]> git.ipfire.org Git - thirdparty/samba.git/log
thirdparty/samba.git
17 years agoMake events robust against their event_context being freed
Volker Lendecke [Sat, 9 Aug 2008 19:39:18 +0000 (21:39 +0200)] 
Make events robust against their event_context being freed

17 years agofix smb_len calculation for chained requests
Volker Lendecke [Sun, 10 Aug 2008 15:53:35 +0000 (17:53 +0200)] 
fix smb_len calculation for chained requests

I think chain_reply() is one of the most tricky parts of Samba. This recursion
needs to go away, we need to sequentially walk the chain list.

17 years agoFix andx offset calculation for more than 2 chained requests
Volker Lendecke [Sun, 10 Aug 2008 15:37:08 +0000 (17:37 +0200)] 
Fix andx offset calculation for more than 2 chained requests

Untested code is broken code.... Test follows later, it's quite an intrusive
change to libsmb/

17 years agoRemove an unused variable, process.c has its static copy
Volker Lendecke [Sun, 10 Aug 2008 09:33:15 +0000 (11:33 +0200)] 
Remove an unused variable, process.c has its static copy

17 years agonmbd: add support for delayed initial samlogon packages.
Michael Adam [Fri, 8 Aug 2008 23:04:55 +0000 (01:04 +0200)] 
nmbd: add support for delayed initial samlogon packages.

The hosts or networks configured with "init logon delayed hosts"
have their initial samlogon packages (empty username) delayed
by the value configured with "init logon delay" (defaulting
to 100 milliseconds).

This gives the administrator some control over what clients would
consider the preferred logon server: they choose the server that
repsonds most quickly.

Michael

17 years agonmbd_packets: make queue_packet() public.
Michael Adam [Fri, 8 Aug 2008 23:03:06 +0000 (01:03 +0200)] 
nmbd_packets: make queue_packet() public.

Michael

17 years agoloadparm: add two parameters "init logon delay hosts" and "init logon delay"
Michael Adam [Fri, 8 Aug 2008 22:31:48 +0000 (00:31 +0200)] 
loadparm: add two parameters "init logon delay hosts" and "init logon delay"

"init logon delays hosts" takes a list of hosts names or addresses
or networks for which the initial SAMLOGON reply should be delayed
(so other DCs get preferred by XP workstations if there are any).
This option takes the same type of list as "hosts allow" does.

"init logon delay" allows one to configure the delay for the hosts
configured for delayed initial samlogon with "init logon delayed hosts".
The value is interpreted as milliseconds. The default value is 100.

This commit only introduces the parameters.
They will be activated in a subsequent commit.

Michael

17 years agolib/access: make list_match() public.
Michael Adam [Fri, 8 Aug 2008 22:05:38 +0000 (00:05 +0200)] 
lib/access: make list_match() public.

Michael

17 years agolib/access: make client_match() public.
Michael Adam [Fri, 8 Aug 2008 22:03:23 +0000 (00:03 +0200)] 
lib/access: make client_match() public.

Michael

17 years agoOne more build fix. Ensure we have KRB5_AUTH_CONTEXT_USE_SUBKEY defined before we...
Jeremy Allison [Fri, 8 Aug 2008 23:08:11 +0000 (16:08 -0700)] 
One more build fix. Ensure we have KRB5_AUTH_CONTEXT_USE_SUBKEY defined before we compile the new code.
Jeremy.

17 years agoTry and fix the build for systems that don't have krb5_auth_con_set_req_cksumtype().
Jeremy Allison [Fri, 8 Aug 2008 22:15:36 +0000 (15:15 -0700)] 
Try and fix the build for systems that don't have krb5_auth_con_set_req_cksumtype().
Jeremy.

17 years agoMerge branch 'v3-3-test' of ssh://jra@git.samba.org/data/git/samba into v3-3-test
Jeremy Allison [Fri, 8 Aug 2008 21:33:55 +0000 (14:33 -0700)] 
Merge branch 'v3-3-test' of ssh://jra@git.samba.org/data/git/samba into v3-3-test

17 years agoAdd Derrick Schommer's <dschommer@F5.com> kerberos delegation patch. Some
Jeremy Allison [Fri, 8 Aug 2008 21:32:15 +0000 (14:32 -0700)] 
Add Derrick Schommer's <dschommer@F5.com> kerberos delegation patch. Some
work by me and advice by Love.
Jeremy.

