\input texinfo @c -*- texinfo -*-
@c %**start of header
@c $Id$
@setfilename hx509.info
@settitle HX509
@iftex
@afourpaper
@end iftex
@c some sensible characters, please?
@documentencoding UTF-8
@setchapternewpage on
@syncodeindex pg cp
@c %**end of header
@include vars.texi
@set VERSION @value{PACKAGE_VERSION}
@set EDITION 1.0
@ifinfo
@dircategory Security
@direntry
* hx509: (hx509). The X.509 distribution from KTH
@end direntry
@end ifinfo
@c title page
@titlepage
@title HX509
@subtitle X.509 distribution from KTH
@subtitle Edition @value{EDITION}, for version @value{VERSION}
@subtitle 2008
@author Love Hörnquist Åstrand
@iftex
@def@copynext{@vskip 20pt plus 1fil}
@def@copyrightstart{}
@def@copyrightend{}
@end iftex
@ifnottex
@macro copynext
@end macro
@macro copyrightstart
@end macro
@macro copyrightend
@end macro
@end ifnottex
@page
@copyrightstart
Copyright (c) 1994-2019 Kungliga Tekniska Högskolan
(Royal Institute of Technology, Stockholm, Sweden).
All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
1. Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
2. Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
3. Neither the name of the Institute nor the names of its contributors
may be used to endorse or promote products derived from this software
without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE INSTITUTE AND CONTRIBUTORS ``AS IS'' AND
ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
ARE DISCLAIMED. IN NO EVENT SHALL THE INSTITUTE OR CONTRIBUTORS BE LIABLE
FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
SUCH DAMAGE.
@copynext
Copyright (c) 1988, 1990, 1993
The Regents of the University of California. All rights reserved.
Redistribution and use in source and binary forms, with or without
modification, are permitted provided that the following conditions
are met:
1. Redistributions of source code must retain the above copyright
notice, this list of conditions and the following disclaimer.
2. Redistributions in binary form must reproduce the above copyright
notice, this list of conditions and the following disclaimer in the
documentation and/or other materials provided with the distribution.
3. Neither the name of the University nor the names of its contributors
may be used to endorse or promote products derived from this software
without specific prior written permission.
THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
SUCH DAMAGE.
@copynext
Copyright 1992 Simmule Turner and Rich Salz. All rights reserved.
This software is not subject to any license of the American Telephone
and Telegraph Company or of the Regents of the University of California.
Permission is granted to anyone to use this software for any purpose on
any computer system, and to alter it and redistribute it freely, subject
to the following restrictions:
1. The authors are not responsible for the consequences of use of this
software, no matter how awful, even if they arise from flaws in it.
2. The origin of this software must not be misrepresented, either by
explicit claim or by omission. Since few users ever read sources,
credits must appear in the documentation.
3. Altered versions must be plainly marked as such, and must not be
misrepresented as being the original software. Since few users
ever read sources, credits must appear in the documentation.
4. This notice may not be removed or altered.
@copynext
IMath is Copyright 2002-2005 Michael J. Fromberger
You may use it subject to the following Licensing Terms:
Permission is hereby granted, free of charge, to any person obtaining
a copy of this software and associated documentation files (the
"Software"), to deal in the Software without restriction, including
without limitation the rights to use, copy, modify, merge, publish,
distribute, sublicense, and/or sell copies of the Software, and to
permit persons to whom the Software is furnished to do so, subject to
the following conditions:
The above copyright notice and this permission notice shall be
included in all copies or substantial portions of the Software.
THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY
CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
@copyrightend
@end titlepage
@macro manpage{man, section}
@cite{\man\(\section\)}
@end macro
@c Less filling! Tastes great!
@iftex
@parindent=0pt
@global@parskip 6pt plus 1pt
@global@chapheadingskip = 15pt plus 4pt minus 2pt
@global@secheadingskip = 12pt plus 3pt minus 2pt
@global@subsecheadingskip = 9pt plus 2pt minus 2pt
@end iftex
@ifinfo
@paragraphindent 0
@end ifinfo
@ifnottex
@node Top, Introduction, (dir), (dir)
@top Heimdal
@end ifnottex
This manual is for version @value{VERSION} of hx509.
