Kerberos for Windows (or KfW for short) is the integrated release of Kerberos version 4 and Kerberos version 5 (or Kerberos 4 and Kerberos 5 for short -- also Kerberos v4 and Kerberos v5).
Kerberos 4 and Kerberos 5 refers to the versions 4 and 5 of the Kerberos protocol. (A protocol refers to how data is transmitted on a network.) KfW includes both of these. KfW 2.1.1, in particular, includes the MIT Kerberos v5 Release 1.2.2 and KfW's own release of Kerberos v4.
Each Windows NT/2000 login session can access only its own credentials cache. Even Windows NT/2000 security impersonation will not allow a process to access the cache of the user the process is impersonating. This is by design.
This implementation has a much smaller memory footprint and is far more robust than the fleavius implementation. It can support a very large number of caches and credentials. The only downside to the LRPC implementation is that it is currently about 35% slower than the fleavius implementation.
For more information, read the krbcc32
Architecture and krbcc32
Implementation documentation at
athena/auth/krbcc/doc/architecture.txt and
athena/auth/krbcc/doc/implementation.txt.
FILE:), please read the krbcc32
Architecture documentation.
Building KfW is supported on Windows NT 4.0, Windows 2000, and Windows XP. While buildin on Windows 9x/Me might work, it is not supported.
First, make sure that you have Microsoft Visual C++ 6.0, a recent release of the Microsoft Platform SDK (July 2000 or higher is known to work), ActiveState Perl (build 517 or higher), sed, gawk, cat, and rm in your PATH. You can get sed, gawk, cat, and rm from the Cygwin distribution. Also make sure that your INCLUDE path includes the Microsoft Platform SDK before the Microsoft Visual C++ include files and that perl has been installed so that .pl files are automatically executed with perl. You will probably also need to be using the default system shell (cmd or command, depending on whether you're running NT/2000 or 9x/Me) so that the Makefiles work properly.
Then, go into the athena directory and type
..\scripts\build.pl --softdirs..\scripts\build.pl --softdirs NODEBUG=1..\scripts\build.pl -?If you are running 9x/Me, you will need to invoke the build.pl script using perl. For example:..\scripts\build.pl --docs
[NOTE: does not work in 9x/Me]
perl ..\scripts\build.pl [args]athena/auth/krb5/src/Makefile, fix your
PATH, and try again.
To make your life easier, you might try putting the
scripts directory in your path. (Under 9x/Me, that will
not help you since you always need to invoke the script through perl.)
If you use the build.pl script, the targets should
get copied into the target directory at the same level as
the athena directory. You can go into
target\bin\i386\dbg (replacing i386 with the
right CPU and dbg with rel if building
release targets), and run the binaries. The debug symbols for the
debug build also get placed there in case you need to debug.
We recommend to developers that the binaries have a single Kerberos directory that is in the PATH.
Kerberos 5 needs a single configuration file: krb5.ini.
You can put it in the same directory as the DLL and everything will
work fine. You can also put it in the Windows directory. You can
even point to an arbitrary file by setting the
KRB5_CONFIG environment variable.
Kerberos 4 needs two configuration files, typically called
krb.con and krbrealm.con. You can put these
files in the same directory as the DLL and everything should work.
You can also set KRB4_KRB.REALMS or KRB4_KRB.CONF to override each
file. Or you can set KRB4_CONFIG to force Kerberos 4 to look for both
files in a particular directory. If you do none of these, this is
where Kerberos 4 will search:
(*) Note: If you put the files in the DLL's directory, this part of the search is what will take you there. If you have another config file earlier in the search, that will take precedence, so be careful.
Note: You can use Leash32 to edit the Kerberos 4 configuration files.
It is anticipated that most sites using Kerberos version 4 on Windows
also will have an existing UNIX Kerberos infrastructure. For that
reason, we kept the format of the krb.con identical to
the UNIX krb.conf and the format of
krbrealm.con identical to the UNIX
krb.realms. For many users, the easiest way to configure
these files for use at their local sites will be to ftp the
corresponding files from a local UNIX machine that is already properly
configured.