17 years agobuild: fix a no previous prototype warning when building without ldap/gssapia
Michael Adam [Fri, 8 Aug 2008 21:30:19 +0000 (23:30 +0200)] 
build: fix a no previous prototype warning when building without ldap/gssapia

move prototype of dns_create_update_request() to appropriate section in dns.h

Michael

17 years agolibnet samsync ldif: fix the build without LDAP.
Michael Adam [Fri, 8 Aug 2008 21:03:51 +0000 (23:03 +0200)] 
libnet samsync ldif: fix the build without LDAP.

Michael

17 years agousing NGROUPS_MAX instead of 32 for the max group value in rep_initgroups() subroutin...
Yannick Bergeron [Fri, 8 Aug 2008 17:32:15 +0000 (13:32 -0400)] 
using NGROUPS_MAX instead of 32 for the max group value in rep_initgroups() subroutine in lib/replace/replace.c

17 years agoAdd simple async wrappers around send, recv and connect
Volker Lendecke [Fri, 8 Aug 2008 16:30:57 +0000 (18:30 +0200)] 
Add simple async wrappers around send, recv and connect

To be used later :-)

17 years agoFix bug #5675 with a varient of Tim Waugh's patch,
Jeremy Allison [Fri, 8 Aug 2008 00:55:57 +0000 (17:55 -0700)] 
Fix bug #5675 with a varient of Tim Waugh's patch,
as proposed by James Peach.
Jeremy.

17 years agoFix "might be used uninitialized" warnings.
Jeremy Allison [Fri, 8 Aug 2008 00:49:19 +0000 (17:49 -0700)] 
Fix "might be used uninitialized" warnings.
Jeremy.

17 years agoFix a build failure on host sunX
Volker Lendecke [Thu, 7 Aug 2008 16:18:10 +0000 (18:18 +0200)] 
Fix a build failure on host sunX

17 years agoSolve an IBM XL C/C++ compiler error encountered in get_exit_code() auth_errors array...
Yannick Bergeron [Wed, 6 Aug 2008 17:23:00 +0000 (13:23 -0400)] 
Solve an IBM XL C/C++ compiler error encountered in get_exit_code() auth_errors array initialization in client/smbspool.c

17 years agoWHATSNEW: Start WHATSNEW for 3.3.0pre1.
Karolin Seeger [Wed, 6 Aug 2008 12:26:41 +0000 (14:26 +0200)] 
WHATSNEW: Start WHATSNEW for 3.3.0pre1.

Karolin

17 years agolibnetapi: fix build of shared library after libnet_join changes.
Michael Adam [Wed, 6 Aug 2008 11:56:52 +0000 (13:56 +0200)] 
libnetapi: fix build of shared library after libnet_join changes.

This needs create_builtin_administrators() and create_builtin_users()
from token_utils now. Did not pop up because the only users of the
shared lib currently are the examples in lib/netapi/examples/
which are not automatically built.

Michael

17 years agofixed permissions on ctdb databases
Andrew Tridgell [Wed, 6 Aug 2008 06:35:43 +0000 (16:35 +1000)] 
fixed permissions on ctdb databases

17 years agofixed a fd leak when trying to regain contact to a domain controller
Andrew Tridgell [Wed, 6 Aug 2008 04:02:45 +0000 (14:02 +1000)] 
fixed a fd leak when trying to regain contact to a domain controller
in winbind

When a w2k3 DC is rebooted the 139/445 ports come up before the
udp/389 cldap port. During this brief period, winbind manages to
connect to 139/445 but not to udp 389. It then enters a tight loop
where it leaks one fd each time. In a couple of seconds it runs out of
file descriptors, and leaves winbind crippled after the DC does
finally come up

17 years agodbwrap: add comment describing behaviour of dbwrap_change_int32_atomic().
Michael Adam [Tue, 5 Aug 2008 21:38:56 +0000 (23:38 +0200)] 
dbwrap: add comment describing behaviour of dbwrap_change_int32_atomic().

Michael

17 years agosecrets: fix replacemend random seed generator (security issue).
Michael Adam [Tue, 5 Aug 2008 21:14:05 +0000 (23:14 +0200)] 
secrets: fix replacemend random seed generator (security issue).

This is a regression introduced by the change to dbwrap.
The replacement dbwrap_change_int32_atomic() does not
correctly mimic the behaviour of tdb_change_int32_atomic():
The intended behaviour is to use *oldval  as an initial
value when the entry does not yet exist in the db and to
return the old value in *oldval.

