls_privkey_t pkey, unsigned int usage, unsigned int flags)

int gnutls_pubkey_import (gnutls_pubkey_t key, const gnutls_datum_t * data, gnutls_x509_crt_fmt_t format)

int gnutls_pubkey_export (gnutls_pubkey_t key, gnutls_x509_crt_fmt_t format, void * output_data, size_t * output_data_size)

Function: int gnutls_pubkey_export2 (gnutls_pubkey_t key, gnutls_x509_crt_fmt_t format, gnutls_datum_t * out)

key: Holds the certificate

format: the format of output params. One of PEM or DER.

out: will contain a certificate PEM or DER encoded

This function will export the public key to DER or PEM format. The contents of the exported data is the SubjectPublicKeyInfo X.509 structure.

The output buffer will be allocated using gnutls_malloc() .

If the structure is PEM encoded, it will have a header of "BEGIN CERTIFICATE".

Returns: In case of failure a negative error code will be returned, and 0 on success.

Since: 3.1.3

Other helper functions that allow directly importing from raw X.509 or OpenPGP structures are shown below.

int gnutls_pubkey_import_x509_raw (gnutls_pubkey_t pkey, const gnutls_datum_t * data, gnutls_x509_crt_fmt_t format, unsigned int flags)

An important function is gnutls_pubkey_import_url which will import public keys from URLs that identify objects stored in tokens (see Smart cards and HSMs and Trusted Platform Module). A function to check for a supported by GnuTLS URL is gnutls_url_is_supported.

Function: unsigned gnutls_url_is_supported (const char * url)

url: A URI to be tested

Check whether the provided url is supported. Depending on the system libraries GnuTLS may support pkcs11, tpmkey or other URLs.

Returns: return non-zero if the given URL is supported, and zero if it is not known.

Since: 3.1.0

Additional functions are available that will return information over a public key, such as a unique key ID, as well as a function that given a public key fingerprint would provide a memorable sketch.

Note that gnutls_pubkey_get_key_id calculates a SHA1 digest of the public key as a DER-formatted, subjectPublicKeyInfo object. Other implementations use different approaches, e.g., some use the “common method” described in section 4.2.1.2 of [RFC5280] which calculates a digest on a part of the subjectPublicKeyInfo object.

int gnutls_pubkey_get_pk_algorithm (gnutls_pubkey_t key, unsigned int * bits)

int gnutls_pubkey_get_preferred_hash_algorithm (gnutls_pubkey_t key, gnutls_digest_algorithm_t * hash, unsigned int * mand)

int gnutls_pubkey_get_key_id (gnutls_pubkey_t key, unsigned int flags, unsigned char * output_data, size_t * output_data_size)

int gnutls_random_art (gnutls_random_art_t type, const char * key_type, unsigned int key_size, void * fpr, size_t fpr_size, gnutls_datum_t * art)

To export the key-specific parameters, or obtain a unique key ID the following functions are provided.

int gnutls_pubkey_export_rsa_raw (gnutls_pubkey_t key, gnutls_datum_t * m, gnutls_datum_t * e)

int gnutls_pubkey_export_dsa_raw (gnutls_pubkey_t key, gnutls_datum_t * p, gnutls_datum_t * q, gnutls_datum_t * g, gnutls_datum_t * y)

int gnutls_pubkey_export_ecc_raw (gnutls_pubkey_t key, gnutls_ecc_curve_t * curve, gnutls_datum_t * x, gnutls_datum_t * y)

int gnutls_pubkey_export_ecc_x962 (gnutls_pubkey_t key, gnutls_datum_t * parameters, gnutls_datum_t * ecpoint)

5.1.2 Private keys

An abstract gnutls_privkey_t can be initialized using the functions below. It can be imported through an existing structure like gnutls_x509_privkey_t, but unlike public keys it cannot be exported. That is to allow abstraction over keys stored in hardware that makes available only operations.

int gnutls_privkey_import_x509 (gnutls_privkey_t pkey, gnutls_x509_privkey_t key, unsigned int flags)

int gnutls_privkey_import_pkcs11 (gnutls_privkey_t pkey, gnutls_pkcs11_privkey_t key, unsigned int flags)

Other helper functions that allow directly importing from raw X.509 or OpenPGP structures are shown below. Again, as with public keys, private keys can be imported from a hardware module using URLs.

Function: int gnutls_privkey_import_url (gnutls_privkey_t key, const char * url, unsigned int flags)

key: A key of type gnutls_privkey_t

url: A PKCS 11 url

flags: should be zero

This function will import a PKCS11 or TPM URL as a private key. The supported URL types can be checked using gnutls_url_is_supported() .

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since: 3.1.0

int gnutls_privkey_import_x509_raw (gnutls_privkey_t pkey, const gnutls_datum_t * data, gnutls_x509_crt_fmt_t format, const char * password, unsigned int flags)

int gnutls_privkey_get_pk_algorithm (gnutls_privkey_t key, unsigned int * bits)

gnutls_privkey_type_t gnutls_privkey_get_type (gnutls_privkey_t key)

int gnutls_privkey_status (gnutls_privkey_t key)

In order to support cryptographic operations using an external API, the following function is provided. This allows for a simple extensibility API without resorting to PKCS #11.

Function: int gnutls_privkey_import_ext3 (gnutls_privkey_t pkey, void * userdata, gnutls_privkey_sign_func sign_fn, gnutls_privkey_decrypt_func decrypt_fn, gnutls_privkey_deinit_func deinit_fn, gnutls_privkey_info_func info_fn, unsigned int flags)

pkey: The private key

userdata: private data to be provided to the callbacks

sign_fn: callback for signature operations

decrypt_fn: callback for decryption operations

deinit_fn: a deinitialization function

info_fn: returns info about the public key algorithm (should not be NULL )

flags: Flags for the import

This function will associate the given callbacks with the gnutls_privkey_t type. At least one of the two callbacks must be non-null. If a deinitialization function is provided then flags is assumed to contain GNUTLS_PRIVKEY_IMPORT_AUTO_RELEASE .

Note that the signing function is supposed to "raw" sign data, i.e., without any hashing or preprocessing. In case of RSA the DigestInfo will be provided, and the signing function is expected to do the PKCS 1 1.5 padding and the exponentiation.

The info_fn must provide information on the algorithms supported by this private key, and should support the flags GNUTLS_PRIVKEY_INFO_PK_ALGO and GNUTLS_PRIVKEY_INFO_SIGN_ALGO . It must return -1 on unknown flags.

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since: 3.4.0

5.1.3 Operations

The abstract key types can be used to access signing and signature verification operations with the underlying keys.

Function: int gnutls_pubkey_verify_data2 (gnutls_pubkey_t pubkey, gnutls_sign_algorithm_t algo, unsigned int flags, const gnutls_datum_t * data, const gnutls_datum_t * signature)

pubkey: Holds the public key

algo: The signature algorithm used

flags: Zero or an OR list of gnutls_certificate_verify_flags

data: holds the signed data

signature: contains the signature

This function will verify the given signed data, using the parameters from the certificate.

Returns: In case of a verification failure GNUTLS_E_PK_SIG_VERIFY_FAILED is returned, and zero or positive code on success. For known to be insecure signatures this function will return GNUTLS_E_INSUFFICIENT_SECURITY unless the flag GNUTLS_VERIFY_ALLOW_BROKEN is specified.