@menu
* Introduction::
* What are X.509 and PKIX ?::
* Setting up a CA::
* CMS signing and encryption::
* Certificate matching::
* Software PKCS 11 module::
* Creating a CA certificate::
* Issuing certificates::
* Issuing CRLs::
* Application requirements::
* CMS background::
* Matching syntax::
* How to use the PKCS11 module::
@detailmenu
--- The Detailed Node Listing ---
Setting up a CA
@c * Issuing certificates::
* Creating a CA certificate::
* Issuing certificates::
* Issuing CRLs::
@c * Issuing a proxy certificate::
@c * Creating a user certificate::
@c * Validating a certificate::
@c * Validating a certificate path::
* Application requirements::
CMS signing and encryption
* CMS background::
Certificate matching
* Matching syntax::
Software PKCS 11 module
* How to use the PKCS11 module::
@end detailmenu
@end menu
@node Introduction, What are X.509 and PKIX ?, Top, Top
@chapter Introduction
A Public Key Infrastructure (PKI) is an authentication mechanism based on
entities having certified cryptographic public keys and corresponding private
(secret) keys.
The ITU-T PKI specifications are designated "x.509", while the IETF PKI
specifications (PKIX) are specified by a number of Internet RFCs and are based
on x.509.
The goals of a PKI (as stated in
RFC 5280) is to meet
@emph{the needs of deterministic, automated identification, authentication, access control, and authorization}.
The administrator should be aware of certain terminologies as explained by the aforementioned
RFC before attemping to put in place a PKI infrastructure. Briefly, these are:
@itemize @bullet
@item CA
Certificate Authority
@item RA
Registration Authority, i.e., an optional system to which a CA delegates certain management functions.
@item Certificate
A binary document that names an entity and its public key and which is signed
by an issuing CA.
@item CRL Issuer
An optional system to which a CA delegates the publication of certificate revocation lists.
@item Repository
A system or collection of distributed systems that stores certificates and CRLs
and serves as a means of distributing these certificates and CRLs to end entities
@end itemize
hx509 (Heimdal x509 support) is a near complete X.509/PKIX stack that can
handle CMS messages (crypto system used in S/MIME and Kerberos PK-INIT)
and basic certificate processing tasks, path construction, path
validation, OCSP and CRL validation, PKCS10 message construction, CMS
Encrypted (shared secret encrypted), CMS SignedData (certificate
signed), and CMS EnvelopedData (certificate encrypted).
hx509 can use PKCS11 tokens, PKCS12 files, PEM files, and/or DER encoded
files.
hx509 consists of a library (libhx509) and a command-line utility (hxtool), as
well as a RESTful, HTTPS-based service that implements an online CA.
@node What are X.509 and PKIX ?, Setting up a CA, Introduction, Top
@chapter What are X.509 and PKIX, PKIX, PKCS7 and CMS ?
X.509 was created by CCITT (later ITU-T) for the X.500 directory
service. Today, X.509 discussions and implementations commonly reference
the IETF's PKIX Certificate and CRL Profile of the X.509 v3 certificate
standard, as specified in RFC 3280.
ITU continues to develop the X.509 standard together with the IETF in a
rather complicated dance.
X.509 is a public key based security system that has associated data
stored within a so called certificate. Initially, X.509 was a strict
hierarchical system with one root. However, ever evolving requiments and
technology advancements saw the inclusion of multiple policy roots,
bridges and mesh solutions.
x.509 can also be used as a peer to peer system, though often seen as a
common scenario.
@section Type of certificates
There are several flavors of certificate in X.509.
@itemize @bullet
@item Trust anchors
Trust anchors are strictly not certificates, but commonly stored in a
certificate format as they become easier to manage. Trust anchors are
the keys that an end entity would trust to validate other certificates.
This is done by building a path from the certificate you want to
validate to to any of the trust anchors you have.
@item End Entity (EE) certificates
End entity certificates are the most common types of certificates. End
entity certificates cannot issue (sign) certificate themselves and are generally
used to authenticate and authorize users and services.
@item Certification Authority (CA) certificates
Certificate authority certificates have the right to issue additional
certificates (be it sub-ordinate CA certificates to build an trust anchors
or end entity certificates). There is no limit to how many certificates a CA
may issue, but there might other restrictions, like the maximum path
depth.
@item Proxy certificates
Remember the statement "End Entity certificates cannot issue
certificates"? Well that statement is not entirely true. There is an
extension called proxy certificates defined in RFC3820, that allows
certificates to be issued by end entity certificates. The service that
receives the proxy certificates must have explicitly turned on support
for proxy certificates, so their use is somewhat limited.