The krb.con file contains configuration information
describing the Kerberos realm and the Kerberos key distribution center
(KDC) servers for known realms.
krb.con contains the name of the local realm in the first
line, followed by lines indicating realm/host entries. The first token
is a realm name, and the second is a hostname of a host running a KDC
for that realm. The words "admin server" following the hostname
indicate that the host also provides an administrative database server
which is contacted when changing a user's password. For example:
ATHENA.MIT.EDU ATHENA.MIT.EDU kerberos.mit.edu admin server ATHENA.MIT.EDU kerberos-1.mit.edu ATHENA.MIT.EDU kerberos-2.mit.edu LCS.MIT.EDU kerberos.lcs.mit.edu admin server
If this were your krb.con file and you wanted to change
the default local realm to CIT.CORNELL.EDU you would edit
it to look like:
CIT.CORNELL.EDU CIT.CORNELL.EDU kerberos.cit.cornell.edu admin server ATHENA.MIT.EDU kerberos.mit.edu admin server ATHENA.MIT.EDU kerberos-1.mit.edu ATHENA.MIT.EDU kerberos-2.mit.edu LCS.MIT.EDU kerberos.lcs.mit.edu admin server
The krbrealm.con file is the host-to-Kerberos realm
translation file. This provides a translation from a local hostname to
the Kerberos realm name for the services provided by that host.
Each line of the translation file is in one the following forms
(domain_name should be of the form .XXX.YYY, e.g.,
.LCS.MIT.EDU):
host_name kerberos_realm domain_name kerberos_realm
If a hostname exactly matches the host_name field in a line of the first form, the corresponding realm is the realm of the host. If a hostname does not match any host_name in the file, but its domain exactly matches the domain_name field in a line of the second form, the corresponding realm is the realm of the host.
If no translation entry applies, the host's realm is considered to be the hostname's domain portion converted to uppercase.
Note: Leash32 can be used to edit the Kerberos 5 configuration file.
See the Kerberos 5 documentation in athena/auth/krb5/doc for further documentation on Kerberos 5.
The Kerberos DLLs need to know what port to use to talk to the Kerberos server. Kerberos 4 now defaults to ports 750 (kerberos 750/udp kdc) and 751 (kerberos-master 751/tcp) if there are no kerberos or kerberos-master entries in the services file. Kerberos 5 also has proper defaults (port 88 with a fallback to 750) in case the services file is missing the entries for kerberos and kerberos-sec.
If your site uses non-standard ports, you will still need a services file appropriate for your site.
The new default for Kerberos 4 and 5 is to store their tickets in memory.
You can specify the name of the ticket file and the directory in which
it is stored via the environment variables KRBTKFILE
(krb4) and KRB5CCNAME (krb5). The krb4 credentials
always go into memory. The ticket string you see always has an
"API:" in front of it.
There are also registry settings for these locations. Playing with Leash32 will reveal where they are (look in HKCU\Software\MIT\Kerberos4 and Kerberos5). You can set machine-wide values by playing with these settings in HKLM.
Kerberos 5 does support using file-based tickets, but we do not recommend this, as it is less secure.
Why Kerberos cares what time it is...
Kerberos authentication uses time stamps as part of its protocol. When the clocks of the Kerberos server and your computer are too far out of synchronization, you cannot authenticate properly. Both the Kerberos server and the Kerberos client depend on having clocks that are synchronized within a certain margin. This margin is normally 5 minutes.
The date and time on the machine running Kerberos must be "accurately" set. If the date or time is off "too far", Kerberos authentication will not work.
You can synchronize your clock using Leash32. It allows you to set the name of the host to which you will synchronize. It saves this information in the registry (under HKCU\Software\MIT\Leash32 -- you can set machine-wide defaults in HKLM).
By default, the server that the libraries contact when synchronizing
the time is time. The domain name has been left off on
purpose. If local system administrators create a machine with a CNAME
of time within the local domain the clients will contact this machine
by default.
If local system administrators are opposed to doing this for some
reason, you can edit the resource LSH_TIME_HOST in the
leashw32.dll to the name appropriate for your local site.
You can also edit the header files from the source distribution and
recompile for your local site. However, this is not recommended. You
can also tweak the registry setting Leash32 uses.
You can also avoid this problem by running a local, properly configured, NTP program on your machine.
The command line options for Leash32 are:
-kinit, -i only perform a kinit and then exit Leash
The options for kinit, klist, and kdestroy can be viewed by typing the
name of the utility followed by -? (e.g., kinit
-?).
Below is Jeffrey Altmans's explanation of the problem. We put in a patch he submitted for the first 2 items in his 3-item solution below.
The design of the Kclient interface is so simplified that applications cannot easily get access to necessary information. This has resulted in the need for the Kerberos configuration page in Eudora which includes fields for:
Realm: CC.COLUMBIA.EDU Service name: pop Service format: %1.%4@%3The specification of a Realm here without a "Principal name" is interesting.
Some information on my system. The local realm is KRB5.COLUMBIA.EDU and the POP server is pop.cc.columbia.edu which is in realm CC.COLUMBIA.EDU. With a ticket manager I retrieve a TGT for CC.COLUMBIA.EDU. Kstatus.exe reports that I am authenticated. Kstatus uses the kclnt32.dll interface just as Eudora does. However, when I ask Eudora to check my mail an interesting thing happens:
GetTicketForService() is called which results in a dialog being displayed
Please enter your Network ID and password. Network ID: jaltman Password:So I enter my password for CC.COLUMBIA.EDU and am told sorry but the password is incorrect.
