\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 <