document the gnutls_pcert_st

[gnutls.git] / doc / cha-gtls-app.texi

@node How to use GnuTLS in applications
@chapter How to use @acronym{GnuTLS} in applications

@menu
* Introduction to the library::
* Preparation::
* Session initialization::
* Associating the credentials::
* Setting up the transport layer::
* TLS handshake::
* Data transfer and termination::
* Handling alerts::
* Priority Strings::
* Advanced and other topics::
* Using the cryptographic library::
* Selecting cryptographic key sizes::
@end menu

@node Introduction to the library
@section Introduction

@menu
* General idea::
* Error handling::
* Common types::
* Debugging and auditing::
* Thread safety::
* Callback functions::
@end menu

@node General idea
@subsection General idea

A brief description of how @acronym{GnuTLS} works internally is shown
at @ref{fig:gnutls-design}. This section may become more clear after
having read the rest of this section.
As shown in the figure, there is a read-only global state that is
initialized once by the global initialization function.  This global
structure, among others, contains the memory allocation functions
used, and structures needed for the @acronym{ASN.1} parser.  This
structure is never modified by any @acronym{GnuTLS} function, except
for the deinitialization function which frees all allocated memory
and is called after the program has permanently
finished using @acronym{GnuTLS}.

@float Figure,fig:gnutls-design
@image{gnutls-internals,12cm}
@caption{High level design of GnuTLS.}
@end float

The credentials structures are used by the authentication methods, such
as certificate authentication. They store certificates, privates keys,
and other information that is needed to prove the identity to the peer,
and/or verify the indentity of the peer. The information stored in
the credentials structures is initialized once and then can be 
shared by many @acronym{TLS} sessions.

A @acronym{GnuTLS} session contains all the required information
to handle one secure connection. The session communicates with the
peers using the provided functions of the transport layer.
Every session has a unique session ID shared with the peer.

Since TLS sessions can be resumed, servers need a
database back-end to hold the session's parameters.  Every
@acronym{GnuTLS} session after a successful handshake calls the
appropriate back-end function (see @ref{resume})
to store the newly negotiated session. The session
database is examined by the server just after having received the
client hello@footnote{The first message in a @acronym{TLS} handshake},
and if the session ID sent by the client, matches a stored session,
the stored session will be retrieved, and the new session will be a
resumed one, and will share the same session ID with the previous one.

@node Error handling
@subsection Error handling

In @acronym{GnuTLS} most functions return an integer type as a result.
In almost all cases a zero or a positive number means success, and a
negative number indicates failure, or a situation that some action has
to be taken. Thus negative error codes may be fatal or not.

Fatal errors terminate the connection immediately and further sends
and receives will be disallowed.  Such an example is
@code{GNUTLS_@-E_@-DECRYPTION_@-FAILED}. Non-fatal errors may warn about
something, i.e., a warning alert was received, or indicate the some
action has to be taken. This is the case with the error code
@code{GNUTLS_@-E_@-REHANDSHAKE} returned by @funcref{gnutls_record_recv}.
This error code indicates that the server requests a re-handshake. The
client may ignore this request, or may reply with an alert.  You can
test if an error code is a fatal one by using the
@funcref{gnutls_error_is_fatal}.
All errors can be converted to a descriptive string using @funcref{gnutls_strerror}.

If any non fatal errors, that require an action, are to be returned by
a function, these error codes will be documented in the function's
reference. For example the error codes @code{GNUTLS_@-E_@-WARNING_@-ALERT_@-RECEIVED} and @code{GNUTLS_@-E_@-FATAL_@-ALERT_@-RECEIVED}
that may returned when receiving data, should be handled by notifying the
user of the alert (as explained in @ref{Handling alerts}).
See @ref{Error codes}, for a description of the available error codes.

@node Common types
@subsection Common types

Several functions in @acronym{GnuTLS} use @code{gnutls_datum_t} which is
convenient way to combine a pointer to data and data's size. Its definition is
shown below.
@verbatim
  typedef struct
  {
    unsigned char *data;
    unsigned int size;
  } gnutls_datum_t;
@end verbatim

Other functions that require data for scattered read use a structure similar
to @code{struct iovec} typically used by @funcintref{readv}. It is shown
below.
@verbatim
  typedef struct
  {
    void *iov_base;             /* Starting address */
    size_t iov_len;             /* Number of bytes to transfer */
  } giovec_t;
@end verbatim


@node Debugging and auditing
@subsection Debugging and auditing

In many cases things may not go as expected and further information,
to assist debugging, from @acronym{GnuTLS} is desired. 
Those are the cases where the @funcref{gnutls_global_set_log_level} and
@funcref{gnutls_global_set_log_function} are to be used. Those will print
verbose information on the @acronym{GnuTLS} functions internal flow.

@showfuncB{gnutls_global_set_log_level,gnutls_global_set_log_function}

When debugging is not required, important issues, such as detected
attacks on the protocol still need to be logged. This is provided
by the logging function set by
@funcref{gnutls_global_set_audit_log_function}. The provided function
will receive an message and the corresponding
TLS session. The session information might be used to derive IP addresses
or other information about the peer involved.

@showfuncdesc{gnutls_global_set_audit_log_function}

@node Thread safety
@subsection Thread safety
@cindex thread safety

The @acronym{GnuTLS} library is thread safe by design, meaning that
objects of the library such as TLS sessions, can be safely divided across
threads as long as a single thread accesses a single object. This is
sufficient to support a server which handles several sessions per thread.
If, however, an object needs to be shared across threads then access must be 
protected with a mutex. Read-only access to objects, for example the
credentials holding structures, is also thread-safe. 

The random generator of the cryptographic back-end, is not thread safe and requires
mutex locks which are setup by @acronym{GnuTLS}.
Applications can either call @funcref{gnutls_global_init} which will initialize the default
operating system provided locks (i.e. @code{pthreads} on GNU/Linux and
@code{CriticalSection} on Windows), or manually specify the locking system using
the function @funcref{gnutls_global_set_mutex} before calling @funcref{gnutls_global_init}. 
Setting mutexes manually is recommended
only for applications that have full control of the underlying libraries. If this
is not the case, the use of the operating system defaults is recommended. An example of 
non-native thread usage is shown below.

@example
#include <gnutls/gnutls.h>

int main()
@{
   /* When the system mutexes are not to be used 
    * gnutls_global_set_mutex() must be called explicitly
    */
   gnutls_global_set_mutex (mutex_init, mutex_deinit, 
                            mutex_lock, mutex_unlock);
   gnutls_global_init();
@}
@end example

@showfuncdesc{gnutls_global_set_mutex}

@node Callback functions
@subsection Callback functions
@cindex callback functions

There are several cases where @acronym{GnuTLS} may need out of
band input from your program. This is now implemented using some
callback functions, which your program is expected to register.