Proxy certificates can be limited by policies stored in the certificate to
what they can be used for. This allows users to delegate the proxy
certificate to services (by sending over the certificate and private
key) so the service can access services on behalf of the user.
One example of this would be a print service. The user wants to print a
large job in the middle of the night when the printer isn't used that
much, so the user creates a proxy certificate with the policy that it
can only be used to access files related to this print job, creates the
print job description and send both the description and proxy
certificate with key over to print service. Later at night when the
print service initializes (without any user intervention), access to the files
for the print job is granted via the proxy certificate. As a result of (in-place)
policy limitations, the certificate cannot be used for any other purposes.
@end itemize
@section Building a path
Before validating a certificate path (or chain), the path needs to be
constructed. Given a certificate (EE, CA, Proxy, or any other type),
the path construction algorithm will try to find a path to one of the
trust anchors.
The process starts by looking at the issuing CA of the certificate, by
Name or Key Identifier, and tries to find that certificate while at the
same time evaluting any policies in-place.
@node Setting up a CA, Creating a CA certificate, What are X.509 and PKIX ?, Top
@chapter Setting up a CA
Do not let information overload scare you off! If you are simply testing
or getting started with a PKI infrastructure, skip all this and go to
the next chapter (see: @pxref{Creating a CA certificate}).
Creating a CA certificate should be more the just creating a
certificate, CA's should define a policy. Again, if you are simply
testing a PKI, policies do not matter so much. However, when it comes to
trust in an organisation, it will probably matter more whom your users
and sysadmins will find it acceptable to trust.
At the same time, try to keep things simple, it's not very hard to run a
Certificate authority and the process to get new certificates should be simple.
You may find it helpful to answer the following policy questions for
your organization at a later stage:
@itemize @bullet
@item How do you trust your CA.
@item What is the CA responsibility.
@item Review of CA activity.
@item How much process should it be to issue certificate.
@item Who is allowed to issue certificates.
@item Who is allowed to requests certificates.
@item How to handle certificate revocation, issuing CRLs and maintain OCSP services.
@end itemize
@node Creating a CA certificate, Issuing certificates, Setting up a CA, Top
@section Creating a CA certificate
This section describes how to create a CA certificate and what to think
about.
@subsection Lifetime CA certificate
You probably want to create a CA certificate with a long lifetime, 10
years at the very minimum. This is because you don't want to push out the
certificate (as a trust anchor) to all you users again when the old
CA certificate expires. Although a trust anchor can't really expire, not all
software works in accordance with published standards.
Keep in mind the security requirements might be different 10-20 years
into the future. For example, SHA1 is going to be withdrawn in 2010, so
make sure you have enough buffering in your choice of digest/hash
algorithms, signature algorithms and key lengths.
@subsection Create a CA certificate
This command below can be used to generate a self-signed CA certificate.
@example
hxtool issue-certificate \
--self-signed \
--issue-ca \
--generate-key=rsa \
--subject="CN=CertificateAuthority,DC=test,DC=h5l,DC=se" \
--lifetime=10years \
--certificate="FILE:ca.pem"
@end example
@subsection Extending the lifetime of a CA certificate
You just realised that your CA certificate is going to expire soon and
that you need replace it with a new CA. The easiest way to do that
is to extend the lifetime of your existing CA certificate.
The example below will extend the CA certificate's lifetime by 10 years.
You should compare this new certificate if it contains all the
special tweaks as the old certificate had.
@example
hxtool issue-certificate \
--self-signed \
--issue-ca \
--lifetime="10years" \
--template-certificate="FILE:ca.pem" \
--template-fields="serialNumber,notBefore,subject,SPKI" \
--ca-private-key=FILE:ca.pem \
--certificate="FILE:new-ca.pem"
@end example
@subsection Subordinate CA
This example below creates a new subordinate certificate authority.
@example
hxtool issue-certificate \
--ca-certificate=FILE:ca.pem \
--issue-ca \
--generate-key=rsa \
--subject="CN=CertificateAuthority,DC=dev,DC=test,DC=h5l,DC=se" \
--certificate="FILE:dev-ca.pem"
@end example
@node Issuing certificates, Issuing CRLs, Creating a CA certificate, Top
@section Issuing certificates
First you'll create a CA certificate, after that you have to deal with
your users and servers and issue certificates to them.