Since: 3.0

Function: int gnutls_pubkey_verify_hash2 (gnutls_pubkey_t key, gnutls_sign_algorithm_t algo, unsigned int flags, const gnutls_datum_t * hash, const gnutls_datum_t * signature)

key: Holds the public key

algo: The signature algorithm used

flags: Zero or an OR list of gnutls_certificate_verify_flags

hash: holds the hash digest to be verified

signature: contains the signature

This function will verify the given signed digest, using the parameters from the public key. Note that unlike gnutls_privkey_sign_hash() , this function accepts a signature algorithm instead of a digest algorithm. You can use gnutls_pk_to_sign() to get the appropriate value.

Returns: In case of a verification failure GNUTLS_E_PK_SIG_VERIFY_FAILED is returned, and zero or positive code on success.

Since: 3.0

Function: int gnutls_pubkey_encrypt_data (gnutls_pubkey_t key, unsigned int flags, const gnutls_datum_t * plaintext, gnutls_datum_t * ciphertext)

key: Holds the public key

flags: should be 0 for now

plaintext: The data to be encrypted

ciphertext: contains the encrypted data

This function will encrypt the given data, using the public key. On success the ciphertext will be allocated using gnutls_malloc() .

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since: 3.0

Function: int gnutls_privkey_sign_data (gnutls_privkey_t signer, gnutls_digest_algorithm_t hash, unsigned int flags, const gnutls_datum_t * data, gnutls_datum_t * signature)

signer: Holds the key

hash: should be a digest algorithm

flags: Zero or one of gnutls_privkey_flags_t

data: holds the data to be signed

signature: will contain the signature allocated with gnutls_malloc()

This function will sign the given data using a signature algorithm supported by the private key. Signature algorithms are always used together with a hash functions. Different hash functions may be used for the RSA algorithm, but only the SHA family for the DSA keys.

You may use gnutls_pubkey_get_preferred_hash_algorithm() to determine the hash algorithm.

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since: 2.12.0

Function: int gnutls_privkey_sign_hash (gnutls_privkey_t signer, gnutls_digest_algorithm_t hash_algo, unsigned int flags, const gnutls_datum_t * hash_data, gnutls_datum_t * signature)

signer: Holds the signer’s key

hash_algo: The hash algorithm used

flags: Zero or one of gnutls_privkey_flags_t

hash_data: holds the data to be signed

signature: will contain newly allocated signature

This function will sign the given hashed data using a signature algorithm supported by the private key. Signature algorithms are always used together with a hash functions. Different hash functions may be used for the RSA algorithm, but only SHA-XXX for the DSA keys.

You may use gnutls_pubkey_get_preferred_hash_algorithm() to determine the hash algorithm.

Note that if GNUTLS_PRIVKEY_SIGN_FLAG_TLS1_RSA flag is specified this function will ignore hash_algo and perform a raw PKCS1 signature.

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since: 2.12.0

Function: int gnutls_privkey_decrypt_data (gnutls_privkey_t key, unsigned int flags, const gnutls_datum_t * ciphertext, gnutls_datum_t * plaintext)

key: Holds the key

flags: zero for now

ciphertext: holds the data to be decrypted

plaintext: will contain the decrypted data, allocated with gnutls_malloc()

This function will decrypt the given data using the algorithm supported by the private key.

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since: 2.12.0

Signing existing structures, such as certificates, CRLs, or certificate requests, as well as associating public keys with structures is also possible using the key abstractions.

Function: int gnutls_x509_crq_set_pubkey (gnutls_x509_crq_t crq, gnutls_pubkey_t key)

crq: should contain a gnutls_x509_crq_t type

key: holds a public key

This function will set the public parameters from the given public key to the request. The key can be deallocated after that.

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since: 2.12.0

Function: int gnutls_x509_crt_set_pubkey (gnutls_x509_crt_t crt, gnutls_pubkey_t key)

crt: should contain a gnutls_x509_crt_t type

key: holds a public key

This function will set the public parameters from the given public key to the certificate. The key can be deallocated after that.

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since: 2.12.0

int gnutls_x509_crt_privkey_sign (gnutls_x509_crt_t crt, gnutls_x509_crt_t issuer, gnutls_privkey_t issuer_key, gnutls_digest_algorithm_t dig, unsigned int flags)

int gnutls_x509_crl_privkey_sign (gnutls_x509_crl_t crl, gnutls_x509_crt_t issuer, gnutls_privkey_t issuer_key, gnutls_digest_algorithm_t dig, unsigned int flags)

int gnutls_x509_crq_privkey_sign (gnutls_x509_crq_t crq, gnutls_privkey_t key, gnutls_digest_algorithm_t dig, unsigned int flags)

5.2 System and application-specific keys

5.2.1 System-specific keys

In several systems there are keystores which allow to read, store and use certificates and private keys. For these systems GnuTLS provides the system-key API in gnutls/system-keys.h. That API provides the ability to iterate through all stored keys, add and delete keys as well as use these keys using a URL which starts with "system:". The format of the URLs is system-specific. The systemkey tool is also provided to assist in listing keys and debugging.

The systems supported via this API are the following.

• Windows Cryptography API (CNG)

Function: int gnutls_system_key_iter_get_info (gnutls_system_key_iter_t * iter, unsigned cert_type, char ** cert_url, char ** key_url, char ** label, gnutls_datum_t * der, unsigned int flags)

iter: an iterator of the system keys (must be set to NULL initially)

cert_type: A value of gnutls_certificate_type_t which indicates the type of certificate to look for

cert_url: The certificate URL of the pair (may be NULL )

key_url: The key URL of the pair (may be NULL )

label: The friendly name (if any) of the pair (may be NULL )

der: if non-NULL the DER data of the certificate

flags: should be zero

This function will return on each call a certificate and key pair URLs, as well as a label associated with them, and the DER-encoded certificate. When the iteration is complete it will return GNUTLS_E_REQUESTED_DATA_NOT_AVAILABLE .

Typically cert_type should be GNUTLS_CRT_X509 .

All values set are allocated and must be cleared using gnutls_free() ,

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since: 3.4.0

void gnutls_system_key_iter_deinit (gnutls_system_key_iter_t iter)

int gnutls_system_key_add_x509 (gnutls_x509_crt_t crt, gnutls_x509_privkey_t privkey, const char * label, char ** cert_url, char ** key_url)

int gnutls_system_key_delete (const char * cert_url, const char * key_url)

5.2.2 Application-specific keys

For systems where GnuTLS doesn’t provide a system specific store, it may often be desirable to define a custom class of keys that are identified via URLs and available to GnuTLS calls such as gnutls_certificate_set_x509_key_file2. Such keys can be registered using the API in gnutls/urls.h. The function which registers such keys is gnutls_register_custom_url.

Function: int gnutls_register_custom_url (const gnutls_custom_url_st * st)

st: A gnutls_custom_url_st structure

Register a custom URL. This will affect the following functions: gnutls_url_is_supported() , gnutls_privkey_import_url() , gnutls_pubkey_import_url, gnutls_x509_crt_import_url() and all functions that depend on them, e.g., gnutls_certificate_set_x509_key_file2() .