An example of this type of functions are the push and pull callbacks
which are used to specify the functions that will retrieve and send
data to the transport layer.

@showfuncB{gnutls_transport_set_push_function,gnutls_transport_set_pull_function}

Other callback functions may require more complicated input and data
to be allocated. Such an example is 
@funcref{gnutls_srp_set_server_credentials_function}.
All callbacks should allocate and free memory using 
@funcintref{gnutls_malloc} and @funcintref{gnutls_free}.


@node Preparation
@section Preparation

To use @acronym{GnuTLS}, you have to perform some changes to your
sources and your build system. The necessary changes are explained in
the following subsections.

@menu
* Headers::
* Initialization::
* Version check::
* Building the source::
@end menu

@node Headers
@subsection Headers

All the data types and functions of the @acronym{GnuTLS} library are
defined in the header file @file{gnutls/gnutls.h}.  This must be
included in all programs that make use of the @acronym{GnuTLS}
library.

@node Initialization
@subsection Initialization

GnuTLS must be initialized before it can be used.  The library is
initialized by calling @funcref{gnutls_global_init}.  The resources
allocated by the initialization process can be released if the
application no longer has a need to call GnuTLS functions, this is
done by calling @funcref{gnutls_global_deinit}.

In order to take advantage of the internationalization features in
GnuTLS, such as translated error messages, the application must set
the current locale using @code{setlocale} before initializing GnuTLS.

@node Version check
@subsection Version check

It is often desirable to check that the version of `gnutls' used is
indeed one which fits all requirements.  Even with binary
compatibility new features may have been introduced but due to problem
with the dynamic linker an old version is actually used.  So you may
want to check that the version is okay right after program start-up.
See the function @funcref{gnutls_check_version}.

@node Building the source
@subsection Building the source

If you want to compile a source file including the
@file{gnutls/gnutls.h} header file, you must make sure that the
compiler can find it in the directory hierarchy.  This is accomplished
by adding the path to the directory in which the header file is
located to the compilers include file search path (via the @option{-I}
option).

However, the path to the include file is determined at the time the
source is configured.  To solve this problem, the library uses the
external package @command{pkg-config} that knows the path to the
include file and other configuration options.  The options that need
to be added to the compiler invocation at compile time are output by
the @option{--cflags} option to @command{pkg-config gnutls}.  The
following example shows how it can be used at the command line:

@example
gcc -c foo.c `pkg-config gnutls --cflags`
@end example

Adding the output of @samp{pkg-config gnutls --cflags} to the
compilers command line will ensure that the compiler can find the
@file{gnutls/gnutls.h} header file.

A similar problem occurs when linking the program with the library.
Again, the compiler has to find the library files.  For this to work,
the path to the library files has to be added to the library search
path (via the @option{-L} option).  For this, the option
@option{--libs} to @command{pkg-config gnutls} can be used.  For
convenience, this option also outputs all other options that are
required to link the program with the library (for instance, the
@samp{-ltasn1} option).  The example shows how to link @file{foo.o}
with the library to a program @command{foo}.

@example
gcc -o foo foo.o `pkg-config gnutls --libs`
@end example

Of course you can also combine both examples to a single command by
specifying both options to @command{pkg-config}:

@example
gcc -o foo foo.c `pkg-config gnutls --cflags --libs`
@end example

When a program uses the GNU autoconf system, then the following
line or similar can be used to detect the presence of GnuTLS.

@example
PKG_CHECK_MODULES([LIBGNUTLS], [gnutls >= 3.0.0])

AC_SUBST([LIBGNUTLS_CFLAGS])
AC_SUBST([LIBGNUTLS_LIBS])
@end example

@node Session initialization
@section Session initialization

In the previous sections we have discussed the global initialization
required for GnuTLS as well as the initialization required for each
authentication method's credentials (see @ref{Authentication}).
In this section we elaborate on the TLS or DTLS session initiation.
Each session is initialized using @funcref{gnutls_init} which among
others is used to specify the type of the connection (server or client), 
and the underlying protocol type, i.e., datagram (UDP) or reliable (TCP).

@showfuncdesc{gnutls_init}

After the session initialization details on the allowed ciphersuites
and protocol versions should be set using the priority functions
such as @funcref{gnutls_priority_set_direct}. We elaborate on them
in @ref{Priority Strings}.
The credentials used for the key exchange method, such as certificates 
or usernames and passwords should also be associated with the session
current session using @funcref{gnutls_credentials_set}. 

@showfuncdesc{gnutls_credentials_set}

@node Associating the credentials
@section Associating the credentials

@menu
* Certificate credentials::
* SRP credentials::
* PSK credentials::
* Anonymous credentials::
@end menu

Each authentication method is associated with a key exchange method, and a credentials type. 
The contents of the credentials is method-dependent, e.g. certificates
for certificate authentication and should be initialized and associated
with a session (see @funcref{gnutls_credentials_set}).  A mapping of the key exchange methods
with the credential types is shown in @ref{tab:key-exchange-cred}.

@float Table,tab:key-exchange-cred
@multitable @columnfractions .25 .25 .2 .2

@headitem Authentication method @tab Key exchange @tab Client credentials @tab Server credentials

@item Certificate
@tab @code{KX_RSA},
@code{KX_DHE_RSA},
@code{KX_DHE_DSS},
@code{KX_ECDHE_RSA},
@code{KX_ECDHE_ECDSA},
@code{KX_RSA_EXPORT}
@tab @code{CRD_CERTIFICATE}
@tab @code{CRD_CERTIFICATE}

@item Password and certificate
@tab @code{KX_SRP_RSA}, @code{KX_SRP_DSS}
@tab @code{CRD_SRP}
@tab @code{CRD_CERTIFICATE}, @code{CRD_SRP}

@item Password
@tab @code{KX_SRP}
@tab @code{CRD_SRP}
@tab @code{CRD_SRP}

@item Anonymous
@tab @code{KX_ANON_DH},
@code{KX_ANON_ECDH}
@tab @code{CRD_ANON}
@tab @code{CRD_ANON}

@item Pre-shared key
@tab @code{KX_PSK},
@code{KX_DHE_PSK}, @code{KX_ECDHE_PSK}
@tab @code{CRD_PSK}
@tab @code{CRD_PSK}

@end multitable
@caption{Key exchange algorithms and the corresponding credential types.}
@end float

@node Certificate credentials
@subsection Certificates
@subsubheading Server certificate authentication

When using certificates the server is required to have at least one
certificate and private key pair. Clients may not hold such
a pair, but a server could require it. In this section we discuss
general issues applying to both client and server certificates. The next
section will elaborate on issues arising from client authentication only.