@c I think this section needs a bit of clarity. Can I add a separate
@c section which explains CSRs as well?
@itemize @bullet
@item Do all the work themself
Generate the key for the user. This has the problme that the the CA
knows the private key of the user. For a paranoid user this might leave
feeling of disconfort.
@item Have the user do part of the work
Receive PKCS10 certificate requests fromusers. PKCS10 is a request for a
certificate. The user may specify what DN they want as well as provide
a certificate signing request (CSR). To prove the user have the key,
the whole request is signed by the private key of the user.
@end itemize
@subsection Name space management
@c The explanation given below is slightly unclear. I will re-read the
@c RFC and document accordingly
What people might want to see.
Re-issue certificates just because people moved within the organization.
Expose privacy information.
Using Sub-component name (+ notation).
@subsection Certificate Revocation, CRL and OCSP
Certificates that a CA issues may need to be revoked at some stage. As
an example, an employee leaves the organization and does not bother
handing in his smart card (or even if the smart card is handed back --
the certificate on it must no longer be acceptable to services; the
employee has left).
You may also want to revoke a certificate for a service which is no
longer being offered on your network. Overlooking these scenarios can
lead to security holes which will quickly become a nightmare to deal
with.
There are two primary protocols for dealing with certificate
revokation. Namely:
@itemize @bullet
@item Certificate Revocation List (CRL)
@item Online Certificate Status Protocol (OCSP)
@end itemize
If however the certificate in qeustion has been destroyed, there is no
need to revoke the certificate because it can not be used by someone
else. This matter since for each certificate you add to CRL, the
download time and processing time for clients are longer.
CRLs and OCSP responders however greatly help manage compatible services
which may authenticate and authorize users (or services) on an on-going
basis. As an example, VPN connectivity established via certificates for
connecting clients would require your VPN software to make use of a CRL
or an OCSP service to ensure revoked certificates belonging to former
clients are not allowed access to (formerly subscribed) network
services.
@node Issuing CRLs, Application requirements, Issuing certificates, Top
@section Issuing CRLs
Create an empty CRL with no certificates revoked. Default expiration
value is one year from now.
@example
hxtool crl-sign \
--crl-file=crl.der \
--signer=FILE:ca.pem
@end example
Create a CRL with all certificates in the directory
@file{/path/to/revoked/dir} included in the CRL as revoked. Also make
it expire one month from now.
@example
hxtool crl-sign \
--crl-file=crl.der \
--signer=FILE:ca.pem \
--lifetime='1 month' \
DIR:/path/to/revoked/dir
@end example
@node Application requirements, CMS signing and encryption, Issuing CRLs, Top
@section Application requirements
Application place different requirements on certificates. This section
tries to expand what they are and how to use hxtool to generate
certificates for those services.
@subsection HTTPS - server
@example
hxtool issue-certificate \
--subject="CN=www.test.h5l.se,DC=test,DC=h5l,DC=se" \
--type="https-server" \
--hostname="www.test.h5l.se" \
--hostname="www2.test.h5l.se" \
...
@end example
@subsection HTTPS - client
@example
hxtool issue-certificate \
--subject="UID=testus,DC=test,DC=h5l,DC=se" \
--type="https-client" \
...
@end example
@subsection S/MIME - email
There are two things that should be set in S/MIME certificates, one or
more email addresses and an extended eku usage (EKU), emailProtection.
The email address format used in S/MIME certificates is defined in
RFC2822, section 3.4.1 and it should be an ``addr-spec''.
There are two ways to specifify email address in certificates. The old
way is in the subject distinguished name, @emph{this should not be used}. The
new way is using a Subject Alternative Name (SAN).
Even though the email address is stored in certificates, they don't need
to be, email reader programs are required to accept certificates that
doesn't have either of the two methods of storing email in certificates
-- in which case, the email client will try to protect the user by
printing the name of the certificate instead.
S/MIME certificate can be used in another special way. They can be
issued with a NULL subject distinguished name plus the email in SAN,
this is a valid certificate. This is used when you wont want to share
more information then you need to.
hx509 issue-certificate supports adding the email SAN to certificate by
using the --email option, --email also gives an implicit emailProtection
eku. If you want to create an certificate without an email address, the
option --type=email will add the emailProtection EKU.
@example
hxtool issue-certificate \
--subject="UID=testus-email,DC=test,DC=h5l,DC=se" \
--type=email \
--email="testus@@test.h5l.se" \
...