The provided structure and callback functions must be valid throughout the lifetime of the process. The registration of an existing URL type will fail with GNUTLS_E_INVALID_REQUEST . Since GnuTLS 3.5.0 this function can be used to override the builtin URLs.

This function is not thread safe.

Returns: returns zero if the given structure was imported or a negative value otherwise.

Since: 3.4.0

The input to this function are three callback functions as well as the prefix of the URL, (e.g., "mypkcs11:") and the length of the prefix. The types of the callbacks are shown below, and are expected to use the exported gnutls functions to import the keys and certificates. E.g., a typical import_key callback should use gnutls_privkey_import_ext3.

typedef int (*gnutls_privkey_import_url_func)(gnutls_privkey_t pkey,
                                              const char *url,
                                              unsigned flags);

typedef int (*gnutls_x509_crt_import_url_func)(gnutls_x509_crt_t pkey,
                                               const char *url,
                                               unsigned flags);

/* The following callbacks are optional */

/* This is to enable gnutls_pubkey_import_url() */
typedef int (*gnutls_pubkey_import_url_func)(gnutls_pubkey_t pkey,
					     const char *url, unsigned flags);

/* This is to allow constructing a certificate chain. It will be provided
 * the initial certificate URL and the certificate to find its issuer, and must
 * return zero and the DER encoding of the issuer's certificate. If not available,
 * it should return GNUTLS_E_REQUESTED_DATA_NOT_AVAILABLE. */
typedef int (*gnutls_get_raw_issuer_func)(const char *url, gnutls_x509_crt_t crt,
					  gnutls_datum_t *issuer_der, unsigned flags);

typedef struct custom_url_st {
        const char *name;
        unsigned name_size;
        gnutls_privkey_import_url_func import_key;
        gnutls_x509_crt_import_url_func import_crt;
        gnutls_pubkey_import_url_func import_pubkey;
	gnutls_get_raw_issuer_func get_issuer;
} gnutls_custom_url_st;

5.3 Smart cards and HSMs

In this section we present the smart-card and hardware security module (HSM) support in GnuTLS using PKCS #11 [PKCS11]. Hardware security modules and smart cards provide a way to store private keys and perform operations on them without exposing them. This decouples cryptographic keys from the applications that use them and provide an additional security layer against cryptographic key extraction. Since this can also be achieved in software components such as in Gnome keyring, we will use the term security module to describe any cryptographic key separation subsystem.

PKCS #11 is plugin API allowing applications to access cryptographic operations on a security module, as well as to objects residing on it. PKCS #11 modules exist for hardware tokens such as smart cards⁹, cryptographic tokens, as well as for software modules like Gnome Keyring. The objects residing on a security module may be certificates, public keys, private keys or secret keys. Of those certificates and public/private key pairs can be used with GnuTLS. PKCS #11’s main advantage is that it allows operations on private key objects such as decryption and signing without exposing the key. In GnuTLS the PKCS #11 functionality is available in gnutls/pkcs11.h.

Figure 5.1: PKCS #11 module usage.

5.3.1 Initialization

To allow all GnuTLS applications to transparently access smart cards and tokens, PKCS #11 is automatically initialized during the first call of a PKCS #11 related function, in a thread safe way. The default initialization process, utilizes p11-kit configuration, and loads any appropriate PKCS #11 modules. The p11-kit configuration files¹⁰ are typically stored in /etc/pkcs11/modules/. For example a file that will instruct GnuTLS to load the OpenSC module, could be named /etc/pkcs11/modules/opensc.module and contain the following:

module: /usr/lib/opensc-pkcs11.so

If you use these configuration files, then there is no need for other initialization in GnuTLS, except for the PIN and token callbacks (see next section). In several cases, however, it is desirable to limit badly behaving modules (e.g., modules that add an unacceptable delay on initialization) to single applications. That can be done using the “enable-in:” option followed by the base name of applications that this module should be used.

It is also possible to manually initialize or even disable the PKCS #11 subsystem if the default settings are not desirable or not available (see PKCS11 Manual Initialization for more information).

Note that, PKCS #11 modules behave in a peculiar way after a fork; they require a reinitialization of all the used PKCS #11 resources. While GnuTLS automates that process, there are corner cases where it is not possible to handle it correctly in an automated way¹¹. For that, it is recommended not to mix fork() and PKCS #11 module usage. It is recommended to initialize and use any PKCS #11 resources in a single process.

Older versions of GnuTLS required to call gnutls_pkcs11_reinit after a fork() call; since 3.3.0 this is no longer required.

5.3.2 Manual initialization of user-specific modules

In systems where one cannot rely on a globally available p11-kit configuration to be available, it is still possible to utilize PKCS #11 objects. That can be done by loading directly the PKCS #11 shared module in the application using gnutls_pkcs11_add_provider, after having called gnutls_pkcs11_init specifying the GNUTLS_PKCS11_FLAG_MANUAL flag.

Function: int gnutls_pkcs11_add_provider (const char * name, const char * params)

name: The filename of the module

params: should be NULL or a known string (see description)

This function will load and add a PKCS 11 module to the module list used in gnutls. After this function is called the module will be used for PKCS 11 operations.

When loading a module to be used for certificate verification, use the string ’trusted’ as params .

Note that this function is not thread safe.

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since: 2.12.0

In that case, the application will only have access to the modules explictly loaded. If the GNUTLS_PKCS11_FLAG_MANUAL flag is specified and no calls to gnutls_pkcs11_add_provider are made, then the PKCS #11 functionality is effectively disabled.

Function: int gnutls_pkcs11_init (unsigned int flags, const char * deprecated_config_file)

flags: An ORed sequence of GNUTLS_PKCS11_FLAG_ *

deprecated_config_file: either NULL or the location of a deprecated configuration file

This function will initialize the PKCS 11 subsystem in gnutls. It will read configuration files if GNUTLS_PKCS11_FLAG_AUTO is used or allow you to independently load PKCS 11 modules using gnutls_pkcs11_add_provider() if GNUTLS_PKCS11_FLAG_MANUAL is specified.

You don’t need to call this function since GnuTLS 3.3.0 because it is being called during the first request PKCS 11 operation. That call will assume the GNUTLS_PKCS11_FLAG_AUTO flag. If another flags are required then it must be called independently prior to any PKCS 11 operation.

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since: 2.12.0

5.3.3 Accessing objects that require a PIN

Objects stored in token such as a private keys are typically protected from access by a PIN or password. This PIN may be required to either read the object (if allowed) or to perform operations with it. To allow obtaining the PIN when accessing a protected object, as well as probe the user to insert the token the following functions allow to set a callback.

void gnutls_pkcs11_set_token_function (gnutls_pkcs11_token_callback_t fn, void * userdata)

void gnutls_pkcs11_set_pin_function (gnutls_pin_callback_t fn, void * userdata)

int gnutls_pkcs11_add_provider (const char * name, const char * params)

gnutls_pin_callback_t gnutls_pkcs11_get_pin_function (void ** userdata)

The callback is of type gnutls_pin_callback_t and will have as input the provided userdata, the PIN attempt number, a URL describing the token, a label describing the object and flags. The PIN must be at most of pin_max size and must be copied to pin variable. The function must return 0 on success or a negative error code otherwise.

typedef int (*gnutls_pin_callback_t) (void *userdata, int attempt,
                                      const char *token_url,
                                      const char *token_label,
                                      unsigned int flags,
                                      char *pin, size_t pin_max);

The flags are of gnutls_pin_flag_t type and are explained below.

