IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
most of the changes are fixes to make all the ldb code compile without
warnings on gcc4. Unfortunately That required a lot of casts :-(
I have also added the start of an 'operational' module, which will
replace the timestamp module, plus add support for some other
operational attributes
In ldb_msg_*() I added some new utility functions to make the
operational module sane, and remove the 'ldb' argument from the
ldb_msg_add_*() functions. That argument was only needed back in the
early days of ldb when we didn't use the hierarchical talloc and thus
needed a place to get the allocation function from. Now its just a
pain to pass around everywhere.
Also added a ldb_debug_set() function that calls ldb_debug() plus sets
the result using ldb_set_errstring(). That saves on some awkward
coding in a few places.
- ensure we don't add attributes twice, should a user ask for the
attribute twice. Do this in such a way that we don't become O(n^2)
- removed some unused code
to msync/mmap not changing the mtime of the file. This patch ensures
that for successfully completed transactions we update the mtime.
I don't do this on all tdb writes as its too expensive, but doing it
just on transactions is bearable, as those cost quite a lot anyway.
* rename the composite helper functions from comp_* to composite_*
* Move the lsa initialization to wb_connect_lsa.c
* Equip smb_composite_connect with a fallback_to_anonymous
The latter two simplify wb_init_domain.c quite a bit.
Volker
this is the compression algorithm used by w2k3 for DsGetNCChanges().
This algorithm isn't known yet, but it seems to be some sort of Lempel-Ziv
algorithm.
metze
metze
r8017@SERNOX: metze | 2005-06-30 13:44:23 +0200
create the SAMBA_4_0 branch for the Summer Of Code Project
metze
r8730@SERNOX: brad | 2005-07-24 03:09:48 +0200
Branching Samba 4
r8731@SERNOX: brad | 2005-07-24 06:39:00 +0200
added 'make installmisc' to howto.txt
added existing 'compression' option to level8 drsuapi torture test
added new 'neighbour_writeable' option to level8 drsuapi torture test
r8732@SERNOX: brad | 2005-07-24 06:42:38 +0200
added metze's dssync patch as source/torture/rpc/dssync.c
r8739@SERNOX: brad | 2005-07-25 00:24:46 +0200
added a test called RPC-DSSYNC to config.mk
hacking at dssync.c in an attempt to make it compile
r8754@SERNOX: brad | 2005-07-25 15:19:21 +0200
Changing dssync.c to use ldb routines for accessing ldap rather than raw ldap calls.
r8765@SERNOX: brad | 2005-07-26 03:35:38 +0200
more ldb changes to test_CompleteJoin(), it mostly kind of almost works now!
r8766@SERNOX: brad | 2005-07-26 03:56:00 +0200
Trying to fix the crazy nesting in the branch
r8769@SERNOX: brad | 2005-07-26 04:48:29 +0200
merging latest changes
r8770@SERNOX: brad | 2005-07-26 04:53:43 +0200
removing nested branch
r8793@SERNOX: jerry | 2005-07-27 05:04:57 +0200
merging on of Brad missing changes from the nested 4.0 branch debacle
r8794@SERNOX: jerry | 2005-07-27 05:14:42 +0200
syncing up with the main 4_0 branch for Brad
r8842@SERNOX: brad | 2005-07-29 00:26:30 +0200
merging changes from branches/SAMBA_4_0
r8850@SERNOX: brad | 2005-07-29 21:07:57 +0200
Bringing my tree up to date
r8851@SERNOX: brad | 2005-07-30 00:48:04 +0200
making dssync.c more ldb-centric, reverted samlogon.c from rev. 8845 to get my branch to compile again.
r8856@SERNOX: brad | 2005-07-30 03:20:33 +0200
I think I have the ldb code down in test_CompleteJoin (not complete yet though)
r8860@SERNOX: brad | 2005-07-30 07:08:13 +0200
Changed comments to C style /**/ (thanks Richard), some more changes to test_CompleteJoin().
r8862@SERNOX: brad | 2005-07-31 04:45:32 +0200
Bringing the SOC/SAMBA_4_0 branch up to date.