@showfuncB{gnutls_certificate_allocate_credentials,gnutls_certificate_free_credentials}

After the credentials structures are initialized, the certificate 
and key pair must be loaded. This occurs before any @acronym{TLS} 
session is initialized, and the same structures are reused for multiple sessions.
Depending on the certificate type different loading functions
are available, as shown below.
For @acronym{X.509} certificates, the functions will
accept and use a certificate chain that leads to a trusted
authority. The certificate chain must be ordered in such way that every
certificate certifies the one before it. The trusted authority's
certificate need not to be included since the peer should possess it
already.

@showfuncC{gnutls_certificate_set_x509_key_mem,gnutls_certificate_set_x509_key,gnutls_certificate_set_x509_key_file}

@showfuncD{gnutls_certificate_set_openpgp_key_mem,gnutls_certificate_set_openpgp_key,gnutls_certificate_set_openpgp_key_file,gnutls_certificate_set_key}

If multiple certificates are used with the functions above each
client's request will be served with the certificate that matches the
requested name (see @ref{Server name indication}).

As an alternative to loading from files or buffers, a callback may be used for the 
server or the client to specify the certificate and the key at the handshake time.
In that case a certificate should be selected according the peer's signature
algorithm preferences. To get those preferences use
@funcref{gnutls_sign_algorithm_get_requested}. Both functions are shown below.

@showfuncC{gnutls_certificate_set_retrieve_function,gnutls_certificate_set_retrieve_function2,gnutls_sign_algorithm_get_requested}

The functions above do not handle the requested server name automatically.
A server would need to check the name requested by the client
using @funcref{gnutls_server_name_get}, and serve the appropriate
certificate. Note that some of these functions require the @code{gnutls_pcert_st} structure to be
filled in. Helper functions to make the required structures are listed below.

@verbatim
typedef struct gnutls_pcert_st
{
  gnutls_pubkey_t pubkey;
  gnutls_datum_t cert;
  gnutls_certificate_type_t type;
} gnutls_pcert_st;
@end verbatim

@showfuncE{gnutls_pcert_import_x509,gnutls_pcert_import_openpgp,gnutls_pcert_import_x509_raw,gnutls_pcert_import_openpgp_raw,gnutls_pcert_deinit}

In a handshake, the negotiated cipher suite depends on the
certificate's parameters, so some key exchange methods might not be
available with all certificates. @acronym{GnuTLS} will disable
ciphersuites that are not compatible with the key, or the enabled
authentication methods.  For example keys marked as sign-only, will
not be able to access the plain RSA ciphersuites, that require
decryption. It is not recommended to use RSA keys for both
signing and encryption. If possible use a different key for the
@code{DHE-RSA} which uses signing and @code{RSA} that requires decryption.
All the key exchange methods shown in @ref{tab:key-exchange} are
available in certificate authentication.


@subsubheading Client certificate authentication

If a certificate is to be requested from the client during the handshake, the server
will send a certificate request message. This behavior is controlled @funcref{gnutls_certificate_server_set_request}.
The request contains a list of the acceptable by the server certificate signers. This list
is constructed using the trusted certificate authorities of the server.
In cases where the server supports a large number of certificate authorities
it makes sense not to advertise all of the names to save bandwidth. That can
be controlled using the function @funcref{gnutls_certificate_send_x509_rdn_sequence}. 
This however will have the side-effect of not restricting the client to certificates
signed by server's acceptable signers.

@showfuncdesc{gnutls_certificate_server_set_request}

@showfuncdesc{gnutls_certificate_send_x509_rdn_sequence}


@subsubheading Client or server certificate verification

Certificate verification is possible by loading the trusted
authorities into the credentials structure by using
the following functions, applicable to X.509 and OpenPGP certificates.

@showfuncB{gnutls_certificate_set_x509_trust_file,gnutls_certificate_set_openpgp_keyring_file}

The peer's certificate is not automatically verified and one 
should call @funcref{gnutls_certificate_verify_peers2}
after a successful handshake to verify the certificate's signature.
Alternative the verification can occur during the handshake
by using @funcref{gnutls_certificate_set_verify_function}.

In order to report a detailed verification output, an alternative
way has to be used. For that, one should call @funcref{gnutls_certificate_get_peers} 
to obtain the raw certificate of the peer and verify it using the
functions discussed in @ref{X.509 certificates}.

@showfuncdesc{gnutls_certificate_verify_peers2}

@showfuncdesc{gnutls_certificate_set_verify_function}



@node SRP credentials
@subsection SRP

The initialization functions in SRP credentials differ between
client and server.
Clients supporting @acronym{SRP} should set the username and password
prior to connection, to the credentials structure.
Alternatively @funcref{gnutls_srp_set_client_credentials_function}
may be used instead, to specify a callback function that should return the
SRP username and password.
The callback is called once during the @acronym{TLS} handshake.

@showfuncE{gnutls_srp_allocate_server_credentials,gnutls_srp_allocate_client_credentials,gnutls_srp_free_server_credentials,gnutls_srp_free_client_credentials,gnutls_srp_set_client_credentials}

@showfuncdesc{gnutls_srp_set_client_credentials_function}

In server side the default behavior of @acronym{GnuTLS} is to read
the usernames and @acronym{SRP} verifiers from password files. These
password file format is compatible the with the @emph{Stanford srp libraries}
format.  If a different password file format is to be used, then 
@funcref{gnutls_srp_set_server_credentials_function} should be called,
to set an appropriate callback. 

@showfuncdesc{gnutls_srp_set_server_credentials_file}

@showfuncdesc{gnutls_srp_set_server_credentials_function}


@node PSK credentials
@subsection PSK
The initialization functions in PSK credentials differ between
client and server.

@showfuncD{gnutls_psk_allocate_server_credentials,gnutls_psk_allocate_client_credentials,gnutls_psk_free_server_credentials,gnutls_psk_free_client_credentials}

Clients supporting @acronym{PSK} should supply the username and key
before a TLS session is established.  Alternatively 
@funcref{gnutls_psk_set_client_credentials_function} can be used to
specify a callback function. This has the
advantage that the callback will be called only if @acronym{PSK} has
been negotiated.

@showfuncA{gnutls_psk_set_client_credentials}

@showfuncdesc{gnutls_psk_set_client_credentials_function}

In server side the default behavior of @acronym{GnuTLS} is to read
the usernames and @acronym{PSK} keys from a password file. The
password file should contain usernames and keys in hexadecimal
format. The name of the password file can be stored to the credentials
structure by calling @funcref{gnutls_psk_set_server_credentials_file}.  If
a different password file format is to be used, then
a callback should be set instead by @funcref{gnutls_psk_set_server_credentials_function}.