Note that due to limitations of PKCS #11 there are issues when multiple libraries are sharing a module. To avoid this problem GnuTLS uses p11-kit that provides a middleware to control access to resources over the multiple users.

To avoid conflicts with multiple registered callbacks for PIN functions, gnutls_pkcs11_get_pin_function may be used to check for any previously set functions. In addition context specific PIN functions are allowed, e.g., by using functions below.

void gnutls_certificate_set_pin_function (gnutls_certificate_credentials_t cred, gnutls_pin_callback_t fn, void * userdata)

void gnutls_pubkey_set_pin_function (gnutls_pubkey_t key, gnutls_pin_callback_t fn, void * userdata)

void gnutls_privkey_set_pin_function (gnutls_privkey_t key, gnutls_pin_callback_t fn, void * userdata)

void gnutls_pkcs11_obj_set_pin_function (gnutls_pkcs11_obj_t obj, gnutls_pin_callback_t fn, void * userdata)

void gnutls_x509_crt_set_pin_function (gnutls_x509_crt_t crt, gnutls_pin_callback_t fn, void * userdata)

5.3.4 Reading objects

All PKCS #11 objects are referenced by GnuTLS functions by URLs as described in [PKCS11URI]. This allows for a consistent naming of objects across systems and applications in the same system. For example a public key on a smart card may be referenced as:

pkcs11:token=Nikos;serial=307521161601031;model=PKCS%2315; \
manufacturer=EnterSafe;object=test1;type=public;\
id=32f153f3e37990b08624141077ca5dec2d15faed

while the smart card itself can be referenced as:

pkcs11:token=Nikos;serial=307521161601031;model=PKCS%2315;manufacturer=EnterSafe

Objects stored in a PKCS #11 token can typically be extracted if they are not marked as sensitive. Usually only private keys are marked as sensitive and cannot be extracted, while certificates and other data can be retrieved. The functions that can be used to enumerate and access objects are shown below.

int gnutls_pkcs11_obj_list_import_url4 (gnutls_pkcs11_obj_t ** p_list, unsigned int * n_list, const char * url, unsigned int flags)

int gnutls_pkcs11_obj_import_url (gnutls_pkcs11_obj_t obj, const char * url, unsigned int flags)

int gnutls_pkcs11_obj_export_url (gnutls_pkcs11_obj_t obj, gnutls_pkcs11_url_type_t detailed, char ** url)

Function: int gnutls_pkcs11_obj_get_info (gnutls_pkcs11_obj_t obj, gnutls_pkcs11_obj_info_t itype, void * output, size_t * output_size)

obj: should contain a gnutls_pkcs11_obj_t type

itype: Denotes the type of information requested

output: where output will be stored

output_size: contains the maximum size of the output and will be overwritten with actual

This function will return information about the PKCS11 certificate such as the label, id as well as token information where the key is stored. When output is text it returns null terminated string although output_size contains the size of the actual data only.

Returns: GNUTLS_E_SUCCESS (0) on success or a negative error code on error.

Since: 2.12.0

int gnutls_x509_crt_import_pkcs11 (gnutls_x509_crt_t crt, gnutls_pkcs11_obj_t pkcs11_crt)

int gnutls_x509_crt_import_url (gnutls_x509_crt_t crt, const char * url, unsigned int flags)

int gnutls_x509_crt_list_import_pkcs11 (gnutls_x509_crt_t * certs, unsigned int cert_max, gnutls_pkcs11_obj_t * const objs, unsigned int flags)

Properties of the physical token can also be accessed and altered with GnuTLS. For example data in a token can be erased (initialized), PIN can be altered, etc.

int gnutls_pkcs11_token_init (const char * token_url, const char * so_pin, const char * label)

int gnutls_pkcs11_token_get_url (unsigned int seq, gnutls_pkcs11_url_type_t detailed, char ** url)

int gnutls_pkcs11_token_get_info (const char * url, gnutls_pkcs11_token_info_t ttype, void * output, size_t * output_size)

int gnutls_pkcs11_token_get_flags (const char * url, unsigned int * flags)

int gnutls_pkcs11_token_set_pin (const char * token_url, const char * oldpin, const char * newpin, unsigned int flags)

The following examples demonstrate the usage of the API. The first example will list all available PKCS #11 tokens in a system and the latter will list all certificates in a token that have a corresponding private key.

int i;
char* url;

gnutls_global_init();

for (i=0;;i++) 
  {
    ret = gnutls_pkcs11_token_get_url(i, &url);
    if (ret == GNUTLS_E_REQUESTED_DATA_NOT_AVAILABLE)
      break;

    if (ret < 0)
      exit(1);
		
    fprintf(stdout, "Token[%d]: URL: %s\n", i, url);
    gnutls_free(url);
  }
gnutls_global_deinit();

/* This example code is placed in the public domain. */

#include <config.h>
#include <gnutls/gnutls.h>
#include <gnutls/pkcs11.h>
#include <stdio.h>
#include <stdlib.h>

#define URL "pkcs11:URL"

int main(int argc, char **argv)
{
        gnutls_pkcs11_obj_t *obj_list;
        gnutls_x509_crt_t xcrt;
        unsigned int obj_list_size = 0;
        gnutls_datum_t cinfo;
        int ret;
        unsigned int i;

        ret = gnutls_pkcs11_obj_list_import_url4(&obj_list, &obj_list_size, URL,
                                                GNUTLS_PKCS11_OBJ_FLAG_CRT|
                                                GNUTLS_PKCS11_OBJ_FLAG_WITH_PRIVKEY);
        if (ret < 0)
                return -1;

        /* now all certificates are in obj_list */
        for (i = 0; i < obj_list_size; i++) {

                gnutls_x509_crt_init(&xcrt);

                gnutls_x509_crt_import_pkcs11(xcrt, obj_list[i]);

                gnutls_x509_crt_print(xcrt, GNUTLS_CRT_PRINT_FULL, &cinfo);

                fprintf(stdout, "cert[%d]:\n %s\n\n", i, cinfo.data);

                gnutls_free(cinfo.data);
                gnutls_x509_crt_deinit(xcrt);
        }

        for (i = 0; i < obj_list_size; i++)
                gnutls_pkcs11_obj_deinit(obj_list[i]);
	gnutls_free(obj_list);

        return 0;
}

5.3.5 Writing objects

With GnuTLS you can copy existing private keys and certificates to a token. Note that when copying private keys it is recommended to mark them as sensitive using the GNUTLS_PKCS11_OBJ_FLAG_MARK_SENSITIVE to prevent its extraction. An object can be marked as private using the flag GNUTLS_PKCS11_OBJ_FLAG_MARK_PRIVATE, to require PIN to be entered before accessing the object (for operations or otherwise).