The effect was that:
1. get_rand_seed() always returns sys_getpid() in *new_seed
   instead of the incremented seed from the secrets.tdb.
2. the seed stored in the tdb is always starting at 0 instead
   of sys_getpid() + 1 and incremented in subsequent calls.

In principle this is a security issue, but i think the danger is
low, since this is only used as a fallback when there is no useable
/dev/urandom, and this is at most called on startup or via
reinit_after_fork.

Michael

17 years agodbwrap: add comment describing behaviour of dbwrap_change_uint32_atomic().
Michael Adam [Tue, 5 Aug 2008 21:13:06 +0000 (23:13 +0200)] 
dbwrap: add comment describing behaviour of dbwrap_change_uint32_atomic().

Michael

17 years agoidmap_tdb2: fix a race condition in idmap_tdb2_allocate_id().
Michael Adam [Tue, 5 Aug 2008 20:38:44 +0000 (22:38 +0200)] 
idmap_tdb2: fix a race condition in idmap_tdb2_allocate_id().

The race is a regression introduced by the change to dbwrap.
It might have led to two concurrent processes returning the same id.

This fix is achieved by changing dbwrap_change_uint32_atomic() to
match the original behaviour of tdb_change_uint32_atomic(), which
is the following: *oldval is used as initial value when
the value does not yet exist and that the old value should be
returned in *oldval.

dbwrap_change_uint32_atomic() is used (only) in idmap_tdb2.c,
to get new ids.

Michael

17 years agoregistry: use _bystring wrappers to dbwrap_trans_(store|delete).
Michael Adam [Mon, 4 Aug 2008 21:30:16 +0000 (23:30 +0200)] 
registry: use _bystring wrappers to dbwrap_trans_(store|delete).

Michael

17 years agoBuilding cifs.upcall is giving this build warning:
Steve French [Tue, 5 Aug 2008 20:36:11 +0000 (15:36 -0500)] 
Building cifs.upcall is giving this build warning:

   client/cifs.upcall.c:205: warning: function declaration isn’t a prototype

This patch fixes this by properly declaring usage() args as void.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <sfrench@samba.org>
17 years agocifs.upcall: fix manpage and comments
Steve French [Tue, 5 Aug 2008 18:27:07 +0000 (13:27 -0500)] 
cifs.upcall: fix manpage and comments

The "cifs.resolver" key type has been changed to "dns_resolver". Fix
the comments at the top of cifs.upcall and the manpage accordingly.

Signed-off-by: Jeff Layton <jlayton@redhat.com>
Signed-off-by: Steve French <sfrench@samba.org>
---
 docs-xml/manpages-3/cifs.upcall.8.xml |    4 ++--
 source/client/cifs.upcall.c           |    8 ++++----
 2 files changed, 6 insertions(+), 6 deletions(-)

17 years agoBacking out most of changeset 5222b8db3fb692e5071bfd1b41849a8eb0a17995
Steve French [Tue, 5 Aug 2008 18:15:46 +0000 (13:15 -0500)] 
Backing out most of changeset 5222b8db3fb692e5071bfd1b41849a8eb0a17995
(so parsing for domain parameter in mount.cifs matches online help)
and rephrasing original code to make it more clear.

The check for "domain" was meant to allow for "dom" or "DOM" and the
option ("dom") described in the help (e.g. "/sbin/mount.cifs -?") is the
shorter ("dom") form.  The reason that the string we compare against
is larger was to improve readability (we could compare against "dom"
but note /* "domain" or "DOMAIN" or "dom" or "DOM" */ but it seemed
terser to just show the larger string in the strcmp target.   The
change to "workgoup" from workg* (anything which begins with "workg"
doesn't matter - it is a minor behavior change - but probably few
scripts depend on the "alias" for this option).

Rework code so that it is clearer what we are comparing against.

17 years agoman pages: Improve description of boolean values in smb.conf.5.
Karolin Seeger [Tue, 5 Aug 2008 12:20:32 +0000 (14:20 +0200)] 
man pages: Improve description of boolean values in smb.conf.5.

This fixes bug #5378.
Thanks Morton K. Poulsen <morten+bugzilla.samba.org [at] afdelingp.dk>
for reporting!

Karolin

17 years agoman pages: Add documentation about smbclient command "rename".
Karolin Seeger [Tue, 5 Aug 2008 12:10:11 +0000 (14:10 +0200)] 
man pages: Add documentation about smbclient command "rename".