The server can help the client chose a suitable username and password,
by sending a hint. Note that there is no common profile for the PSK hint and applications
are discouraged to use it.
A server, may specify the hint by calling
@funcref{gnutls_psk_set_server_credentials_hint}.  The client can retrieve
the hint, for example in the callback function, using
@funcref{gnutls_psk_client_get_hint}.

@showfuncdesc{gnutls_psk_set_server_credentials_file}

@showfuncC{gnutls_psk_set_server_credentials_function,gnutls_psk_set_server_credentials_hint,gnutls_psk_client_get_hint}

@node Anonymous credentials
@subsection Anonymous
The key exchange methods for anonymous authentication
might require Diffie-Hellman parameters to be generated by the server and
associated with an anonymous credentials structure. Check
@ref{Parameter generation} for more information.
The initialization functions for the credentials are shown below.

@showfuncD{gnutls_anon_allocate_server_credentials,gnutls_anon_allocate_client_credentials,gnutls_anon_free_server_credentials,gnutls_anon_free_client_credentials}



@node Setting up the transport layer
@section Setting up the transport layer

The next step is to setup the underlying transport layer details. The
Berkeley sockets are implicitly used by GnuTLS, thus a
call to @funcref{gnutls_transport_set_ptr2} would be sufficient to
specify the socket descriptor.

@showfuncdesc{gnutls_transport_set_ptr2}

@showfuncA{gnutls_transport_set_ptr}

If however another transport layer than TCP is selected, then
the following functions have to be specified.

@showfuncdesc{gnutls_transport_set_push_function}
@showfuncdesc{gnutls_transport_set_vec_push_function}
@showfuncdesc{gnutls_transport_set_pull_function}

The functions above accept a callback function which
should return the number of bytes written, or -1 on
error and should set @code{errno} appropriately.
In some environments, setting @code{errno} is unreliable. For example
Windows have several errno variables in different CRTs, or in other
systems it may be a non thread-local variable.  If this is a concern to
you, call @funcref{gnutls_transport_set_errno} with the intended errno
value instead of setting @code{errno} directly.

@showfuncdesc{gnutls_transport_set_errno}

@acronym{GnuTLS} currently only interprets the EINTR, EAGAIN and EMSGSIZE errno
values and returns the corresponding @acronym{GnuTLS} error codes:
@itemize
@item @code{GNUTLS_E_INTERRUPTED} 
@item @code{GNUTLS_E_AGAIN}
@item @code{GNUTLS_E_LARGE_PACKET}
@end itemize
The EINTR and EAGAIN values are returned by interrupted system calls, 
or when non blocking IO is used.  All @acronym{GnuTLS} functions can be 
resumed (called again), if any of the above error codes is returned. The
EMSGSIZE value is returned when attempting to send a large datagram.

In the case of DTLS it is also desirable to override the generic 
transport functions with functions that emulate the operation
of @code{recvfrom} and @code{sendto}. In addition
@acronym{DTLS} requires timers during the receive of a handshake
message, set using the @funcref{gnutls_transport_set_pull_timeout_function} 
function. To check the retransmission timers the function
@funcref{gnutls_dtls_get_timeout} is provided, which returns the time
remaining until the next retransmission, or better the time until 
@funcref{gnutls_handshake} should be called again.

@showfuncdesc{gnutls_transport_set_pull_timeout_function}
@showfuncdesc{gnutls_dtls_get_timeout}

@menu
* Asynchronous operation::
* DTLS sessions::
@end menu

@node Asynchronous operation
@subsection Asynchronous operation
@acronym{GnuTLS} can be used with asynchronous socket or event-driven programming.
The approach is similar to using Berkeley sockets under such an environment.
The blocking, due to network interaction, calls such as
@funcref{gnutls_handshake}, @funcref{gnutls_record_recv},
can be set to non-blocking by setting the underlying sockets to non-blocking.
If other push and pull functions are setup, then they should behave the same
way as @funcintref{recv} and @funcintref{send} when used in a non-blocking
way, i.e., set errno to @code{EAGAIN}. Since, during a TLS protocol session 
@acronym{GnuTLS} does not block except for network interaction, the non blocking
@code{EAGAIN} errno will be propagated and @acronym{GnuTLS} functions 
will return the @code{GNUTLS_E_AGAIN} error code. Such calls can be resumed the 
same way as a system call would. 
The only exception is @funcref{gnutls_record_send},
which if interrupted subsequent calls need not to include the data to be
sent (can be called with NULL argument).

The @funcintref{select} system call can also be used in combination with the
@acronym{GnuTLS} functions. @funcintref{select} allows monitoring of sockets
and notifies on them being ready for reading or writing data. Note however
that this system call cannot notify on data present in @acronym{GnuTLS}
read buffers, it is only applicable to the kernel sockets API. Thus if
you are using it for reading from a @acronym{GnuTLS} session, make sure
that any cached data are read completely. That can be achieved by checking there 
are no data waiting to be read (using @funcref{gnutls_record_check_pending}), 
either before the @funcintref{select} system call, or after a call to
@funcref{gnutls_record_recv}. @acronym{GnuTLS} does not keep a write buffer,
thus when writing no additional actions are required.

Although in the TLS protocol implementation each call to receive or send
function implies to restoring the same function that was interrupted, in
the DTLS protocol this requirement isn't true.
There are cases where a retransmission is required, which are indicated by
a received message and thus @funcref{gnutls_record_get_direction} must be called 
to decide which operation to restore, i.e., send or receive.
@showfuncdesc{gnutls_record_get_direction}

Moreover, to prevent blocking from DTLS' retransmission timers to block a
handshake, the @funcref{gnutls_init} function should be called with the
@code{GNUTLS_NONBLOCK} flag set (see @ref{Session initialization}).


@node DTLS sessions
@subsection DTLS sessions

Because datagram TLS can operate over connections where the peer
of a server cannot be reliably verified, functionality is available to prevent
denial of service attacks. @acronym{GnuTLS} requires a server
to generate a secret key that is used to sign a cookie@footnote{A key of 128 bits or 16 bytes should be sufficient for this purpose.}. 
That cookie is sent to the client using @funcref{gnutls_dtls_cookie_send}, and 
the client must reply using the correct cookie. The server side
should verify the initial message sent by client using @funcref{gnutls_dtls_cookie_verify}.
If successful the session should be initialized and associated with
the cookie using @funcref{gnutls_dtls_prestate_set}, before proceeding to
the handshake.

@showfuncD{gnutls_key_generate,gnutls_dtls_cookie_send,gnutls_dtls_cookie_verify,gnutls_dtls_prestate_set}

Note that the above apply to server side only and they are not mandatory to be
used. Not using them, however, allows denial of service attacks.
The client side cookie handling is part of @funcref{gnutls_handshake}. 