Function: int gnutls_pkcs11_copy_x509_privkey2 (const char * token_url, gnutls_x509_privkey_t key, const char * label, const gnutls_datum_t * cid, unsigned int key_usage, unsigned int flags)

token_url: A PKCS 11 URL specifying a token

key: A private key

label: A name to be used for the stored data

cid: The CKA_ID to set for the object -if NULL, the ID will be derived from the public key

key_usage: One of GNUTLS_KEY_*

flags: One of GNUTLS_PKCS11_OBJ_* flags

This function will copy a private key into a PKCS 11 token specified by a URL. It is highly recommended flags to contain GNUTLS_PKCS11_OBJ_FLAG_MARK_SENSITIVE unless there is a strong reason not to.

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since: 3.4.0

Function: int gnutls_pkcs11_copy_x509_crt2 (const char * token_url, gnutls_x509_crt_t crt, const char * label, const gnutls_datum_t * cid, unsigned int flags)

token_url: A PKCS 11 URL specifying a token

crt: The certificate to copy

label: The name to be used for the stored data

cid: The CKA_ID to set for the object -if NULL, the ID will be derived from the public key

flags: One of GNUTLS_PKCS11_OBJ_FLAG_*

This function will copy a certificate into a PKCS 11 token specified by a URL. Valid flags to mark the certificate: GNUTLS_PKCS11_OBJ_FLAG_MARK_TRUSTED , GNUTLS_PKCS11_OBJ_FLAG_MARK_SENSITIVE , GNUTLS_PKCS11_OBJ_FLAG_MARK_PRIVATE , GNUTLS_PKCS11_OBJ_FLAG_MARK_CA , GNUTLS_PKCS11_OBJ_FLAG_MARK_ALWAYS_AUTH .

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since: 3.4.0

Function: int gnutls_pkcs11_delete_url (const char * object_url, unsigned int flags)

object_url: The URL of the object to delete.

flags: One of GNUTLS_PKCS11_OBJ_* flags

This function will delete objects matching the given URL. Note that not all tokens support the delete operation.

Returns: On success, the number of objects deleted is returned, otherwise a negative error value.

Since: 2.12.0

5.3.6 Using a PKCS #11 token with TLS

It is possible to use a PKCS #11 token to a TLS session, as shown in ex-pkcs11-client. In addition the following functions can be used to load PKCS #11 key and certificates by specifying a PKCS #11 URL instead of a filename.

int gnutls_certificate_set_x509_trust_file (gnutls_certificate_credentials_t cred, const char * cafile, gnutls_x509_crt_fmt_t type)

int gnutls_certificate_set_x509_key_file2 (gnutls_certificate_credentials_t res, const char * certfile, const char * keyfile, gnutls_x509_crt_fmt_t type, const char * pass, unsigned int flags)

5.3.7 Verifying certificates over PKCS #11

The PKCS #11 API can be used to allow all applications in the same operating system to access shared cryptographic keys and certificates in a uniform way, as in Figure 5.1. That way applications could load their trusted certificate list, as well as user certificates from a common PKCS #11 module. Such a provider is the p11-kit trust storage module¹² and it provides access to the trusted Root CA certificates in a system. That provides a more dynamic list of Root CA certificates, as opposed to a static list in a file or directory.

That store, allows for blacklisting of CAs or certificates, as well as categorization of the Root CAs (Web verification, Code signing, etc.), in addition to restricting their purpose via stapled extensions¹³. GnuTLS will utilize the p11-kit trust module as the default trust store if configured to; i.e., if ’–with-default-trust-store-pkcs11=pkcs11:’ is given to the configure script.

5.3.8 Invoking p11tool

Program that allows operations on PKCS #11 smart cards and security modules.

To use PKCS #11 tokens with GnuTLS the p11-kit configuration files need to be setup. That is create a .module file in /etc/pkcs11/modules with the contents ’module: /path/to/pkcs11.so’. Alternatively the configuration file /etc/gnutls/pkcs11.conf has to exist and contain a number of lines of the form ’load=/usr/lib/opensc-pkcs11.so’.

You can provide the PIN to be used for the PKCS #11 operations with the environment variables GNUTLS_PIN and GNUTLS_SO_PIN.

This section was generated by AutoGen, using the agtexi-cmd template and the option descriptions for the p11tool program. This software is released under the GNU General Public License, version 3 or later.

p11tool help/usage (--help)

This is the automatically generated usage text for p11tool.

The text printed is the same whether selected with the help option (--help) or the more-help option (--more-help). more-help will print the usage text by passing it through a pager program. more-help is disabled on platforms without a working fork(2) function. The PAGER environment variable is used to select the program, defaulting to more. Both will exit with a status code of 0.

p11tool - GnuTLS PKCS #11 tool
Usage:  p11tool [ -<flag> [<val>] | --<name>[{=| }<val>] ]... [url]


Tokens:

       --list-tokens          List all available tokens
       --list-token-urls      List the URLs available tokens
       --list-mechanisms      List all available mechanisms in a token
       --initialize           Initializes a PKCS #11 token
       --initialize-pin       Initializes/Resets a PKCS #11 token user PIN
       --initialize-so-pin    Initializes/Resets a PKCS #11 token security officer PIN
       --set-pin=str          Specify the PIN to use on token operations
       --set-so-pin=str       Specify the Security Officer's PIN to use on token initialization

Object listing:

       --list-all             List all available objects in a token
       --list-all-certs       List all available certificates in a token
       --list-certs           List all certificates that have an associated private key
       --list-all-privkeys    List all available private keys in a token
       --list-privkeys        an alias for the 'list-all-privkeys' option
       --list-keys            an alias for the 'list-all-privkeys' option
       --list-all-trusted     List all available certificates marked as trusted
       --export               Export the object specified by the URL
                                - prohibits these options:
                                export-stapled
                                export-chain
                                export-pubkey
       --export-stapled       Export the certificate object specified by the URL
                                - prohibits these options:
                                export
                                export-chain
                                export-pubkey
       --export-chain         Export the certificate specified by the URL and its chain of trust
                                - prohibits these options:
                                export-stapled
                                export
                                export-pubkey
       --export-pubkey        Export the public key for a private key
                                - prohibits these options:
                                export-stapled
                                export
                                export-chain
       --info                 List information on an available object in a token
       --trusted              an alias for the 'mark-trusted' option
       --distrusted           an alias for the 'mark-distrusted' option

Key generation:

       --generate-rsa         Generate an RSA private-public key pair
       --generate-dsa         Generate a DSA private-public key pair
       --generate-ecc         Generate an ECDSA private-public key pair
       --bits=num             Specify the number of bits for the key generate
       --curve=str            Specify the curve used for EC key generation
       --sec-param=str        Specify the security level

Writing objects:

       --set-id=str           Set the CKA_ID (in hex) for the specified by the URL object
                                - prohibits the option 'write'
       --set-label=str        Set the CKA_LABEL for the specified by the URL object
                                - prohibits these options:
                                write
                                set-id
       --write                Writes the loaded objects to a PKCS #11 token
       --delete               Deletes the objects matching the given PKCS #11 URL
       --label=str            Sets a label for the write operation
       --id=str               Sets an ID for the write operation
       --mark-wrap            Marks the generated key to be a wrapping key
                                - disabled as '--no-mark-wrap'
       --mark-trusted         Marks the object to be written as trusted
                                - prohibits the option 'mark-distrusted'
                                - disabled as '--no-mark-trusted'
       --mark-distrusted      When retrieving objects, it requires the objects to be distrusted
(blacklisted)
                                - prohibits the option 'mark-trusted'
       --mark-decrypt         Marks the object to be written for decryption
                                - disabled as '--no-mark-decrypt'
       --mark-sign            Marks the object to be written for signature generation
                                - disabled as '--no-mark-sign'
       --mark-ca              Marks the object to be written as a CA
                                - disabled as '--no-mark-ca'
       --mark-private         Marks the object to be written as private
                                - disabled as '--no-mark-private'
       --ca                   an alias for the 'mark-ca' option
       --private              an alias for the 'mark-private' option
       --secret-key=str       Provide a hex encoded secret key
       --load-privkey=file    Private key file to use
                                - file must pre-exist
       --load-pubkey=file     Public key file to use
                                - file must pre-exist
       --load-certificate=file Certificate file to use
                                - file must pre-exist

Other options:

   -d, --debug=num            Enable debugging
                                - it must be in the range:
                                  0 to 9999
       --outfile=str          Output file
       --login                Force (user) login to token
                                - disabled as '--no-login'
       --so-login             Force security officer login to token
                                - disabled as '--no-so-login'
       --admin-login          an alias for the 'so-login' option
       --test-sign            Tests the signature operation of the provided object
       --generate-random=num  Generate random data
   -8, --pkcs8                Use PKCS #8 format for private keys
       --inder                Use DER/RAW format for input
                                - disabled as '--no-inder'
       --inraw                an alias for the 'inder' option
       --outder               Use DER format for output certificates, private keys, and DH parameters
                                - disabled as '--no-outder'
       --outraw               an alias for the 'outder' option
       --provider=file        Specify the PKCS #11 provider library
       --detailed-url         Print detailed URLs
                                - disabled as '--no-detailed-url'
       --only-urls            Print a compact listing using only the URLs
       --batch                Disable all interaction with the tool

Version, usage and configuration options:

   -v, --version[=arg]        output version information and exit
   -h, --help                 display extended usage information and exit
   -!, --more-help            extended usage information passed thru pager

Options are specified by doubled hyphens and their name or by a single
hyphen and the flag character.
Operands and options may be intermixed.  They will be reordered.

Program that allows operations on PKCS #11 smart cards and security
modules.

To use PKCS #11 tokens with GnuTLS the p11-kit configuration files need to
be setup.  That is create a .module file in /etc/pkcs11/modules with the
contents 'module: /path/to/pkcs11.so'.  Alternatively the configuration
file /etc/gnutls/pkcs11.conf has to exist and contain a number of lines of
the form 'load=/usr/lib/opensc-pkcs11.so'.

You can provide the PIN to be used for the PKCS #11 operations with the
environment variables GNUTLS_PIN and GNUTLS_SO_PIN.

token-related-options options

Tokens.

list-token-urls option.

This is the “list the urls available tokens” option. This is a more compact version of –list-tokens.

set-pin option.

This is the “specify the pin to use on token operations” option. This option takes a string argument. Alternatively the GNUTLS_PIN environment variable may be used.

set-so-pin option.

This is the “specify the security officer’s pin to use on token initialization” option. This option takes a string argument. Alternatively the GNUTLS_SO_PIN environment variable may be used.

object-list-related-options options

Object listing.

list-all-privkeys option.

This is the “list all available private keys in a token” option. Lists all the private keys in a token that match the specified URL.

list-privkeys option.

This is an alias for the list-all-privkeys option, see the list-all-privkeys option documentation.

list-keys option.

This is an alias for the list-all-privkeys option, see the list-all-privkeys option documentation.

export-stapled option.

This is the “export the certificate object specified by the url” option.

This option has some usage constraints. It:

• must not appear in combination with any of the following options: export, export-chain, export-pubkey.

Exports the certificate specified by the URL while including any attached extensions to it. Since attached extensions are a p11-kit extension, this option is only available on p11-kit registered trust modules.

export-chain option.

This is the “export the certificate specified by the url and its chain of trust” option.

This option has some usage constraints. It:

• must not appear in combination with any of the following options: export-stapled, export, export-pubkey.

Exports the certificate specified by the URL and generates its chain of trust based on the stored certificates in the module.

export-pubkey option.

This is the “export the public key for a private key” option.

This option has some usage constraints. It:

• must not appear in combination with any of the following options: export-stapled, export, export-chain.

Exports the public key for the specified private key

trusted option.

This is an alias for the mark-trusted option, see the mark-trusted option documentation.

distrusted option.

This is an alias for the mark-distrusted option, see the mark-distrusted option documentation.

keygen-related-options options

Key generation.

generate-rsa option.

This is the “generate an rsa private-public key pair” option. Generates an RSA private-public key pair on the specified token. Should be combined with –sec-param or –bits.

generate-dsa option.

This is the “generate a dsa private-public key pair” option. Generates a DSA private-public key pair on the specified token. Should be combined with –sec-param or –bits.

generate-ecc option.

This is the “generate an ecdsa private-public key pair” option. Generates an ECDSA private-public key pair on the specified token. Should be combined with –curve, –sec-param or –bits.

bits option.

This is the “specify the number of bits for the key generate” option. This option takes a number argument. For applications which have no key-size restrictions the –sec-param option is recommended, as the sec-param levels will adapt to the acceptable security levels with the new versions of gnutls.

curve option.

This is the “specify the curve used for ec key generation” option. This option takes a string argument. Supported values are secp192r1, secp224r1, secp256r1, secp384r1 and secp521r1.

sec-param option.

This is the “specify the security level” option. This option takes a string argument Security parameter. This is alternative to the bits option. Available options are [low, legacy, medium, high, ultra].

write-object-related-options options

Writing objects.

set-id option.

This is the “set the cka_id (in hex) for the specified by the url object” option. This option takes a string argument.

This option has some usage constraints. It:

• must not appear in combination with any of the following options: write.

Modifies or sets the CKA_ID in the specified by the URL object. The ID should be specified in hexadecimal format without a ’0x’ prefix.

set-label option.

This is the “set the cka_label for the specified by the url object” option. This option takes a string argument.

This option has some usage constraints. It:

• must not appear in combination with any of the following options: write, set-id.

Modifies or sets the CKA_LABEL in the specified by the URL object

write option.

This is the “writes the loaded objects to a pkcs #11 token” option. It can be used to write private, public keys, certificates or secret keys to a token. Must be combined with one of –load-privkey, –load-pubkey, –load-certificate option.

id option.

This is the “sets an id for the write operation” option. This option takes a string argument. Sets the CKA_ID to be set by the write operation. The ID should be specified in hexadecimal format without a ’0x’ prefix.

mark-wrap option.

This is the “marks the generated key to be a wrapping key” option.

This option has some usage constraints. It:

• can be disabled with –no-mark-wrap.

Marks the generated key with the CKA_WRAP flag.

mark-trusted option.

This is the “marks the object to be written as trusted” option.

This option has some usage constraints. It:

• can be disabled with –no-mark-trusted.
• must not appear in combination with any of the following options: mark-distrusted.

Marks the object to be generated/written with the CKA_TRUST flag.

mark-distrusted option.