r8863@SERNOX: brad | 2005-07-31 20:00:41 +0200
Updated some missing files from the branch
r8864@SERNOX: brad | 2005-07-31 20:25:50 +0200
Removing autogenerated files from branch
r8865@SERNOX: brad | 2005-07-31 20:43:58 +0200
last of the unneeded files in SOC/SAMBA_4_0
r9004@SERNOX: brad | 2005-08-03 18:51:23 +0200
r5214@buttercup: j0j0 | 2005-08-03 10:44:30 -0600
r@buttercup: j0j0 | 2005-08-02 22:54:13 -0600
creating a local branch of branches/SAMBA_4_0
r9013@SERNOX: brad | 2005-08-03 20:57:48 +0200
r5228@buttercup: j0j0 | 2005-08-03 13:00:11 -0600
Fixing differences between this branch and /branches/SAMBA_4_0
r9014@SERNOX: brad | 2005-08-03 21:18:05 +0200
r5231@buttercup: j0j0 | 2005-08-03 13:23:12 -0600
Updating config.mk so that smbtorture builds again
r9061@SERNOX: brad | 2005-08-04 18:17:36 +0200
r5249@buttercup: j0j0 | 2005-08-03 21:01:02 -0600
Start using libnet_Join() for DC join.
r9062@SERNOX: brad | 2005-08-04 18:17:47 +0200
r5250@buttercup: j0j0 | 2005-08-04 10:21:34 -0600
Some more work towards performing a dc join.
r9064@SERNOX: brad | 2005-08-04 18:53:51 +0200
r5253@buttercup: j0j0 | 2005-08-04 10:53:00 -0600
Fixed a bug (passing a TALLOC_CTX to libnet_context_init() )
r9069@SERNOX: brad | 2005-08-04 21:59:55 +0200
r5279@buttercup: j0j0 | 2005-08-04 14:04:55 -0600
Some more work on the domain join
r9117@SERNOX: brad | 2005-08-05 16:50:26 +0200
r5281@buttercup: j0j0 | 2005-08-05 08:55:58 -0600
Committing minor changes before merge
r9180@SERNOX: brad | 2005-08-07 17:25:25 +0200
r5314@buttercup: j0j0 | 2005-08-07 09:30:12 -0600
Reworked libnet_join to use two join levels, AUTOMATIC and SPECIFIED.
r9181@SERNOX: brad | 2005-08-07 17:25:36 +0200
r5315@buttercup: j0j0 | 2005-08-07 09:31:22 -0600
Working with libnet_Join(), code cleanup needed in the near future.
r9192@SERNOX: brad | 2005-08-07 21:40:22 +0200
r5373@buttercup: j0j0 | 2005-08-07 13:46:09 -0600
Some code cleanup to make things a little more readable.
r9249@SERNOX: brad | 2005-08-12 01:31:48 +0200
r5375@buttercup: j0j0 | 2005-08-11 17:38:44 -0600
Split libnet_JoinDomain() into libnet_JoinDomain() and libnet_JoinADSDomain().
r9256@SERNOX: brad | 2005-08-12 04:55:11 +0200
r5413@buttercup: j0j0 | 2005-08-11 21:02:27 -0600
Clean up libnet_JoinADSDomain() a little, added a comment to the test_join struct.
r9314@SERNOX: brad | 2005-08-16 03:53:20 +0200
r5436@buttercup: j0j0 | 2005-08-15 20:01:21 -0600
libnet_JoinDomain() should honour LIBNET_JOIN_TORTURE now.
torture_join_domain() should properly use libnet_JoinDomain().
dssync.c uses torture_join_domain() again.
r9351@SERNOX: brad | 2005-08-17 07:15:31 +0200
r5438@buttercup: j0j0 | 2005-08-16 23:23:58 -0600
Removed LIBNET_JOIN_TORTURE level, as it became unnecessary once libnet_Join_primary_domain() handled netbios names better.
Corrected libnet_JoinDomain() and libnet_JoinADSDomain().
r9352@SERNOX: brad | 2005-08-17 07:24:49 +0200
r5440@buttercup: j0j0 | 2005-08-16 23:33:25 -0600
Fixed a typo.
r9354@SERNOX: metze | 2005-08-17 10:28:25 +0200
remove object files from svn
metze
r9376@SERNOX: brad | 2005-08-18 05:15:48 +0200
r5476@buttercup: j0j0 | 2005-08-17 21:24:33 -0600
Proof that I shouldn't code when i'm tired (silly bugfixes).
r9405@SERNOX: brad | 2005-08-19 22:50:10 +0200
r5500@buttercup: j0j0 | 2005-08-19 14:56:25 -0600
Get dssync.c compiling again after merge (ldb_dn changes from rev. 9391).
r9407@SERNOX: brad | 2005-08-20 03:22:42 +0200
r5502@buttercup: j0j0 | 2005-08-19 19:28:22 -0600
libnet/libnet_join.c
Some more fixes so ldb uses ldb_dn's.
torture/rpc/dssync.c
Some debugging printf()'s.
ldb_dn fixes.
torture/rpc/testjoin.c
Change torture_join_domain() to use libnet_JoinDomain() rather than libnet_Join().