Datagrams are typically restricted by a maximum transfer unit (MTU). For that
both client and server side should set the correct maximum transfer unit for
the layer underneath @acronym{GnuTLS}. This will allow proper fragmentation
of DTLS messages and prevent messages from being silently discarded by the
transport layer. The ``correct'' maximum transfer unit can be obtained through
a path MTU discovery mechanism @xcite{RFC4821}.

@showfuncC{gnutls_dtls_set_mtu,gnutls_dtls_get_mtu,gnutls_dtls_get_data_mtu}


@node TLS handshake
@section TLS handshake
Once a session has been initialized and a network
connection has been set up, TLS and DTLS protocols
perform a handshake. The handshake is the actual key
exchange.

@showfuncdesc{gnutls_handshake}

The handshake process doesn't ensure the verification
of the peer's identity. When certificates are in use,
this can be done, either after the handshake is complete, or during 
the handshake if @funcref{gnutls_certificate_set_verify_function}
has been used. In both cases the @funcref{gnutls_certificate_verify_peers2} function can be
used to verify the peer's certificate (see @ref{Certificate authentication}
for more information).

@showfuncA{gnutls_certificate_verify_peers2}

@node Data transfer and termination
@section Data transfer and termination
Once the handshake is complete and peer's identity
has been verified data can be exchanged. The available
functions resemble the POSIX @code{recv} and @code{send}
functions. It is suggested to use @funcref{gnutls_error_is_fatal}
to check whether the error codes returned by these functions are
fatal for the protocol or can be ignored.

@showfuncdesc{gnutls_record_send}

@showfuncdesc{gnutls_record_recv}

@showfuncdesc{gnutls_error_is_fatal}

Although, in the TLS protocol the receive function can be called
at any time, when DTLS is used the GnuTLS receive functions must be
called once a message is available for reading, even if no data are
expected. This is because in DTLS various (internal) actions
may be required due to retransmission timers. Moreover,
an extended receive function is shown below, which allows the extraction
of the message's sequence number. Due to the unreliable nature of the
protocol, this field allows distinguishing out-of-order messages.

@showfuncdesc{gnutls_record_recv_seq}

The @funcref{gnutls_record_check_pending} helper function is available to 
allow checking whether data are available to be read in a @acronym{GnuTLS} session 
buffers. Note that this function complements but does not replace @funcintref{select},
i.e., @funcref{gnutls_record_check_pending} reports no data to be read, @funcintref{select}
should be called to check for data in the network buffers.

@showfuncdesc{gnutls_record_check_pending}
@showfuncA{gnutls_record_get_direction}

Once a TLS or DTLS session is no longer needed, it is
recommended to use @funcref{gnutls_bye} to terminate the
session. That way the peer is notified securely about the
intention of termination, which allows distinguishing it
from a malicious connection termination.
A session can be deinitialized with the @funcref{gnutls_deinit} function.

@showfuncdesc{gnutls_bye}
@showfuncdesc{gnutls_deinit}

@node Handling alerts
@section Handling alerts
During a TLS connection alert messages may be exchanged by the
two peers. Those messages may be fatal, meaning the connection
must be terminated afterwards, or warning when something needs
to be reported to the peer, but without interrupting the session.
The error codes @code{GNUTLS_E_@-WARNING_@-ALERT_@-RECEIVED}
or @code{GNUTLS_E_@-FATAL_@-ALERT_@-RECEIVED} signal those alerts
when received, and may be returned by all GnuTLS functions that receive 
data from the peer, being @funcref{gnutls_handshake} and @funcref{gnutls_record_recv}.

If those error codes are received the alert and its level should be logged
or reported to the peer using the functions below.

@showfuncdesc{gnutls_alert_get}
@showfuncdesc{gnutls_alert_get_name}

The peer may also be warned or notified of a fatal issue
by using one of the functions below. All the available alerts
are listed in @ref{The Alert Protocol}.

@showfuncdesc{gnutls_alert_send}
@showfuncdesc{gnutls_error_to_alert}


@node Priority Strings
@section Priority strings
@cindex Priority strings

In order to specify cipher suite preferences on a TLS session
there are priority functions that accept a string
specifying the enabled for the handshake algorithms.
That string may contain a single initial keyword such as
in @ref{tab:prio-keywords} and may be followed by
additional algorithm or special keywords.

@showfuncB{gnutls_priority_set_direct,gnutls_priority_set}

@float Table,tab:prio-keywords
@multitable @columnfractions .20 .70
@headitem Keyword @tab Description
@item PERFORMANCE @tab
All the known to be secure ciphersuites are enabled,
limited to 128 bit ciphers and sorted by terms of speed
performance. The message authenticity security level is of 64 bits or more.

@item NORMAL @tab
Means all the known to be secure ciphersuites. The ciphers are sorted by security
margin, although the 256-bit ciphers are included as a fallback only.
The message authenticity security level is of 64 bits or more.

@item SECURE128 @tab
Means all known to be secure ciphersuites that offer a 
security level 128-bit or more and a message authenticity
security level of 80 bits or more.

@item SECURE192 @tab
Means all the known to be secure ciphersuites that offer a 
security level 192-bit or more and a message authenticity
security level of 128 bits or more.

@item SECURE256 @tab
Currently alias for SECURE192.

@item SUITEB128 @tab
Means all the NSA Suite B cryptography (RFC5430) ciphersuites
with an 128 bit security level.

@item SUITEB192 @tab
Means all the NSA Suite B cryptography (RFC5430) ciphersuites
with an 192 bit security level.

@item EXPORT @tab
Means all ciphersuites are enabled, including the
low-security 40 bit ciphers.

@item NONE @tab
Means nothing is enabled.  This disables even protocols and
compression methods. It should be followed by the
algorithms to be enabled.

@end multitable
@caption{Supported initial keywords.}
@end float

Unless the initial keyword is "NONE" the defaults (in preference
order) are for TLS protocols TLS 1.2, TLS1.1, TLS1.0, SSL3.0; for
compression NULL; for certificate types X.509.
In key exchange algorithms when in NORMAL or SECURE levels the
perfect forward secrecy algorithms take precedence of the other
protocols.  In all cases all the supported key exchange algorithms
are enabled@footnote{Except for the RSA-EXPORT which is only enabled in
EXPORT level.}.

Note that the SECURE levels distinguish between overall security level and
message authenticity security level. That is because the message
authenticity security level requires the adversary to break
the algorithms at real-time during the protocol run, whilst 
the overall security level refers to off-line adversaries 
(e.g. adversaries breaking the ciphertext years after it was captured).

The NONE keyword, if used, must followed by keywords specifying 
the algorithms and protocols to be enabled. The other initial keywords may be 
followed by such keywords.
The order with which every algorithm or protocol
is specified is significant. Algorithms specified before others
will take precedence. The supported algorithms and protocols
are shown in @ref{tab:prio-algorithms}. 
To avoid collisions in order to specify a compression algorithm in
the priority string you have to prefix it with "COMP-", protocol versions
with "VERS-", signature algorithms with "SIGN-" and certificate types with "CTYPE-". 
All other algorithms don't need a prefix. Each specified keyword can
be prefixed with any of the following characters.