@end example
An example of an certificate without and subject distinguished name with
an email address in a SAN.
@example
hxtool issue-certificate \
--subject="" \
--type=email \
--email="testus@@test.h5l.se" \
...
@end example
@subsection PK-INIT
A PK-INIT infrastructure allows users and services to pick up kerberos
credentials (tickets) based on their certificate. This, for example,
allows users to authenticate to their desktops using smartcards while
acquiring kerberos tickets in the process.
As an example, an office network which offers centrally controlled
desktop logins, mail, messaging (xmpp) and openafs would give users
single sign-on facilities via smartcard based logins. Once the kerberos
ticket has been acquired, all kerberized services would immediately
become accessible based on deployed security policies.
Let's go over the process of initializing a demo PK-INIT framework:
@example
hxtool issue-certificate \
--type="pkinit-kdc" \
--pk-init-principal="krbtgt/TEST.H5L.SE@@TEST.H5L.SE" \
--hostname=kerberos.test.h5l.se \
--ca-certificate="FILE:ca.pem,ca.key" \
--generate-key=rsa \
--certificate="FILE:kdc.pem" \
--subject="cn=kdc"
@end example
How to create a certificate for a user.
@example
hxtool issue-certificate \
--type="pkinit-client" \
--pk-init-principal="user@@TEST.H5L.SE" \
--ca-certificate="FILE:ca.pem,ca.key" \
--generate-key=rsa \
--subject="cn=Test User" \
--certificate="FILE:user.pem"
@end example
The --type field can be specified multiple times. The same certificate
can hence house extensions for both pkinit-client as well as S/MIME.
To use the PKCS11 module, please see the section:
@pxref{How to use the PKCS11 module}.
More about how to configure the KDC, see the documentation in the
Heimdal manual to set up the KDC.
@subsection XMPP/Jabber
The jabber server certificate should have a dNSname that is the same as
the user entered into the application, not the same as the host name of
the machine.
@example
hxtool issue-certificate \
--subject="CN=xmpp1.test.h5l.se,DC=test,DC=h5l,DC=se" \
--hostname="xmpp1.test.h5l.se" \
--hostname="test.h5l.se" \
...
@end example
The certificate may also contain a jabber identifier (JID) that, if the
receiver allows it, authorises the server or client to use that JID.
When storing a JID inside the certificate, both for server and client,
it's stored inside a UTF8String within an otherName entity inside the
subjectAltName, using the OID id-on-xmppAddr (1.3.6.1.5.5.7.8.5).
To read more about the requirements, see RFC3920, Extensible Messaging
and Presence Protocol (XMPP): Core.
hxtool issue-certificate have support to add jid to the certificate
using the option @kbd{--jid}.
@example
hxtool issue-certificate \
--subject="CN=Love,DC=test,DC=h5l,DC=se" \
--jid="lha@@test.h5l.se" \
...
@end example
@node CMS signing and encryption, CMS background, Application requirements, Top
@chapter CMS signing and encryption
CMS is the Cryptographic Message System that among other, is used by
S/MIME (secure email) and Kerberos PK-INIT. It's an extended version of
the RSA, Inc standard PKCS7.
@node CMS background, Certificate matching, CMS signing and encryption, Top
@section CMS background
@node Certificate matching, Matching syntax, CMS background, Top
@chapter Certificate matching
To match certificates hx509 have a special query language to match
certifictes in queries and ACLs.
@node Matching syntax, Software PKCS 11 module, Certificate matching, Top
@section Matching syntax
This is the language definitions somewhat slopply descriped:
@example
expr = TRUE,
FALSE,
! expr,
expr AND expr,
expr OR expr,
( expr )
compare
compare =
word == word,
word != word,
word IN ( word [, word ...])
word IN %@{variable.subvariable@}
word =
STRING,
%@{variable@}
@end example
@node Software PKCS 11 module, How to use the PKCS11 module, Matching syntax, Top
@chapter Software PKCS 11 module
PKCS11 is a standard created by RSA, Inc to support hardware and
software encryption modules. It can be used by smartcard to expose the
crypto primitives inside without exposing the crypto keys.
Hx509 includes a software implementation of PKCS11 that runs within the
memory space of the process and thus exposes the keys to the
application.
@node How to use the PKCS11 module, , Software PKCS 11 module, Top
@section How to use the PKCS11 module
@example
$ cat > ~/.soft-pkcs11.rc <