Some more debugging statements.
I'm not sure why, but GUID_all_zero(user_handle.uuid) is returning true in torture_leave_domain() when called it from torture_destroy_context() in torture/rpc/dssync.c.
That's what i'm working out now.
r9427@SERNOX: brad | 2005-08-20 18:38:29 +0200
r5504@buttercup: j0j0 | 2005-08-20 10:44:52 -0600
Some bugfixes.
Removed a bunch of debugging code.
torture_leave_domain() works again! not 100% perfect yet though...
r9428@SERNOX: brad | 2005-08-20 19:09:26 +0200
r5506@buttercup: j0j0 | 2005-08-20 11:15:54 -0600
Restructure torture_join_domain() so that it joins itself, removes itself, and joins itself to the domain again to ensure that its account information is all current and as expected.
r9452@SERNOX: brad | 2005-08-21 19:33:51 +0200
r5508@buttercup: j0j0 | 2005-08-21 11:40:36 -0600
Bugfixes, trying to get things straight between contexts.
r9467@SERNOX: brad | 2005-08-22 04:00:48 +0200
r5510@buttercup: j0j0 | 2005-08-21 20:06:55 -0600
Another round of bugfixing.
r9521@SERNOX: brad | 2005-08-23 15:26:44 +0200
r5596@buttercup: j0j0 | 2005-08-23 07:33:06 -0600
Merging changes
r9524@SERNOX: metze | 2005-08-23 16:09:42 +0200
- fix the build caused by changes in the main samba4 tree,
- add an option "dssync:german=yes" to allow me to run against my german w2k3 server
this should be replaces by CLDAP calls to get the Default-First-Site-Name dynamicly
- remove some temporary comments, as DsAddEntry works now
metze
r9528@SERNOX: metze | 2005-08-23 18:22:22 +0200
the RPC-DSSYNC test is now able to fetch the whole tree,
including the unicodePwd, ntPwdHistory fields
metze
r9559@SERNOX: brad | 2005-08-24 04:11:47 +0200
r5612@buttercup: j0j0 | 2005-08-23 20:19:12 -0600
Some fixes around using talloc in a hierarchical fashion.
Still not right, but better.
r9564@SERNOX: brad | 2005-08-24 05:43:11 +0200
r5614@buttercup: j0j0 | 2005-08-23 21:50:38 -0600
Gave libnet_JoinADSDomain() its own tmp_ctx rather than passing it from libnet_JoinDomain() as a parameter (yuk).
As a side effect, it proves that my bug lies in libnet_JoinDomain(), not libnet_JoinADSDomain().
r9565@SERNOX: brad | 2005-08-24 06:09:46 +0200
r5616@buttercup: j0j0 | 2005-08-23 22:17:12 -0600
Small fix, if r->out.error_string and r2->samr_handle.out.error_string weren't set to NULL, torture_join_domain() would segfault on the second join.
r9630@SERNOX: brad | 2005-08-26 06:42:50 +0200
Commented out the parts of the dssync test which perform the dc join and create/remove associated ldap entries.
Commented out the test for the 'german' dssync option, because now we detect the Site-Name using CLDAP. If cldap_netlogon() does not return ok, the code defaults to 'Default-First-Site-Name'.
r9670@SERNOX: brad | 2005-08-27 02:30:11 +0200
Added a patch from metze.
To showcase what i've learned today, i've created two new parameters which can be set at runtime, drsuapi:last_usn and drsuapi:partition.
drsuapi:last_usn takes an integer representing the USN of the last recieved replication update for a particular partition (uses the domain dn if drsuapi:parition isn't set).
That value is passed in the DsGetNCChanges() call so that only info which has been updated since that point in time is returned. If this option is not set, 0 is used by default, and all updates for that partition are returned.
drsuapi:partition takes a string dn and uses that as the name of the AD partition to replicate.
Some debugging output was also added.
r9723@SERNOX: brad | 2005-08-29 01:07:51 +0200
Added some copyright notices.
Changed some things in net_join.c to try and figure out why 'net join <domain> bdc' segfaults.