What happened here?
First, Eudora does not include a field for the principal name to go along with the realm. So it does not call SetUserName() which should have been set to
jaltman@CC.COLUMBIA.EDUto match the realm specified in Eudora for the POP host. Eudora needs the realm so it can construct the service ticket name
pop.mailhub@CC.COLUMBIA.EDUbut without setting the username it has no mechanism to report the desired realm to kclnt32.dll.
GetTicketForService() calls an internal function to verify the TGT. But because no realm has been set and data is not shared between process boundaries this instance of kclnt32.dll has no idea that the ticket manager realm is CC.COLUMBIA.EDU. And instead of attempting to load the Ticket File Realm it calls krb_get_lrealm(). So it tries to verify a TGT
krbtgt.KRB5.COLUMBIA.EDU@KRB5.COLUMBIA.EDUwhich does not exist. So it destroys the existing tickets and then attempts to get a new TGT for
jaltmanin the default realm which again is KRB5.COLUMBIA.EDU. But of course kclnt32.dll does not display the realm in the dialog box so the user has no idea that kclnt32.dll is confused.
In order to correct this situation the following needs to be done to kclnt32.dll:
- all calls to krb_get_lrealm() should be preceded by an attempt to retrieve the realm from the ticket file with krb_get_tf_realm(). Only if krb_get_tf_realm() fails should krb_get_lream() be called.
- if a realm is specified in the GetTicketForService() request that realm should be used for the verification of the TGT.
- the dialog box displayed by UserInfo() should append the realm to the szNetID (if it is not already part of the string) when setting the default value for the box. This will indicate to the user which realm s/he is being authenticated against.
Then the author's of Eudora should add a principal name to the configuration for Kerberos and call SetUserName() to set the principal and realm to the values needed to authenticate against the POP server.
Leash DLL timesync code:
1. Use TIMEHOST environment value if defined.
2. Otherwise, use value from registry
(HKCU\Software\MIT\Leash\Settings,timehost) if present.
3. Otherwise, use value from registry
(HKLM\Software\MIT\Leash\Settings,timehost) if present.
4. Otherwise, use resource string if present.
5. Otherwise, default to #defined value "time".
Kerberos 4:
A. location of krbrealm & krbconf:
1. First, check for environment overrides:
a. Use %KRB4_KRB.REALMS% as full filename for realms file if defined.
a. Use %KRB4_KRB.CONF% as full filename for config file if defined.
b. Otherwise, look for krbrealm.con and krb.con in dir %KRB4_CONFIG%.
2. If nothing defined so far, look in registry:
a. HKCU\Software\MIT\Kerberos4,krb.realms for realms full pathname.
a. HKCU\Software\MIT\Kerberos4,krb.conf for config full pathname.
b. HKCU\Software\MIT\Kerberos4,config as dir for both files.
c. HKLM\Software\MIT\Kerberos4,krb.realms for realms full pathname .
c. HKLM\Software\MIT\Kerberos4,krb.conf for config full pathname.
d. HKLM\Software\MIT\Kerberos4,configdir as dir for both files.
3. If any of the above are set, we use it even if the files are not there.
If none of them are set, we use the old krb4 search.
B. ticket file
1. %KRBTKFILE% if defined
2. Registry setting, if setting is present
(HKCU\MIT\Kerberos4,ticketfile)
3. Registry setting, if setting is present
(HKLM\MIT\Kerberos4,ticketfile)
4. Otherwise, "API:krb4cc".
(
If we ever support a file-based cache, we should use this:
4. %TEMP%\ticket.krb, if var defined and dir exists
5. %TMP%\ticket.krb, if var defined and dir exists
6. c:\temp\ticket.krb if c:\temp exists
7. c:\tmp\ticket.krb if c:\tmp exists
8. GetWindowsDirectory()\ticket.krb as a last-ditch default? It's
either that or c:\ticket.krb!
)
Kerberos 5:
A. location of krb5.ini:
1. %KRB5_CONFIG% if defined
2. (HKCU\Software\MIT\kerberos5,config) if defined
3. (HKCU\Software\MIT\kerberos5,config) if defined
4. Otherwise, use GetWindowsDirectory()\krb5.ini
(do this instead of OpenFile to make things more explicit/simple)
B. credentials cache
1. %KRB5CCNAME% if defined
2. (HKCU\Software\MIT\kerberos5,ccname) if defined
3. (HKLM\Software\MIT\kerberos5,ccname) if defined
4. If RegKRB5CCNAME is set under [Files] in kerberos.ini,
look at that path in the registry (code already in krb5 for compat
with Gradient DCE installations, I believe).