This fixes bug #5268.
Thanks to Alexander Franz <a.franz [at] gmx.net> for reporting!

Karolin

17 years agoREADME.Coding: A few minor fixes.
Karolin Seeger [Tue, 5 Aug 2008 10:55:20 +0000 (12:55 +0200)] 
README.Coding: A few minor fixes.

Karolin

17 years agolibnet_keytab: fix the build with heimdal
Stefan Metzmacher [Mon, 4 Aug 2008 12:28:02 +0000 (14:28 +0200)] 
libnet_keytab: fix the build with heimdal

metze

17 years agoclikrb5: don't use krb5_keyblock_init() when no salt is specified
Stefan Metzmacher [Mon, 4 Aug 2008 11:52:18 +0000 (13:52 +0200)] 
clikrb5: don't use krb5_keyblock_init() when no salt is specified

If the caller wants to create a key with no salt we should
not use krb5_keyblock_init() (only used when using heimdal)
because it does sanity checks on the key length.

metze

17 years agocli_request_new() already gave use the req, remove a pointless function call
Volker Lendecke [Fri, 1 Aug 2008 14:05:49 +0000 (16:05 +0200)] 
cli_request_new() already gave use the req, remove a pointless function call

17 years agoFix a typo
Volker Lendecke [Fri, 1 Aug 2008 13:29:06 +0000 (15:29 +0200)] 
Fix a typo

17 years agolibnet dssync: start memory allocation cleanup: use tmp ctx in libnet_dssync().
Michael Adam [Fri, 1 Aug 2008 15:13:42 +0000 (17:13 +0200)] 
libnet dssync: start memory allocation cleanup: use tmp ctx in libnet_dssync().

Don't leak temporary data to callers but use a temporary context
that is freed at the end.

Michael

17 years agolibnet dssync: fix memory allocation for error/result messages.
Michael Adam [Fri, 1 Aug 2008 15:10:59 +0000 (17:10 +0200)] 
libnet dssync: fix memory allocation for error/result messages.

Use the libnet_dssync_context as a talloc context for the
result_message and error_message string members.
Using the passed in mem_ctx makes the implicit assumption
that mem_ctx is at least as long-lived as the libnet_dssync_context,
which is wrong.

Michael

17 years agodssync keytab: add comment header explaining add_to_keytab_entries().
Michael Adam [Fri, 1 Aug 2008 15:09:08 +0000 (17:09 +0200)] 
dssync keytab: add comment header explaining add_to_keytab_entries().

Michael

17 years agolibnet dssync: add my C after dssync keytab changes.
Michael Adam [Fri, 1 Aug 2008 12:26:46 +0000 (14:26 +0200)] 
libnet dssync: add my C after dssync keytab changes.

Michael

17 years agovampire keytab: add command line switch --clean-old-entries .
Michael Adam [Thu, 31 Jul 2008 22:12:18 +0000 (00:12 +0200)] 
vampire keytab: add command line switch --clean-old-entries .

This allows to control cleaning the keytab.
It will only clean old occurences of keys that are replicated in
this run. So if you want to ensure things are cleaned up, combine
this switch with --force-full-repl or --single-obj-repl (+dn list).

Michael

17 years agodssync: add clean_old_entries flag to dssync_ctx.
Michael Adam [Thu, 31 Jul 2008 22:09:28 +0000 (00:09 +0200)] 
dssync: add clean_old_entries flag to dssync_ctx.

Initialize it to false.
And pass it down to the libnet_keytab context in
libnet_dssync_keytab.c:keytab_startup().

Unused yet.

Michael

Note: This might not be not 100% clean design to put this into the
toplevel dssync context while it is keytab specific. But then, on the
other hand, other imaginable backends might want to use this flag, too...

17 years agolibnet keytab: implement cleaning of old entries in libnet_keytab_add().
Michael Adam [Thu, 31 Jul 2008 22:07:40 +0000 (00:07 +0200)] 
libnet keytab: implement cleaning of old entries in libnet_keytab_add().

Triggered by the flag clean_old_entries from the libnet_keytab_contex
(unused yet...).

Michael

17 years agolibnet keytab: add parameter ingnore_kvno to libnet_keytab_remove_entries()
Michael Adam [Thu, 31 Jul 2008 22:05:42 +0000 (00:05 +0200)] 
libnet keytab: add parameter ingnore_kvno to libnet_keytab_remove_entries()

to allow for removing all entries with given principal and enctype without
repecting the kvno (i.e. cleaning "old" entries...)