It occurs when the last talloc_free() happens, so i'm sure it's something to do with the memory fiddling i'm doing in libnet_join.
Added some drsuapi attribute ids that I figured out today.
I put some (many, dry) notes together while doing that, so i'll try to put them up on a blog at samba.org a little later tonight.
r9740@SERNOX: metze | 2005-08-29 16:58:03 +0200
fix up the DsGetNCchanges loop,
and remove misleading comments
metze
r9743@SERNOX: metze | 2005-08-29 17:26:45 +0200
make the logic a bit clearer
metze
r9815@SERNOX: brad | 2005-08-31 02:36:21 +0200
Added cldap_netlogon() AD Site-Name lookup into libnet/libnet_join.c.
Bugfixing rampage in libnet_join.c to resolve misunderstanding of talloc_steal().
libnet_join now creates the CN=<netbios name>,CN=Servers,CN=<site name>,CN=Sites,CN=Configuration,<domain dn> container on a dc join.
r9858@SERNOX: brad | 2005-09-01 03:17:17 +0200
Removed extraneous NDR_ALL subsystem requirement from torture/config.mk.
Added lots of error checking as per metze's advice.
Removed commented out code.
More bug chasing.
r9863@SERNOX: brad | 2005-09-01 05:53:19 +0200
Cleaned up dssync.c, removed the unneeded DsCrackNames() call, removed DC join/leave related stuff.
It no longer looks like my house does!
r9887@SERNOX: metze | 2005-09-01 11:34:03 +0200
- fix dssync:highest_usn parameter handling
- ask for LINKED_ATTRIBUTE replication
metze
r9891@SERNOX: metze | 2005-09-01 14:13:18 +0200
make the code more readable, and fix a few bugs
metze
r9911@SERNOX: brad | 2005-09-01 20:36:27 +0200
Bugfixes in libnet_join.c.
Cleaned up comments.
Added domain_dn_str and account_dn_str to struct libnet_JoinDomain.
Removed struct dcerpc_pipe *samr_pipe and struct policy_handle user_handle from struct libnet_Join.
r9920@SERNOX: brad | 2005-09-01 23:34:13 +0200
Added disclaimer (I can't seem to get libnet_JoinDomain() to keep the samr_pipe and u_handle open past the function call, grrrr....).
r9921@SERNOX: brad | 2005-09-01 23:37:54 +0200
Added copyright statement.
Cleaned up unneeded variables from torture_join_domain().
r9932@SERNOX: brad | 2005-09-02 01:49:42 +0200
Really rushed project notes.
r10841@SERNOX: metze | 2005-10-08 20:01:45 +0200
remove diff to main SAMBA_4_0 branch
metze
r10862@SERNOX: metze | 2005-10-10 10:31:52 +0200
remove the differences between SAMBA_4_0 and SOC/SAMBA_4_0
metze
r10863@SERNOX: metze | 2005-10-10 10:34:26 +0200
fix the build
metze
r10864@SERNOX: metze | 2005-10-10 11:10:08 +0200
remove README file to reduce, diffs to main SAMBA_4_0 branch:
metze
README:
This project was centered around adding a torture test to Samba 4, which used drsuapi_DsGetNCChanges() to retrieve the contents of an Active Directory in the same manner as an Active Directory DC replication event.
As the project unfolded, I also applied some changes to the functionality of the libnet library related to joining a machine account to a domain.
One of the first things that I implemented in this project was a 'neighbour_writeable' option for the RPC-DRSUAPI torture test. The command line to execute this torture test is as follows:
smbtorture --option=drsuapi:neighbour_writeable=True -W <domain name> -U <admin username>%<password> ncacn_ip_tcp:<domain controller dns name> RPC-DRSUAPI
This option provides us with runtime control over the DRSUAPI_DS_REPLICA_NEIGHBOUR_WRITEABLE flag in the struct drsuapi_DsGetNCChanges.in.req.req<level>.replica_flags, allowing us to easily test for differences in the behaviour of AD replication with the switch on or off.
In the course of the project, I also implemented two more flags for the RPC-DSSYNC test. dssync:last_usn takes an integer representing the USN (Universal Serial Number) of the last recieved replication update for a particular partition (uses the domain DN if drsuapi:parition isn't set). That value is passed in the DsGetNCChanges() call so that only info which has been updated since that point in time is returned. If this option is not set, 0 is used by default, and all updates for that partition are returned. dssync:partition takes a string DN and uses that as the name of the AD partition to replicate.