@table @asis
@item '!' or '-' 
appended with an algorithm will remove this algorithm.
@item "+" 
appended with an algorithm will add this algorithm.
@end table

@float Table,tab:prio-algorithms
@multitable @columnfractions .20 .70
@headitem Type @tab Keywords
@item Ciphers @tab
AES-128-CBC, AES-256-CBC, AES-128-GCM, CAMELLIA-128-CBC,
CAMELLIA-256-CBC, ARCFOUR-128, 3DES-CBC ARCFOUR-40. Catch all
name is CIPHER-ALL which will add all the algorithms from NORMAL
priority.

@item Key exchange @tab
RSA, DHE-RSA, DHE-DSS, SRP, SRP-RSA, SRP-DSS,
PSK, DHE-PSK, ECDHE-RSA, ANON-ECDH, ANON-DH, RSA-EXPORT. The
Catch all name is KX-ALL which will add all the algorithms from NORMAL
priority.

@item MAC @tab
MD5, SHA1, SHA256, AEAD (used with
GCM ciphers only). All algorithms from NORMAL priority can be accessed with MAC-ALL.

@item Compression algorithms @tab
COMP-NULL, COMP-DEFLATE. Catch all is COMP-ALL.

@item TLS versions @tab
VERS-SSL3.0, VERS-TLS1.0, VERS-TLS1.1,
VERS-TLS1.2, VERS-DTLS1.0. Catch all is VERS-TLS-ALL.

@item Signature algorithms @tab
SIGN-RSA-SHA1, SIGN-RSA-SHA224, 
SIGN-RSA-SHA256, SIGN-RSA-SHA384, SIGN-RSA-SHA512, SIGN-DSA-SHA1, 
SIGN-DSA-SHA224, SIGN-DSA-SHA256, SIGN-RSA-MD5. Catch all
is SIGN-ALL. This is only valid for TLS 1.2 and later.

@item Elliptic curves @tab
CURVE-SECP192R1, CURVE-SECP224R1, CURVE-SECP256R1, CURVE-SECP384R1, CURVE-SECP521R1. Catch all is CURVE-ALL.

@end multitable
@caption{The supported algorithm keywords in priority strings.}
@end float

Note that the DHE key exchange methods are generally
slower@footnote{It depends on the group used.  Primes with
lesser bits are always faster, but also easier to break.  See @ref{Selecting cryptographic key sizes}
for the acceptable security levels.} than their elliptic curves counterpart
(ECDHE). Moreover the plain Diffie-Hellman key exchange
requires parameters to be generated and associated with a credentials
structure by the server (see @ref{Parameter generation}). 

The available special keywords are shown in @ref{tab:prio-special}. 

@float Table,tab:prio-special
@multitable @columnfractions .45 .45
@headitem Keyword @tab Description

@item %COMPAT @tab
will enable compatibility mode. It might mean that violations
of the protocols are allowed as long as maximum compatibility with
problematic clients and servers is achieved. More specifically this
string would disable TLS record random padding and tolerate packets
over the maximum allowed TLS record.

@item %NO_EXTENSIONS @tab
will prevent the sending of any TLS extensions in client side. Note
that TLS 1.2 requires extensions to be used, as well as safe
renegotiation thus this option must be used with care.

@item %SERVER_PRECEDENCE @tab
The ciphersuite will be selected according to server priorities
and not the client's.

@item %DISABLE_SAFE_RENEGOTIATION @tab
will disable safe renegotiation
completely.  Do not use unless you know what you are doing.
Testing purposes only.

@item %UNSAFE_RENEGOTIATION @tab
will allow handshakes and re-handshakes
without the safe renegotiation extension.  Note that for clients
this mode is insecure (you may be under attack), and for servers it
will allow insecure clients to connect (which could be fooled by an
attacker).  Do not use unless you know what you are doing and want
maximum compatibility.

@item %PARTIAL_RENEGOTIATION @tab
will allow initial handshakes to proceed,
but not re-handshakes.  This leaves the client vulnerable to attack,
and servers will be compatible with non-upgraded clients for
initial handshakes.  This is currently the default for clients and
servers, for compatibility reasons.

@item %SAFE_RENEGOTIATION @tab
will enforce safe renegotiation.  Clients and
servers will refuse to talk to an insecure peer.  Currently this
causes interoperability problems, but is required for full protection.

@item %SSL3_RECORD_VERSION @tab
will use SSL3.0 record version in client hello.
This is the default.

@item %LATEST_RECORD_VERSION @tab
will use the latest TLS version record version in client hello.

@item %VERIFY_ALLOW_SIGN_RSA_MD5 @tab
will allow RSA-MD5 signatures in certificate chains.

@item %VERIFY_ALLOW_X509_V1_CA_CRT @tab
will allow V1 CAs in chains.

@end multitable
@caption{Special priority string keywords.}
@end float

Finally the ciphersuites enabled by any priority string can be
listed using the @code{gnutls-cli} application (see @ref{gnutls-cli Invocation}), 
or by using the priority functions as in @ref{Listing the ciphersuites in a priority string}.

Example priority strings are:
@example
The default priority without the HMAC-MD5:
    "NORMAL:-MD5"

Specifying RSA with AES-128-CBC:
    "NONE:+VERS-TLS-ALL:+MAC-ALL:+RSA:+AES-128-CBC:+SIGN-ALL:+COMP-NULL"

Specifying the defaults except ARCFOUR-128:
    "NORMAL:-ARCFOUR-128"

Enabling the 128-bit secure ciphers, while disabling SSL 3.0 and 
enabling compression:
    "SECURE128:-VERS-SSL3.0:+COMP-DEFLATE"
@end example

@node Advanced and other topics
@section Advanced and other topics

@menu
* Session resumption::
* Parameter generation::
* Keying Material Exporters::
* Channel Bindings::
* Interoperability::
* Compatibility with the OpenSSL library::
@end menu

@node Session resumption
@subsection Session resumption
@cindex resuming sessions
@cindex session resumption

@subsubheading Client side

To reduce time and roundtrips spent in a handshake the client can   
request session resumption from a server that previously shared
a session with. For that the client has to retrieve and store
the session parameters. Before establishing a new session to the same 
server the parameters must be re-associated with the GnuTLS session
using @funcref{gnutls_session_set_data}.

@showfuncC{gnutls_session_get_data,gnutls_session_get_id,gnutls_session_set_data}

Keep in mind that sessions will be expired after some time, depending
on the server, and a server may choose not to resume a session
even when requested to.  The expiration is to prevent temporal session keys
from becoming long-term keys. Also note that as a client you must enable, 
using the priority functions, at least the algorithms used in the last session.