This is called with ignore_kvno == false from libnet_keytab_add_entry() to
keep the original behaviour.

Michael

17 years agolibnet keytab: add flag clean_old_entries to libnet_keytab_context.
Michael Adam [Thu, 31 Jul 2008 22:03:10 +0000 (00:03 +0200)] 
libnet keytab: add flag clean_old_entries to libnet_keytab_context.

Michael

17 years agolibnet keytab: use proper counter type (uint32_t) in libnet_keytab_add().
Michael Adam [Thu, 31 Jul 2008 21:15:35 +0000 (23:15 +0200)] 
libnet keytab: use proper counter type (uint32_t) in libnet_keytab_add().

Michael

17 years agovampire keytab: introduce switch --single-obj-repl.
Michael Adam [Thu, 31 Jul 2008 21:05:45 +0000 (23:05 +0200)] 
vampire keytab: introduce switch --single-obj-repl.

This controls whether single object replication is to be used.
This only has an effect when at least one object dn is given
on the commandline.

NOTE: Now the default is to use normal replication with uptodateness
vectors and use object dns given on the command line as a positive
write filter. Single object replication is only performed when this
new switch is specified.

Michael

17 years agodssync keytab: when not in single object replication mode, use object dn list as...
Michael Adam [Thu, 31 Jul 2008 20:53:41 +0000 (22:53 +0200)] 
dssync keytab: when not in single object replication mode, use object dn list as write filter.

I.e. only the passwords and keys of those objects whose dns are provided
are written to the keytab file. Others are skippded.

Michael

17 years agodssync keytab: support storing kerberos keys from supplemental credentials.
Michael Adam [Thu, 31 Jul 2008 10:25:06 +0000 (12:25 +0200)] 
dssync keytab: support storing kerberos keys from supplemental credentials.

Michael

17 years agolibnet dssync: rename flag single to single_object_replication
Michael Adam [Wed, 30 Jul 2008 15:53:28 +0000 (17:53 +0200)] 
libnet dssync: rename flag single to single_object_replication

So that it is more obvious what this controls.

Michael

17 years agonet rpc vampire: rename --repl-nodiff to --force-full-repl.
Michael Adam [Wed, 30 Jul 2008 15:46:13 +0000 (17:46 +0200)] 
net rpc vampire: rename --repl-nodiff to --force-full-repl.

This more clear.

Michael

17 years agolibnet dssync: rename repl_nodiff flag to force_full_replication.
Michael Adam [Wed, 30 Jul 2008 15:44:22 +0000 (17:44 +0200)] 
libnet dssync: rename repl_nodiff flag to force_full_replication.

Michael

17 years agolibnet dssync: support lists of dns (instead of one dn) for single object replication.
Michael Adam [Wed, 30 Jul 2008 11:02:36 +0000 (13:02 +0200)] 
libnet dssync: support lists of dns (instead of one dn) for single object replication.

Just specify several DNs separated by spaces on the command line of
"net rpc vampire keytab" to get the passwords for each of these
accouns via single object replication.

Michael

17 years agolibnet dssync: move determination of request level into build_request()
Michael Adam [Wed, 30 Jul 2008 10:35:45 +0000 (12:35 +0200)] 
libnet dssync: move determination of request level into build_request()

...where it belongs.

Michael

17 years agolibnet dssync: refactor dsgetncchanges loop out into libnet_dssync_getncchanges().
Michael Adam [Wed, 30 Jul 2008 10:32:30 +0000 (12:32 +0200)] 
libnet dssync: refactor dsgetncchanges loop out into libnet_dssync_getncchanges().

Michael

17 years agolibnet dssync: fix single object replication by adding one check.
Michael Adam [Wed, 30 Jul 2008 10:31:38 +0000 (12:31 +0200)] 
libnet dssync: fix single object replication by adding one check.

Before, this used the old uptodate vector in the request...

Michael

17 years agolibnet dssync: simplify logic of libnet_dssync_process() main loop.
Michael Adam [Wed, 30 Jul 2008 10:00:49 +0000 (12:00 +0200)] 
libnet dssync: simplify logic of libnet_dssync_process() main loop.

Untangle parsing of results and processing.
Make loop logic more obvious.
Call finishing operation after the loop, not inside.

Michael

17 years agolibnet dssync: refactor creation of request out into new function
Michael Adam [Wed, 30 Jul 2008 08:27:00 +0000 (10:27 +0200)] 
libnet dssync: refactor creation of request out into new function

libnet_dssync_build_request().