Based initially on a patch provided to me by one of my mentors, Stephan (metze) Metzmacher, the RPC-DSSYNC test was implemented for this project. Initially functionality was included to perform a DC join prior to initiating replication, but the code was removed when it was realized that replication could indeed take place without being a member of the domain in any way. It has been recently suggested that we may need a DC join after all to get all of the information we may want from the AD replication. This is probably best added using a torture_join_domain() call once the libnet code is able to keep the user policy handle and SAMR RPC pipe open.
The DC join code was taken out of the RPC-DSSYNC and implemented for the most part in the libnet libraries. To test this, the RPC-NETLOGON test was modified to perform a domain join, leave and rejoin. Currently, the test has a fault in that it is unable to leave the domain using the same SAMR RPC pipe and user_policy information as was used for the first join. This is because I was unable to get the code working properly in libnet to provide that functionality. Currently missing from the DC join in libnet is the code to create the CN=NTDS Settings,CN=<DC NETBIOS NAME>,CN=<Site-Name>,CN=Sites,CN=Configuration,<domain DN> container using the dcerpc_drsuapi_DsAddEntry() call. I did not want to implement this functionality in libnet while there were still problems with the code.
I also provided the ability in libnet and the RPC-DSSYNC test to look up the proper site name using the cldap library.
In my investigations, I was unable to find out any information regarding the UnicodePwd attribute, except that the same password is represented differently for two different users in the same directory.
I was also able to resolve and confirm the meaning of some DRSUAPI_ATTRIBUTE ID's.
DRSUAPI_OBJECTCLASS_domain (0xA0042)
DRSUAPI_OBJECTCLASS_domainDNS (0xA0043)
wellKnownObjects (0x9026A)
fSMORoleOwner (0x90171)
name or dc (0x90001)
whenCreated (0x20002)
instanceType (0x20001)
gPLink (0x9037B)
These were added to the IDL for drsuapi (source/librpc/idl/drsuapi.idl).
I would like to thank everyone on the Samba team who worked with me and assisted me with this project, specifically all the work done by Stephan Metzmacher, Andrew Bartlett and Jerry Carter. Working on this project with the Samba team really has been a life changing experience, as corny as that sounds.
I've realized that I was born to be a systems developer, and it has helped confirm in my mind that Open Source (specifically Samba) development is exactly what i've been missing!
I would also like to take this opportunity to thank Chris Dibona and Google for the amazing opportunity. I don't know if I would have taken the leap in other circumstances.
I know these notes sound a little rushed, but it is 23:55 after all! :)
Initialize a domain structure properly. Excerpt from wb_init_domain.c:
/*
* Initialize a domain:
*
* - With schannel credentials, try to open the SMB connection with the machine
* creds. Fall back to anonymous.
*
* - If we have schannel creds, do the auth2 and open the schannel'ed netlogon
* pipe.
*
* - Open LSA. If we have machine creds, try to open with ntlmssp. Fall back
* to schannel and then to anon bind.
*
* - With queryinfopolicy, verify that we're talking to the right domain
*
* A bit complex, but with all the combinations I think it's the best we can
* get. NT4, W2k3SP1 and W2k all have different combinations, but in the end we
* have a signed&sealed lsa connection on all of them.
*
* Is this overkill? In particular the authenticated SMB connection seems a
* bit overkill, given that we do schannel for netlogon and ntlmssp for
* lsa later on w2k3, the others don't do this anyway.
*/
Thanks to Jeremy for his detective work, and to the Samba4 team for providing
such a great infrastructure.
Next step is to connect to SAM. Do it via LDAP if we can, fall back to samr
with all we have.
Volker
start to look sane.
Question: What about providing all winbind commands as irpc interfaces that
are called from the samba3 compatibility layer? This way it would be easy for
other samba components to access its functionality. Does that make sense?
Volker
authenticated session down into LDB. This associates a session info
structure with the open LDB, allowing a future ldb_ntacl module to
allow/deny operations on that basis.
Along the way, I cleaned up a few things, and added new helper functions
to assist. In particular the LSA pipe uses simpler queries for some of
the setup.
In ldap_server, I have removed the 'ldasrv:hacked' module, which hasn't
been worked on (other than making it continue to compile) since January,
and I think the features of this module are being put into ldb anyway.
I have also changed the partitions in ldap_server to be initialised
after the connection, with the private pointer used to associate the ldb
with the incoming session.
Andrew Bartlett