It is highly recommended for clients to enable the session ticket extension using
@funcref{gnutls_session_ticket_enable_client} in order to allow resumption with 
servers that do not store any state.

@showfuncA{gnutls_session_ticket_enable_client}

@showfuncdesc{gnutls_session_is_resumed}

@subsubheading Server side

In order to support resumption a server can store
the session security parameters in a local database or by using session
tickets (see @ref{Session tickets}) to delegate storage to the client. Because
session tickets might not be supported by all clients, servers
could combine the two methods.

A storing server needs to specify callback functions to store, retrieve and delete session data. These can be
registered with the functions below. The stored sessions in the database can be checked using @funcref{gnutls_db_check_entry}
for expiration.

@showfuncD{gnutls_db_set_retrieve_function,gnutls_db_set_store_function,gnutls_db_set_ptr,gnutls_db_set_remove_function}
@showfuncA{gnutls_db_check_entry}

A server utilizing tickets should generate ticket encryption
and authentication keys using @funcref{gnutls_session_ticket_key_generate}.
Those keys should be associated with the GnuTLS session using
@funcref{gnutls_session_ticket_enable_server}.

@showfuncdesc{gnutls_session_ticket_enable_server}
@showfuncdesc{gnutls_session_ticket_key_generate}
@showfuncdesc{gnutls_session_resumption_requested}

A server enabling both session tickets and a storage for session data
would use session tickets when clients support it and the storage otherwise.

@node Parameter generation
@subsection Parameter generation
@cindex parameter generation
@cindex generating parameters

Several TLS ciphersuites require additional parameters that
need to be generated or provided by the application. The
Diffie-Hellman based ciphersuites (ANON-DH or DHE), require
the group parameters to be provided. Those can either be
be generated on the fly using @funcref{gnutls_dh_params_generate2}
or imported from pregenerated data using @funcref{gnutls_dh_params_import_pkcs3}.
The parameters can be used in a @acronym{TLS} session by calling
@funcref{gnutls_certificate_set_dh_params} or
@funcref{gnutls_anon_set_server_dh_params} for anonymous sessions.

@showfuncD{gnutls_dh_params_generate2,gnutls_dh_params_import_pkcs3,gnutls_certificate_set_dh_params,gnutls_anon_set_server_dh_params}

Due to the time-consuming calculations required for the generation
of Diffie-Hellman parameters we suggest against performing generation
of them within an application. The @code{certtool} tool can be used to 
generate or export known safe values that can be stored in code
or in a configuration file to provide the ability to replace. We also
recommend the usage of @funcref{gnutls_sec_param_to_pk_bits} 
(see @ref{Selecting cryptographic key sizes}) to determine
the bit size of the generated parameters.

Note that the information stored in the generated PKCS #3 structure
changed with GnuTLS 3.0.9. Since that version the @code{privateValueLength}
member of the structure is set, allowing the server utilizing the
parameters to use keys of the size of the security parameter. This
provides better performance in key exchange.

The ciphersuites that involve the RSA-EXPORT key exchange require
additional parameters. Those ciphersuites are rarely used today
because they are by design insecure, thus if you have no requirement
for them, the rest of this section can be skipped. The RSA-EXPORT key exchange
requires 512-bit RSA keys to be generated. It is recommended those
parameters to be refreshed (regenerated) in short intervals. The
following functions can be used for these parameters.

@showfuncD{gnutls_rsa_params_generate2,gnutls_certificate_set_rsa_export_params,gnutls_rsa_params_import_pkcs1,gnutls_rsa_params_export_pkcs1}

To allow renewal of the parameters within an application without
accessing the credentials, which are a shared structure,
an alternative interface is available using a callback function.  

@showfuncdesc{gnutls_certificate_set_params_function}


@node Keying Material Exporters
@subsection Keying material exporters
@cindex keying material exporters
@cindex exporting keying material

The TLS PRF can be used by other protocols to derive keys based on
the TLS master secret.  The API to use is @funcref{gnutls_prf}.  The 
function needs to be provided with the label in the parameter 
@code{label}, and the extra data to mix in the
@code{extra} parameter.  Depending on whether you want to mix in the
client or server random data first, you can set the
@code{server_random_first} parameter.

For example, after establishing a TLS session using
@funcref{gnutls_handshake}, you can invoke the TLS PRF with this call:

@example
#define MYLABEL "EXPORTER-FOO"
#define MYCONTEXT "some context data"
char out[32];
rc = gnutls_prf (session, strlen (MYLABEL), MYLABEL, 0,
                 strlen (MYCONTEXT), MYCONTEXT, 32, out);
@end example

If you don't want to mix in the client/server random, there is a 
low-level TLS PRF interface called @funcref{gnutls_prf_raw}.

@node Channel Bindings
@subsection Channel bindings
@cindex channel bindings

In user authentication protocols (e.g., EAP or SASL mechanisms) it is
useful to have a unique string that identifies the secure channel that
is used, to bind together the user authentication with the secure
channel.  This can protect against man-in-the-middle attacks in some
situations.  That unique string is called a ``channel binding''.  For
background and discussion see @xcite{RFC5056}.

In @acronym{GnuTLS} you can extract a channel binding using the
@funcref{gnutls_session_channel_binding} function.  Currently only the
type @code{GNUTLS_CB_TLS_UNIQUE} is supported, which corresponds to
the @code{tls-unique} channel binding for TLS defined in
@xcite{RFC5929}.

The following example describes how to print the channel binding data.
Note that it must be run after a successful TLS handshake.

@example
@{
  gnutls_datum_t cb;
  int rc;

  rc = gnutls_session_channel_binding (session,
                                       GNUTLS_CB_TLS_UNIQUE,
                                       &cb);
  if (rc)
    fprintf (stderr, "Channel binding error: %s\n",
             gnutls_strerror (rc));
  else
    @{
      size_t i;
      printf ("- Channel binding 'tls-unique': ");
      for (i = 0; i < cb.size; i++)
        printf ("%02x", cb.data[i]);
      printf ("\n");
    @}
@}
@end example

@node Interoperability
@subsection Interoperability

The @acronym{TLS} protocols support many ciphersuites, extensions and version
numbers. As a result, few implementations are 
not able to properly interoperate once faced with extensions or version protocols
they do not support and understand. The @acronym{TLS} protocol allows for a
graceful downgrade to the commonly supported options, but practice shows 
it is not always implemented correctly. 