Michael

17 years agovampire keytab: add switch --repl-nodiff to trigger full replication.
Michael Adam [Tue, 29 Jul 2008 20:52:59 +0000 (22:52 +0200)] 
vampire keytab: add switch --repl-nodiff to trigger full replication.

I.e. replication without keeping track of the up to date vector.

Michael

17 years agodssync keytab: store the samaccountname in the keytab for diff replication.
Michael Adam [Tue, 29 Jul 2008 16:07:07 +0000 (18:07 +0200)] 
dssync keytab: store the samaccountname in the keytab for diff replication.

When retreiving a diff replication, the sAMAccountName attribute is usually
not replicated. So in order to build the principle, we need to store the
sAMAccounName in the keytab, referenced  by the DN of the object, so that
it can be retrieved if necessary.

It is stored in the form of SAMACCOUNTNAME/object_dn@dns_domain_name
with kvno=0 and ENCTYPE_NONE.

Michael

17 years agodssync keytab: move handling of removal of duplicates to libnet_keytab_add_entry().
Michael Adam [Tue, 29 Jul 2008 16:05:13 +0000 (18:05 +0200)] 
dssync keytab: move handling of removal of duplicates to libnet_keytab_add_entry().

This makes libnet_keytab_remove_entries static and moves it up.
libnet_keytab_add_entry() now removes the duplicates in advance.
No special handling neede for the UTDV - this is also needed
for other entries...

Michael

17 years agolibnet_keytab: add some debug statements to libnet_keytab_search().
Michael Adam [Tue, 29 Jul 2008 15:54:01 +0000 (17:54 +0200)] 
libnet_keytab: add some debug statements to libnet_keytab_search().

Michael

17 years agodssync keytab: store the UpToDate vector with ENCTYPE_NULL.
Michael Adam [Tue, 29 Jul 2008 13:23:12 +0000 (15:23 +0200)] 
dssync keytab: store the UpToDate vector with ENCTYPE_NULL.

Michael

17 years agolibnet keytab: use libnet_keytab_add_entry() in libnet_keytab_add().
Michael Adam [Tue, 29 Jul 2008 13:21:30 +0000 (15:21 +0200)] 
libnet keytab: use libnet_keytab_add_entry() in libnet_keytab_add().

This will in particular allow us to store ENCTYPE_NULL.

Michael

17 years agolibnet keytab: add function libnet_keytab_add_entry()
Michael Adam [Tue, 29 Jul 2008 13:19:18 +0000 (15:19 +0200)] 
libnet keytab: add function libnet_keytab_add_entry()

This is a stripped down version of smb_krb5_kt_add_entry() that
takes one explicit enctype instead of an array. And it does
not neither salting of keys nor cleanup of old entries.

Michael

17 years agodssync keytab: log the DN of the object to be parsed.
Michael Adam [Tue, 29 Jul 2008 12:15:07 +0000 (14:15 +0200)] 
dssync keytab: log the DN of the object to be parsed.

For debugging purposes.

Michael

17 years agodssync keytab: remove old UpToDateNess vectors from keytab before storing new one.
Michael Adam [Tue, 29 Jul 2008 12:13:37 +0000 (14:13 +0200)] 
dssync keytab: remove old UpToDateNess vectors from keytab before storing new one.

Michael

17 years agolibnet keytab: add function libnet_keytab_remove_entries().
Michael Adam [Tue, 22 Jul 2008 09:39:01 +0000 (11:39 +0200)] 
libnet keytab: add function libnet_keytab_remove_entries().

This can be used to remove entries of given principal, kvno and enctype.

Michael

17 years agolibnet_keytab: cleanup libnet_keytab_search().
Michael Adam [Tue, 29 Jul 2008 12:39:40 +0000 (14:39 +0200)] 
libnet_keytab: cleanup libnet_keytab_search().

Michael

17 years agolibnet keytab: test for matching enctype in libnet_keytab_search().
Michael Adam [Tue, 29 Jul 2008 11:32:17 +0000 (13:32 +0200)] 
libnet keytab: test for matching enctype in libnet_keytab_search().

Michael

17 years agodssync keytab: add parsing and logging of servicePrincipalName-s
Michael Adam [Tue, 29 Jul 2008 10:55:19 +0000 (12:55 +0200)] 
dssync keytab: add parsing and logging of servicePrincipalName-s

As with the userPrincipalName, this is for debugging purposes only (for now..).

