Age | Commit message (Collapse) | Author | Files | Lines |
|
81fe27bcc0148d410ca4617f8759b9df1a5e935c)
|
|
Autobuild-User: Andrew Bartlett <abartlet@samba.org>
Autobuild-Date: Wed Dec 1 00:59:59 CET 2010 on sn-devel-104
|
|
|
|
This allows heimdal to build without gcc, by not using atomic
operations. We don't need heimdal to be atomic in Samba.
|
|
this e_data field in a kerberos error packet tells windows to do clock
skew recovery.
See [MS-KILE] 2.2.1 KERB-ERROR-DATA
Pair-Programmed-With: Andrew Bartlett <abartlet@samba.org>
|
|
We need to use the name that the HDB entry returned, otherwise we
will not canonicalise the reply as requested.
Andrew Bartlett
|
|
By checking the client principal here, we compare the realm based on
the normalised realm, but do so early enough to validate the PAC (and
regenerate it if required).
Andrew Bartlett
|
|
Autobuild-User: Matthias Dieter Wallnöfer <mdw@samba.org>
Autobuild-Date: Mon Nov 15 23:14:57 UTC 2010 on sn-devel-104
|
|
Samba4 may modify the case of the realm in a returned entry, but will no longer modify the case of the prinicipal components.
The easy way to keep this test passing is to consider also what we
need to do to get the krbtgt account for the PAC signing - and to use
krbtgt/<this>/@REALM component to fetch the real krbtgt, and to use
that resutl for realm comparion.
Andrew Bartlett
Autobuild-User: Andrew Bartlett <abartlet@samba.org>
Autobuild-Date: Mon Nov 15 08:47:44 UTC 2010 on sn-devel-104
|
|
|
|
|
|
Heimdal uses HEIMDAL_NORETURN_ATTRIBUTE and HEIMDAL_PRINTF_ATTRIBUTE,
and we need to provide a link between these and Samba's function
attribute handling.
Andrew Bartlett
|
|
5734d03c20e104c8f45533d07f2a2cbbd3224f29)
|
|
This means that no reply packet should be generated, but that instead
the user of the libkdc API should forward the packet to a real KDC,
that has a full database.
Andrew Bartlett
|
|
Autobuild-User: Andrew Bartlett <abartlet@samba.org>
Autobuild-Date: Thu Nov 11 10:37:03 UTC 2010 on sn-devel-104
|
|
Pair-Programmed-With: Andrew Bartlett <abartlet@samba.org>
|
|
The clock skew handling was previously only on properly wrapped
GSSAPI, and was skipped for DCE-style. This allows the ASN.1 errors
from the krb5_rd_req to suggest parsing as a kerberos error packet.
Andrew Bartlett
Autobuild-User: Andrew Tridgell <tridge@samba.org>
Autobuild-Date: Mon Nov 8 07:58:09 UTC 2010 on sn-devel-104
|
|
There are exceptions from the expected behaviour of 'checksum type
matches key type' that we must deal with here, or else we can't serve
DES-only servers.
Andrew Bartlett
|
|
The new waf-based build system now has all the same functionality, and
the old build system has been broken for quite some time.
Autobuild-User: Jelmer Vernooij <jelmer@samba.org>
Autobuild-Date: Sun Oct 31 02:01:44 UTC 2010 on sn-devel-104
|
|
|
|
the lex code in heimdal had a function error_message() which conflicts
with a function from the com_err library. This replaces it with
lex_err_message()
Pair-Programmed-With: Andrew Bartlett <abartlet@samba.org>
|
|
|
|
1bea031b9404b14114b0272ecbe56e60c567af5c)
|
|
42cabfb5b683dbcb97d583c397b897507689e382)
I based this on Matthieu's import of lorikeet-heimdal, and then
updated it to this commit.
Andrew Bartlett
|
|
Some hdb modules (samba4) may change the case of the realm in
a returned result. Use that to determine if it matches the krbtgt
realm also returned from the DB (the DB will return it in the 'right' case)
Andrew Bartlett
|
|
|
|
This was a wonderful bug!
On some Fedora systems, but not on Ubuntu, there is a difference
between UTC and GMT. Heimdal replaced timegm() with _der_timegm()
which did not account for that difference (which is 24 seconds at the
moment). This led to a mutual authentication failure.
Pair-Programmed-With: Andrew Bartlett <abartlet@samba.org>
|
|
If we re-use this context, we overwrite the timestamp while talking
to the KDC and fail the mutual authentiation with the target server.
Andrew Bartlett
|
|
If the host running this code used IPv6 forms for IPv4 addreses
then the check for '.' would not be sufficient to determine that this
isn't a name we should mangle. Instead, check if it can be parsed
as a numeric address first, and only then mangle.
Andrew Bartlett
|
|
In this case, the whole request packet should be forwarded to
a real KDC, with full secrets, as we don't have the password.
This could also be used to implement 'play dead when the LDAP
server is down'.
Andrew Bartlett
|
|
This should allow master key rollover.
(but the real reason is to allow multiple krbtgt accounts, as used by
Active Directory to implement RODC support)
Andrew Bartlett
|
|
When you have a domain search list in resolv.conf, and one of the DNS
servers for a searched domain is uncontactable then we would timeout
resolving DNS names.
Avoid this by adding a '.' to the hostname if the hostname already has
a '.' in it, which we assume to mean it is fully qualified.
|
|
Karolin
|
|
metze
|
|
Karolin
|
|
If we don't do this, the PAC is given for the machine accout, not the
account being impersonated.
Andrew Bartlett
|
|
This allows us to resolve multiple forms of a name, allowing for
example machine$@REALM to get an S4U2Self ticket for
host/machine@REALM.
Andrew Bartlett
|
|
Karolin
|
|
This bug was introduced when kpasswd was migrated to a local getarg()
call, in Heimdal commit 7dd146072cd9b56d660a01f4aa20f8d81be356e8
Andrew Bartlett
|
|
|
|
f4e0dc17709829235f057e0e100d34802d3929ff)
|
|
|
|
a5e675fed7c5db8a7370b77ed0bfa724196aa84d)
|
|
security issues
|
|
|
|
|
|
"strdup" does always create a new object in the memory (through "malloc") which
needs to be freed if it isn't used anymore.
|
|
Karolin
|
|
Andrew using cp like in commit ca12e7bc8ff4a91f2044c0a60550fec902e97a78
is wrong as that removes #include "config.h" and breaks the build on AIX.
metze
|
|
This is a fairly ugly workaround, but then again, strerror_r() is a
very ugly mess.
|