Because there is no way to achieve maximum interoperability with broken peers
without sacrificing security, @acronym{GnuTLS} ignores such peers by default. 
This might not be acceptable in cases where maximum compatibility
is required. Thus we allow enabling compatibility with broken peers using
priority strings (see @ref{Priority Strings}). A conservative priority
string that would disable certain @acronym{TLS} protocol
options that are known to cause compatibility problems, is shown below. 
@verbatim
NORMAL:%COMPAT
@end verbatim

For broken peers that do not tolerate TLS version numbers over TLS 1.0
another priority string is:
@verbatim
NORMAL:-VERS-TLS-ALL:+VERS-TLS1.0:+VERS-SSL3.0:%COMPAT
@end verbatim
This priority string will in addition to above, only enable SSL 3.0 and 
TLS 1.0 as protocols. Note however that
there are known attacks against those protocol versions, especially over
the CBC-mode ciphersuites. To mitigate them another priority string
that only allows the stream cipher ARCFOUR is below.
@verbatim
NORMAL:-VERS-TLS-ALL:+VERS-TLS1.0:+VERS-SSL3.0:-CIPHER-ALL:+ARCFOUR-128:%COMPAT
@end verbatim

@node Compatibility with the OpenSSL library
@subsection Compatibility with the OpenSSL library
@cindex OpenSSL

To ease @acronym{GnuTLS}' integration with existing applications, a
compatibility layer with the OpenSSL library is included
in the @code{gnutls-openssl} library. This compatibility layer is not
complete and it is not intended to completely re-implement the OpenSSL
API with @acronym{GnuTLS}.  It only provides limited source-level
compatibility. 

The prototypes for the compatibility functions are in the
@file{gnutls/openssl.h} header file. The limitations 
imposed by the compatibility layer include:

@itemize

@item Error handling is not thread safe.

@end itemize

@node Using the cryptographic library
@section Using the cryptographic library

@acronym{GnuTLS} is not a low-level cryptographic library, i.e., 
it does not provide access to basic cryptographic primitives. However
it abstracts the internal cryptographic back-end (see @ref{Cryptographic Backend}),
providing symmetric crypto, hash and HMAC algorithms, as well access
to the random number generation.

@menu
* Symmetric cryptography::
* Hash and HMAC functions::
* Random number generation::
@end menu

@node Symmetric cryptography
@subsection Symmetric cryptography
@cindex symmetric cryptography

The available functions to access symmetric crypto algorithms operations
are shown below. The supported algorithms are the algorithms required by the TLS protocol.
They are listed in @ref{tab:ciphers}.

@showfuncE{gnutls_cipher_init,gnutls_cipher_encrypt2,gnutls_cipher_decrypt2,gnutls_cipher_set_iv,gnutls_cipher_deinit}

In order to support authenticated encryption with associated data (AEAD) algorithms the following
functions are provided to set the associated data and retrieve the authentication tag.

@showfuncB{gnutls_cipher_add_auth,gnutls_cipher_tag}

@node Hash and HMAC functions
@subsection Hash and HMAC functions
@cindex hash functions
@cindex HMAC functions

The available operations to access hash functions and hash-MAC (HMAC) algorithms
are shown below. HMAC algorithms provided keyed hash functionality. They supported HMAC algorithms are listed in @ref{tab:macs}.

@showfuncF{gnutls_hmac_init,gnutls_hmac,gnutls_hmac_output,gnutls_hmac_deinit,gnutls_hmac_get_len,gnutls_hmac_fast}

The available functions to access hash functions are shown below. The supported hash functions
are the same as the HMAC algorithms.

@showfuncF{gnutls_hash_init,gnutls_hash,gnutls_hash_output,gnutls_hash_deinit,gnutls_hash_get_len,gnutls_hash_fast}

@node Random number generation
@subsection Random number generation
@cindex random numbers

Access to the random number generator is provided using the @funcref{gnutls_rnd}
function. It allows obtaining random data of various levels.

@showenumdesc{gnutls_rnd_level_t,The random number levels.}
@showfuncdesc{gnutls_rnd}

@node Selecting cryptographic key sizes
@section Selecting cryptographic key sizes
@cindex key sizes

Because many algorithms are involved in TLS, it is not easy to set
a consistent security level.  For this reason in @ref{tab:key-sizes} we
present some correspondence between key sizes of symmetric algorithms
and public key algorithms based on @xcite{ECRYPT}. 
Those can be used to generate certificates with
appropriate key sizes as well as select parameters for Diffie-Hellman and SRP
authentication.

@float Table,tab:key-sizes
@multitable @columnfractions .10 .12 .10 .20 .32

@headitem Security bits @tab RSA, DH and SRP parameter size @tab ECC key size @tab Security parameter @tab Description

@item 80
@tab 1248
@tab 160
@tab @code{LOW}
@tab Very short term protection against agencies

@item 96
@tab 1776
@tab 192
@tab @code{LEGACY}
@tab Legacy standard level

@item 112
@tab 2432
@tab 224
@tab @code{NORMAL}
@tab Medium-term protection

@item 128
@tab 3248
@tab 256
@tab @code{HIGH}
@tab Long term protection

@item 256
@tab 15424
@tab 512
@tab @code{ULTRA}
@tab Foreseeable future

@end multitable
@caption{Key sizes and security parameters.}
@end float

The first column  provides a security parameter in a number of bits. This
gives an indication of the number of combinations to be tried by an adversary
to brute force a key. For example to test all possible keys in a 112 bit security parameter
@math{2^{112}} combinations have to be tried. For today's technology this is infeasible.
The next two columns correlate the security
parameter with actual bit sizes of parameters for DH, RSA, SRP and ECC algorithms.
A mapping to @code{gnutls_sec_param_t} value is given for each security parameter, on
the next column, and finally a brief description of the level.

@c @showenumdesc{gnutls_sec_param_t,The @code{gnutls_sec_@-param_t} enumeration.}

Note, however, that the values suggested here are nothing more than an
educated guess that is valid today. There are no guarantees that an
algorithm will remain unbreakable or that these values will remain
constant in time. There could be scientific breakthroughs that cannot
be predicted or total failure of the current public key systems by
quantum computers. On the other hand though the cryptosystems used in
TLS are selected in a conservative way and such catastrophic
breakthroughs or failures are believed to be unlikely.
The NIST publication SP 800-57 @xcite{NISTSP80057} contains a similar
table.

When using @acronym{GnuTLS} and a decision on bit sizes for a public
key algorithm is required, use of the following functions is  
recommended:

@showfuncdesc{gnutls_sec_param_to_pk_bits}

@showfuncdesc{gnutls_pk_bits_to_sec_param}

Those functions will convert a human understandable security parameter
of @code{gnutls_sec_param_t} type, to a number of bits suitable for a public 
key algorithm.

The following functions will set the minimum acceptable group size for Diffie-Hellman
and SRP authentication. 
@showfuncB{gnutls_dh_set_prime_bits,gnutls_srp_set_prime_bits}