Michael

17 years agodssync keytab: fix comma placement in debug output
Michael Adam [Tue, 29 Jul 2008 10:54:46 +0000 (12:54 +0200)] 
dssync keytab: fix comma placement in debug output

Michael

17 years agodssync keytab: add debugging output when skipping an object.
Michael Adam [Tue, 29 Jul 2008 08:17:15 +0000 (10:17 +0200)] 
dssync keytab: add debugging output when skipping an object.

Michael

17 years agolibnet keytab: add enctype parameter to libnet_keytab_search().
Michael Adam [Tue, 29 Jul 2008 08:16:37 +0000 (10:16 +0200)] 
libnet keytab: add enctype parameter to libnet_keytab_search().

Not really used yet.

Note: callers use ENCTYPE_ARCFOUR_HMAC enctype for UTDV (for now).
This is what is currently stored. This is to be changed
to ENCTYPE_NULL.

Michael

17 years agodssync keytab: add store enctypes in the libnet_keytype_entry structs.
Michael Adam [Mon, 28 Jul 2008 12:42:30 +0000 (14:42 +0200)] 
dssync keytab: add store enctypes in the libnet_keytype_entry structs.

Still unused by the libnet_keytab_add() function.
This will follow.
In preparation of supporting multiple encryption types in libnet_dssync_keytab.

Michael

17 years agolibnet_keytab: add enctype field to libnet_keytab_entry struct.
Michael Adam [Mon, 28 Jul 2008 12:40:54 +0000 (14:40 +0200)] 
libnet_keytab: add enctype field to libnet_keytab_entry struct.

In preparation of supporting more enctyption types in libnet_dssync_keytab.

Michael

17 years agodssync: allow replications of a single obj with net rpc vampire keytab.
Michael Adam [Thu, 17 Jul 2008 22:18:40 +0000 (00:18 +0200)] 
dssync: allow replications of a single obj with net rpc vampire keytab.

This is triggered by setting the new "single" flag in the dssync_context
and filling the "object_dn" member with the dn of the object to be
fetched.

This call is accomplished by specifying the DRSUAPI_EXOP_REPL_OBJ
extended operation in the DsGetNCCHanges request. This variant does
honor an up-to-date-ness vectore passed in, but the answer does not
return a new up-to-dateness vector.

Call this operation as "net rpc vampire keytab /path/keytab object_dn" .

Michael

17 years agodssync: pass uptodateness vector into and out of DsGetNCChanges request.
Michael Adam [Wed, 16 Jul 2008 23:05:06 +0000 (01:05 +0200)] 
dssync: pass uptodateness vector into and out of DsGetNCChanges request.

Also store the new uptodateness vector in the backend after completion
and retrieve the old vector before sending the DsGetNCChanges request.

This effectively accomplishes differential replication.

Michael

17 years agodssync: skip analysis of the msDS_KeyVersionNumber attribute:
Michael Adam [Thu, 17 Jul 2008 11:32:19 +0000 (13:32 +0200)] 
dssync: skip analysis of the msDS_KeyVersionNumber attribute:

It is a calculated attribute that won't get distributed via replication.

Michael

17 years agodssync: either use the req5 or the req8 request, depending on the supported_extenstion
Michael Adam [Thu, 17 Jul 2008 11:05:43 +0000 (13:05 +0200)] 
dssync: either use the req5 or the req8 request, depending on the supported_extenstion

that have been recorded in the remote_info28 in the dssync_context.

Michael

17 years agodssync: record the bind info in the new remote_info28 in libnet_dssync_bind().
Michael Adam [Thu, 17 Jul 2008 11:04:04 +0000 (13:04 +0200)] 
dssync: record the bind info in the new remote_info28 in libnet_dssync_bind().

This extracts the info24 data in case this is what was returned (instead of info28).
E.g. windows 2000 returns info24.

Michael

17 years agodssync: add a drsuapi_DsBindInfo28 struct to the dssync_context struct
Michael Adam [Thu, 17 Jul 2008 11:02:31 +0000 (13:02 +0200)] 
dssync: add a drsuapi_DsBindInfo28 struct to the dssync_context struct

to keep track of what the server told us upon DsBind.

