For compatibility with MIT Kerberos, support automatic acquisition of initiator
credentials if a client keytab is available. The default path on non-Windows is
/var/heimdal/user/%{euid}/client.keytab, but can be overriden with the
KRB5_CLIENT_KTNAME environment variable or the default_client_keytab_name
configuration option. If a client keytab does not exist, or exists but does not
contain the principal for which initiator credentials are being acquired, the
system keytab is tried.
It turns out gss_add_cred() really needed a complete rewrite. It's much
better to first have a gss_duplicate_cred() (which has been needed for
other reasons anyways), and use that when the input_cred_handle is not
GSS_C_NO_CREDENTIAL and output_cred_handle is not NULL, then mutate that
duplicate credential handle (or the input_cred_handle if
output_cred_handle is NULL).
Zero out the DES_cblock structure instead of the (not yet used at this point
in the function) key schedule. The contents could potentially be left
on the stack in the case of an error return from _gssapi_verify_pad().
We generally clear out the cryptographic key and key schedule from
local variables before relinquishing control flow, but this case was
missed. Reported by jhb@FreeBSD.org.
Treat principals of type NT-UNKNOWN as NT-SRV-HST if the first component
of the principal name is "host".
Change-Id: I28fb619379daac827436040e701d4ab7b279852b
gsskrb5_acceptor_start() was making a copy of the global pointer
_gsskrb5_keytab to use later. This invites a race condition where
another thread could call gsskrb5_register_acceptor_identity()
(thus invalidating the target of the copied pointer) before it is
used by gsskrb5_acceptor_start().
So instead, clone the keytab to a new one while protected by the
mutex lock (similar to get_keytab() in acquire_cred.c).
Signed-off-by: Nicolas Williams <nico@twosigma.com>
Some non-GSSAPI implementations that instead try to create compatible packets by wrapping krb5_mk_req()
can trigger a NULL authenticator here. Assume this to be equvilent to specifying an all-zero
channel bindings and some reasonable (fixed) flags.
Original patch by Andrew Bartlett, restructured by Douglas Bagnall
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Solaris'/Illumos' gss_acquire_cred_with_password() does not have
side-effects. MIT and Heimdal have differed, but it's now agreed that
the Solaris/Illumos behavior is correct.
To make a credential obained with gss_acquire_cred_with_password()
available to other processes, use gss_store_cred().
In at least two instances the krb5 cred handle expiration time was misused
as a remaining lifetime. This is not surprising since the field name is
wrong ("lifetime" not "expiration"). This commit fixes the code, the next
commit will rename the field and change its type from OM_uint32 to time_t.
gss_add_cred() with GSS_C_NO_CREDENTIAL as the input_cred_handle should
act like gss_acquire_cred() with desired_mechs containing just the
desired_mech.
In gsskrb5_accept_delegated_token() it is wrong to store the delegated
credentials in the default ccache by default. When the caller does not
provide a target credential handle, we just do nothing and return success.
Test the return value of gsskrb5_accept_delegated_token() against
GSS_S_COMPLETE, rather than 0.
Delegated or other explicit credentials were mishandled, the code only
worked correctly when processing default credentials. In particular
this caused root's default credential cache to be accessed when accepting
delegated credentials in SSH:
ssh_gssapi_accept_ctx() ->
ssh_gssapi_getclient() ->
gss_inquire_cred_by_mech()
When /tmp/krb5cc_0 contained expired tickets, cascaded credentials
stopped working for non-root users!
The first enctype RFC3961 prf output length's bytes are correct because
the little- and big-endian representations of unsigned zero are the
same. The second block of output was wrong because the counter was not
being encoded as big-endian.
This change could break applications. But those applications would not
have been interoperating with other implementations anyways (in
particular: MIT's).
We turn on a few extra warnings and fix the fallout that occurs
when building with --enable-developer. Note that we get different
warnings on different machines and so this will be a work in
progress. So far, we have built on NetBSD/amd64 5.99.64 (which
uses gcc 4.5.3) and Ubuntu 10.04.3 LTS (which uses gcc 4.4.3).
Notably, we fixed
1. a lot of missing structure initialisers,
2. unchecked return values for functions that glibc
marks as __attribute__((warn-unused-result)),
3. made minor modifications to slc and asn1_compile
which can generate code which generates warnings,
and
4. a few stragglers here and there.
We turned off the extended warnings for many programs in appl/ as
they are nearing the end of their useful lifetime, e.g. rsh, rcp,
popper, ftp and telnet.
Interestingly, glibc's strncmp() macro needed to be worked around
whereas the function calls did not.
We have not yet tried this on 32 bit platforms, so there will be
a few more warnings when we do.