This is the “when retrieving objects, it requires the objects to be distrusted (blacklisted)” option.

This option has some usage constraints. It:

• must not appear in combination with any of the following options: mark-trusted.

Ensures that the objects retrieved have the CKA_X_TRUST flag. This is p11-kit trust module extension, thus this flag is only valid with p11-kit registered trust modules.

mark-decrypt option.

This is the “marks the object to be written for decryption” option.

This option has some usage constraints. It:

• can be disabled with –no-mark-decrypt.

Marks the object to be generated/written with the CKA_DECRYPT flag set to true.

mark-sign option.

This is the “marks the object to be written for signature generation” option.

This option has some usage constraints. It:

• can be disabled with –no-mark-sign.

Marks the object to be generated/written with the CKA_SIGN flag set to true.

mark-ca option.

This is the “marks the object to be written as a ca” option.

This option has some usage constraints. It:

• can be disabled with –no-mark-ca.

Marks the object to be generated/written with the CKA_CERTIFICATE_CATEGORY as CA.

mark-private option.

This is the “marks the object to be written as private” option.

This option has some usage constraints. It:

• can be disabled with –no-mark-private.

Marks the object to be generated/written with the CKA_PRIVATE flag. The written object will require a PIN to be used.

ca option.

This is an alias for the mark-ca option, see the mark-ca option documentation.

private option.

This is an alias for the mark-private option, see the mark-private option documentation.

secret-key option.

This is the “provide a hex encoded secret key” option. This option takes a string argument. This secret key will be written to the module if –write is specified.

other-options options

Other options.

debug option (-d).

This is the “enable debugging” option. This option takes a number argument. Specifies the debug level.

so-login option.

This is the “force security officer login to token” option.

This option has some usage constraints. It:

• can be disabled with –no-so-login.

Forces login to the token as security officer (admin).

admin-login option.

This is an alias for the so-login option, see the so-login option documentation.

test-sign option.

This is the “tests the signature operation of the provided object” option. It can be used to test the correct operation of the signature operation. If both a private and a public key are available this operation will sign and verify the signed data.

generate-random option.

This is the “generate random data” option. This option takes a number argument. Asks the token to generate a number of bytes of random bytes.

inder option.

This is the “use der/raw format for input” option.

This option has some usage constraints. It:

• can be disabled with –no-inder.

Use DER/RAW format for input certificates and private keys.

inraw option.

This is an alias for the inder option, see the inder option documentation.

outder option.

This is the “use der format for output certificates, private keys, and dh parameters” option.

This option has some usage constraints. It:

• can be disabled with –no-outder.

The output will be in DER or RAW format.

outraw option.

This is an alias for the outder option, see the outder option documentation.

provider option.

This is the “specify the pkcs #11 provider library” option. This option takes a file argument. This will override the default options in /etc/gnutls/pkcs11.conf

batch option.

This is the “disable all interaction with the tool” option. In batch mode there will be no prompts, all parameters need to be specified on command line.

p11tool exit status

One of the following exit values will be returned:

‘0 (EXIT_SUCCESS)’

Successful program execution.

‘1 (EXIT_FAILURE)’

The operation failed or the command syntax was not valid.

p11tool See Also

certtool (1)

p11tool Examples

To view all tokens in your system use:

$ p11tool --list-tokens

To view all objects in a token use:

$ p11tool --login --list-all "pkcs11:TOKEN-URL"

To store a private key and a certificate in a token run:

$ p11tool --login --write "pkcs11:URL" --load-privkey key.pem \
          --label "Mykey"
$ p11tool --login --write "pkcs11:URL" --load-certificate cert.pem \
          --label "Mykey"

Note that some tokens require the same label to be used for the certificate and its corresponding private key.

To generate an RSA private key inside the token use:

$ p11tool --login --generate-rsa --bits 1024 --label "MyNewKey" \
          --outfile MyNewKey.pub "pkcs11:TOKEN-URL"

The bits parameter in the above example is explicitly set because some tokens only support limited choices in the bit length. The output file is the corresponding public key. This key can be used to general a certificate request with certtool.

certtool --generate-request --load-privkey "pkcs11:KEY-URL" \
   --load-pubkey MyNewKey.pub --outfile request.pem

5.4 Trusted Platform Module (TPM)

In this section we present the Trusted Platform Module (TPM) support in GnuTLS. Note that we recommend against using TPM with this API because it is restricted to TPM 1.2. We recommend instead to use PKCS#11 wrappers for TPM such as CHAPS¹⁴ or opencryptoki¹⁵. These will allow using the standard smart card and HSM functionality (see Smart cards and HSMs) for TPM keys.

There was a big hype when the TPM chip was introduced into computers. Briefly it is a co-processor in your PC that allows it to perform calculations independently of the main processor. This has good and bad side-effects. In this section we focus on the good ones; these are the fact that you can use the TPM chip to perform cryptographic operations on keys stored in it, without accessing them. That is very similar to the operation of a PKCS #11 smart card. The chip allows for storage and usage of RSA keys, but has quite some operational differences from PKCS #11 module, and thus require different handling. The basic TPM operations supported and used by GnuTLS, are key generation and signing. That support is currently limited to TPM 1.2.

The next sections assume that the TPM chip in the system is already initialized and in a operational state. If not, ensure that the TPM chip is enabled by your BIOS, that the tcsd daemon is running, and that TPM ownership is set (by running tpm_takeownership).

In GnuTLS the TPM functionality is available in gnutls/tpm.h.

5.4.1 Keys in TPM

The RSA keys in the TPM module may either be stored in a flash memory within TPM or stored in a file in disk. In the former case the key can provide operations as with PKCS #11 and is identified by a URL. The URL is described in [TPMURI] and is of the following form.

tpmkey:uuid=42309df8-d101-11e1-a89a-97bb33c23ad1;storage=user

It consists from a unique identifier of the key as well as the part of the flash memory the key is stored at. The two options for the storage field are ‘user’ and ‘system’. The user keys are typically only available to the generating user and the system keys to all users. The stored in TPM keys are called registered keys.

The keys that are stored in the disk are exported from the TPM but in an encrypted form. To access them two passwords are required. The first is the TPM Storage Root Key (SRK), and the other is a key-specific password. Also those keys are identified by a URL of the form:

tpmkey:file=/path/to/file

When objects require a PIN to be accessed the same callbacks as with PKCS #11 objects are expected (see Accessing objects that require a PIN). Note that the PIN function may be called multiple times to unlock the SRK and the specific key in use. The label in the key function will then be set to ‘SRK’ when unlocking the SRK key, or to ‘TPM’ when unlocking any other key.

5.4.2 Key generation

All keys used by the TPM must be generated by the TPM. This can be done using gnutls_tpm_privkey_generate.

Function: int gnutls_tpm_privkey_generate (gnutls_pk_algorithm_t pk, unsigned int bits, const char * srk_password, const char * key_password, gnutls_tpmkey_fmt_t format, gnutls_x509_crt_fmt_t pub_format, gnutls_datum_t * privkey, gnutls_datum_t * pubkey, unsigned int flags)

pk: the public key algorithm

bits: the security bits

srk_password: a password to protect the exported key (optional)

key_password: the password for the TPM (optional)

format: the format of the private key

pub_format: the format of the public key

privkey: the generated key

pubkey: the corresponding public key (may be null)

flags: should be a list of GNUTLS_TPM_* flags

This function will generate a private key in the TPM chip. The private key will be generated within the chip and will be exported in a wrapped with TPM’s master key form. Furthermore the wrapped key can be protected with the provided password .