5. Otherwise, if using CCAPI, default to "API:krb5cc".
if no CCAPI, use "FILE:" with:
a. %TEMP%\krb5cc, if var defined and dir exists
b. %TMP%\krb5cc, if var defined and dir exists
c. c:\temp\krb5cc if c:\temp exists
d. c:\tmp\krb5cc if c:\tmp exists
e. GetWindowsDirectory()\krb5cc as a last-ditch default?
it's either that or c:\krb5cc!
DNS Lookups provide Kerberos the ability to determine the Kerberos Realm that a host belongs to and to find the servers associated with a given Realm by using the Domain Name Service instead of or in addition to local configuration files.
DNS Lookups are used under two circumstances:
krb.con file is found for Kerberos 4 or no
krb5.ini file is found for Kerberos 5.
krb.con file or krb5.ini file
contains a command to activate DNS Lookups and the lookup cannot be
answered by data found in the appropriate configuration file.
To activate DNS lookups in a krb.con file place a line
that reads:
.KERBEROS.OPTION. dns
as one of the Realm to Host entries. When DNS Lookups are used the first line may be left blank to indicate that the default realm should be determined by a DNS Lookup.
To activate DNS lookups in a krb5.ini file place:
dns_fallback = true
into the [libdefaults] section. If a "default_realm"
entry is not provided, a DNS lookup will be performed to determine the
default realm.
Host to realm lookups are performed using DNS TXT records. Example records are:
_kerberos.yclept.kermit.columbia.edu. IN TXT "KRB5.COLUMBIA.EDU" _kerberos.columbia.edu. IN TXT "CC.COLUMBIA.EDU"
Realm to server lookups are performed using DNS SRV records. Example records are:
_kerberos._udp.KRB5.COLUMBIA.EDU. IN SRV 0 0 750 yclept.kermit.columbia.edu _kerberos-adm._tcp.KRB5.COLUMBIA.EDU IN SRV 0 0 749 yclept.kermit.columbia.edu _kpasswd._udp.KRB5.COLUMBIA.EDU IN SRV 0 0 464 yclept.kermit.columbia.edu
We encountered a problem at MIT that we felt needed to be addressed even though it broke some backwards compatibility. We found that if someone used a Kerberized application spanning multiple PPP sessions a Kerberos error would be generated and few applications would catch this error and try to get new tickets instead. E.g. Suppose a user starts a PPP connection and then starts Eudora, fetching mail. The user then decides to close down the PPP connection while they read their mail and compose responses. Next they initiate a new PPP connection and incorporate mail again. Note that the user never exited Eudora. Instead of prompting the user for their name and password Eudora will generate and error message. The only way for the user to recover the functionality would be to use Leash, Kview, or kdestroy to destroy their old tickets so that Eudora would get new tickets.
This happened because many ISPs hand out a new IP address to a user each time that user reconnects to the system. Also a Kerberos ticket includes the machines local IP address in an encrypted form this is used by most severs to insure that the ticket has not been copied to another users machine.
Since the local IP address is stored in the ticket it seems that it should be easy to compare this data to the machine's local IP address at the same time that an application is checking to see if the ticket has expired. Unfortunately the IP address in the ticket is encrypted in the server's session key and so is inaccessible to the local machine.
Instead we borrowed an idea from Kerberos version 5 and decided to store the local IP address, unencrypted, in the credential which is cached in the local cache. Within the Kclient function IsCredExpired() or the krbv4wXX.dll function kchktkt we verify that the ticket has not expired and that the local IP address matches the IP address stored in the ticket.
This implies that machines with multiple copies of kclnt32.dll or krbv4w32.dll, of different versions, may encounter unexpected errors when using Kerberized applications. The normal error message generated will be BAD_TKT_FILE_FORMAT or NO_TKT_FILE.
Users of applications that use other vendors Kerberos implementations may also be affected. E.g. some software from FTP, Inc.
Qualcomm has been working with Platinum on a 32-bit Kclient which would supports both Kerberos v4 and v5. From what I have heard this is a commercial implementation. It ignores GSS or other abstraction layers above the Kerberos layer that application developers should write to. It keeps its ticket cache in the DLL, as such it will not share the ticket cache with other Kerberos implementations that may reside on the user's system.
Platinum and Qualcomm decide to add a new API call to the Kclient interface. Eudora uses this new function if it finds a KCLNT32.DLL. In this case it does not use the thunking application KERB16.
We have duplicated this function in our release of KCLNT32 so that Eudora will not GPF. Please DO NOT WRITE APPLICATIONS TO THIS FUNCTION.