Michael

17 years agodssync keytab: wrap printing of the uptodate vector in DEBUGLEVEL >= 10 checks
Michael Adam [Thu, 17 Jul 2008 09:54:32 +0000 (11:54 +0200)] 
dssync keytab: wrap printing of the uptodate vector in DEBUGLEVEL >= 10 checks

Michael

17 years agodssync keytab: add support for keeping track of the up-to-date-ness vector.
Michael Adam [Wed, 16 Jul 2008 22:54:35 +0000 (00:54 +0200)] 
dssync keytab: add support for keeping track of the up-to-date-ness vector.

The startup operation should get the old up-to-date-ness vector from the backend
and the finish operation should store the new vector to the backend after replication.

This adds the change of the signatures of the operations ot the dssync_ops struct
and the implementation for the keytab ops. The up-to-date-ness vector is stored
under the principal constructed as UTDV/$naming_context_dn@$dns_domain_name.

The vector is still uninterpreted in libnet_dssync_process().
This will be the next step...

This code is essentially by Metze.

Michael

17 years agolibnet_keytab: add a libnet_keytab_search() function
Michael Adam [Wed, 16 Jul 2008 22:53:13 +0000 (00:53 +0200)] 
libnet_keytab: add a libnet_keytab_search() function

that searches and fetches an entry from a keytab file by principal and kvno.

This code is by metze.

Michael

17 years agodssync keytab: use add_to_keytab_entries() for pwd history in parse_object().
Michael Adam [Wed, 23 Jul 2008 22:30:07 +0000 (00:30 +0200)] 
dssync keytab: use add_to_keytab_entries() for pwd history in parse_object().

Michael

17 years agodssync keytab: add prefix parameter to add_to_keytab_entries() for flexibility.
Michael Adam [Wed, 16 Jul 2008 21:12:31 +0000 (23:12 +0200)] 
dssync keytab: add prefix parameter to add_to_keytab_entries() for flexibility.

This will allow to construct principals of the form PREFIX/name@domain

Michael

17 years agodssync keytab: add check for success of ADD_TO_ARRAY().
Michael Adam [Wed, 16 Jul 2008 21:10:20 +0000 (23:10 +0200)] 
dssync keytab: add check for success of ADD_TO_ARRAY().

Michael

17 years agodssync keytab: refactor adding entry to keytab_context out into new function
Michael Adam [Wed, 16 Jul 2008 21:08:40 +0000 (23:08 +0200)] 
dssync keytab: refactor adding entry to keytab_context out into new function

add_to_keytab_entries()

Michael

17 years agodssync: replace the processing_fn by startup/process/finish ops.
Michael Adam [Wed, 16 Jul 2008 15:12:04 +0000 (17:12 +0200)] 
dssync: replace the processing_fn by startup/process/finish ops.

This remove static a variable for the keytab context in the keytab
processing function and simplifies the signature. The keytab context
is instead in the new private data member of the dssync_context struct.

This is in preparation of adding support for keeping track of the
up-to-date-ness vector, in order to be able to sync diffs instead
of the whole database.

Michael

17 years agoSamba3 HowTo: Fix duplicate chapter id.
Karolin Seeger [Fri, 1 Aug 2008 12:10:28 +0000 (14:10 +0200)] 
Samba3 HowTo: Fix duplicate chapter id.

Karolin

17 years agonetapi: when using NetApi functions forward net's kerberos setting.
Günther Deschner [Thu, 31 Jul 2008 13:14:14 +0000 (15:14 +0200)] 
netapi: when using NetApi functions forward net's kerberos setting.

Guenther

17 years agonet: Use NetLocalGroupAdd() for adding aliases.
Günther Deschner [Thu, 31 Jul 2008 13:12:09 +0000 (15:12 +0200)] 
net: Use NetLocalGroupAdd() for adding aliases.

Guenther

17 years agonetapi: in NetLocalGroupAdd_r() only set description if necessary.
Günther Deschner [Thu, 31 Jul 2008 13:11:20 +0000 (15:11 +0200)] 
netapi: in NetLocalGroupAdd_r() only set description if necessary.

Guenther

17 years agolibnetunjoin: add use_kerberos flag.
Günther Deschner [Thu, 31 Jul 2008 12:23:23 +0000 (14:23 +0200)] 
libnetunjoin: add use_kerberos flag.

Guenther

17 years agonet: add "-k" switch for kerberos authentication (in preparation for #5416).
Günther Deschner [Wed, 30 Jul 2008 19:37:09 +0000 (21:37 +0200)] 
net: add "-k" switch for kerberos authentication (in preparation for #5416).

Guenther