Note that bits in TPM is quantized value. If the input value is not one of the allowed values, then it will be quantized to one of 512, 1024, 2048, 4096, 8192 and 16384.

Allowed flags are:

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since: 3.1.0

int gnutls_tpm_get_registered (gnutls_tpm_key_list_t * list)

void gnutls_tpm_key_list_deinit (gnutls_tpm_key_list_t list)

int gnutls_tpm_key_list_get_url (gnutls_tpm_key_list_t list, unsigned int idx, char ** url, unsigned int flags)

Function: int gnutls_tpm_privkey_delete (const char * url, const char * srk_password)

url: the URL describing the key

srk_password: a password for the SRK key

This function will unregister the private key from the TPM chip.

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since: 3.1.0

5.4.3 Using keys

Importing keys

The TPM keys can be used directly by the abstract key types and do not require any special structures. Moreover functions like gnutls_certificate_set_x509_key_file2 can access TPM URLs.

int gnutls_privkey_import_tpm_raw (gnutls_privkey_t pkey, const gnutls_datum_t * fdata, gnutls_tpmkey_fmt_t format, const char * srk_password, const char * key_password, unsigned int flags)

int gnutls_pubkey_import_tpm_raw (gnutls_pubkey_t pkey, const gnutls_datum_t * fdata, gnutls_tpmkey_fmt_t format, const char * srk_password, unsigned int flags)

Function: int gnutls_privkey_import_tpm_url (gnutls_privkey_t pkey, const char * url, const char * srk_password, const char * key_password, unsigned int flags)

pkey: The private key

url: The URL of the TPM key to be imported

srk_password: The password for the SRK key (optional)

key_password: A password for the key (optional)

flags: One of the GNUTLS_PRIVKEY_* flags

This function will import the given private key to the abstract gnutls_privkey_t type.

Note that unless GNUTLS_PRIVKEY_DISABLE_CALLBACKS is specified, if incorrect (or NULL) passwords are given the PKCS11 callback functions will be used to obtain the correct passwords. Otherwise if the SRK password is wrong GNUTLS_E_TPM_SRK_PASSWORD_ERROR is returned and if the key password is wrong or not provided then GNUTLS_E_TPM_KEY_PASSWORD_ERROR is returned.

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since: 3.1.0

Function: int gnutls_pubkey_import_tpm_url (gnutls_pubkey_t pkey, const char * url, const char * srk_password, unsigned int flags)

pkey: The public key

url: The URL of the TPM key to be imported

srk_password: The password for the SRK key (optional)

flags: should be zero

This function will import the given private key to the abstract gnutls_privkey_t type.

Note that unless GNUTLS_PUBKEY_DISABLE_CALLBACKS is specified, if incorrect (or NULL) passwords are given the PKCS11 callback functions will be used to obtain the correct passwords. Otherwise if the SRK password is wrong GNUTLS_E_TPM_SRK_PASSWORD_ERROR is returned.

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since: 3.1.0

Listing and deleting keys

The registered keys (that are stored in the TPM) can be listed using one of the following functions. Those keys are unfortunately only identified by their UUID and have no label or other human friendly identifier. Keys can be deleted from permament storage using gnutls_tpm_privkey_delete.

int gnutls_tpm_get_registered (gnutls_tpm_key_list_t * list)

void gnutls_tpm_key_list_deinit (gnutls_tpm_key_list_t list)

int gnutls_tpm_key_list_get_url (gnutls_tpm_key_list_t list, unsigned int idx, char ** url, unsigned int flags)

Function: int gnutls_tpm_privkey_delete (const char * url, const char * srk_password)

url: the URL describing the key

srk_password: a password for the SRK key

This function will unregister the private key from the TPM chip.

Returns: On success, GNUTLS_E_SUCCESS (0) is returned, otherwise a negative error value.

Since: 3.1.0

5.4.4 Invoking tpmtool

Program that allows handling cryptographic data from the TPM chip.

This section was generated by AutoGen, using the agtexi-cmd template and the option descriptions for the tpmtool program. This software is released under the GNU General Public License, version 3 or later.

tpmtool help/usage (--help)

This is the automatically generated usage text for tpmtool.

The text printed is the same whether selected with the help option (--help) or the more-help option (--more-help). more-help will print the usage text by passing it through a pager program. more-help is disabled on platforms without a working fork(2) function. The PAGER environment variable is used to select the program, defaulting to more. Both will exit with a status code of 0.

tpmtool is unavailable - no --help

debug option (-d)

This is the “enable debugging” option. This option takes a number argument. Specifies the debug level.

generate-rsa option

This is the “generate an rsa private-public key pair” option. Generates an RSA private-public key pair in the TPM chip. The key may be stored in file system and protected by a PIN, or stored (registered) in the TPM chip flash.

user option

This is the “any registered key will be a user key” option.

This option has some usage constraints. It:

• must appear in combination with the following options: register.
• must not appear in combination with any of the following options: system.

The generated key will be stored in a user specific persistent storage.

system option

This is the “any registered key will be a system key” option.

This option has some usage constraints. It:

• must appear in combination with the following options: register.
• must not appear in combination with any of the following options: user.

The generated key will be stored in system persistent storage.

test-sign option

This is the “tests the signature operation of the provided object” option. This option takes a string argument url. It can be used to test the correct operation of the signature operation. This operation will sign and verify the signed data.

sec-param option

This is the “specify the security level [low, legacy, medium, high, ultra].” option. This option takes a string argument Security parameter. This is alternative to the bits option. Note however that the values allowed by the TPM chip are quantized and given values may be rounded up.

inder option

This is the “use the der format for keys.” option.

This option has some usage constraints. It:

• can be disabled with –no-inder.

The input files will be assumed to be in the portable DER format of TPM. The default format is a custom format used by various TPM tools

outder option

This is the “use der format for output keys” option.

This option has some usage constraints. It:

• can be disabled with –no-outder.

The output will be in the TPM portable DER format.

tpmtool exit status

One of the following exit values will be returned:

‘0 (EXIT_SUCCESS)’

Successful program execution.

‘1 (EXIT_FAILURE)’

The operation failed or the command syntax was not valid.

tpmtool See Also

p11tool (1), certtool (1)

tpmtool Examples

To generate a key that is to be stored in file system use:

$ tpmtool --generate-rsa --bits 2048 --outfile tpmkey.pem

To generate a key that is to be stored in TPM’s flash use:

$ tpmtool --generate-rsa --bits 2048 --register --user

To get the public key of a TPM key use:

$ tpmtool --pubkey tpmkey:uuid=58ad734b-bde6-45c7-89d8-756a55ad1891;storage=user \
          --outfile pubkey.pem

or if the key is stored in the file system:

$ tpmtool --pubkey tpmkey:file=tmpkey.pem --outfile pubkey.pem

To list all keys stored in TPM use:

$ tpmtool --list

6 How to use GnuTLS in applications