Andrew Tridgell [Tue, 29 Nov 2005 08:55:13 +0000 (08:55 +0000)]
r11955: got rid of the old rootDSE code in the ldap server.
The partitioning logic is still there, but we only have one
partition. If we need partitioning in the future it might be better to
remove this partitioning code and use a partitioning module instead
Andrew Tridgell [Tue, 29 Nov 2005 08:52:41 +0000 (08:52 +0000)]
r11954: add the static rootdse content to the sam ldb,and enable the rootdse
module in @MODULES
Andrew Tridgell [Tue, 29 Nov 2005 08:51:36 +0000 (08:51 +0000)]
r11953: enabled the rootdse module in the ldb modules code
Andrew Tridgell [Tue, 29 Nov 2005 08:50:52 +0000 (08:50 +0000)]
r11952: added a rootdse module. This will replace the existing rootdse code in
the ldap server. The reason for the change is that ldb modules need
some way to get at the static info stored in the rootDSE (such as the
location of the schema) but they can't do that right now
Andrew Tridgell [Mon, 28 Nov 2005 22:53:42 +0000 (22:53 +0000)]
r11949: make sure we ask gensec to give us a session key
andrew, this answers your question on irc about whether the same
session key mechanisms are used in smb2. They are - the RPC-LSA secret
tests pass fine over ncacn_np on SMB2, which means the session key
must be working
Stefan Metzmacher [Mon, 28 Nov 2005 13:15:57 +0000 (13:15 +0000)]
r11941: fix cut'n'paste bug
metze
Andrew Bartlett [Mon, 28 Nov 2005 07:59:46 +0000 (07:59 +0000)]
r11940: Love has clarified why this code does what it does.
Andrew Bartlett
Andrew Bartlett [Sun, 27 Nov 2005 02:03:42 +0000 (02:03 +0000)]
r11931: Add a short README explaining what this directory is all about.
Andrew Bartlett
Andrew Bartlett [Sun, 27 Nov 2005 02:02:44 +0000 (02:02 +0000)]
r11930: Add socket/packet handling code for kpasswdd
Allow ticket requests with only a netbios name to be considered 'null'
addresses, and therefore allowed by default.
Use the netbios address as the workstation name for the allowed
workstations check with krb5.
Andrew Bartlett
Andrew Bartlett [Sun, 27 Nov 2005 02:00:37 +0000 (02:00 +0000)]
r11929: Add static, comments.
Andrew Bartlett
Andrew Bartlett [Sun, 27 Nov 2005 02:00:12 +0000 (02:00 +0000)]
r11928: More Kerberos musings...
Andrew Bartlett
Stefan Metzmacher [Fri, 25 Nov 2005 15:36:47 +0000 (15:36 +0000)]
r11913: if we have a UNIQUE name with more than 1 address,
it becomes implicit an MHOMED record
metze
Stefan Metzmacher [Fri, 25 Nov 2005 15:30:35 +0000 (15:30 +0000)]
r11912: fix nbt_name_registration, there's still some minor stuff todo,
e.g. to return the first address of the 0x1B address as first
address in the 0x1C reply, and handle sgroup merge overflow
of 25 addresses
metze
Stefan Metzmacher [Fri, 25 Nov 2005 13:44:16 +0000 (13:44 +0000)]
r11911: as we pass the owned_released vs. replica test now, run it with make test
metze
Stefan Metzmacher [Fri, 25 Nov 2005 13:43:12 +0000 (13:43 +0000)]
r11910: fix nbt_name_release and nbt_name_query, so that we pass the owned_released vs. replica
winsrepl torture test
metze
Stefan Metzmacher [Fri, 25 Nov 2005 12:03:40 +0000 (12:03 +0000)]
r11908: implement SGROUP merging, that passes the different owner tests
(but only without socket_wrapper, I need to look at that later
and then add the different_owner test to NBT-WINSREPLICATION-QUICK
so that it'll be runned by make test)
metze
Andrew Tridgell [Fri, 25 Nov 2005 11:51:47 +0000 (11:51 +0000)]
r11907: added testing of SMB2 keepalive
Andrew Tridgell [Fri, 25 Nov 2005 11:51:15 +0000 (11:51 +0000)]
r11906: opcode 13 appears to be keepalive. Metze guessed this one :-)
Andrew Tridgell [Fri, 25 Nov 2005 11:33:57 +0000 (11:33 +0000)]
r11905: added SMB2_FLUSH as opcode 7. Thanks to metze and volker for help
brainstorming this one.
Andrew Tridgell [Fri, 25 Nov 2005 11:12:08 +0000 (11:12 +0000)]
r11904: added smb2_tdis() testing
Andrew Tridgell [Fri, 25 Nov 2005 11:11:47 +0000 (11:11 +0000)]
r11903: added smb2_tdis() (opcode 4)
Andrew Tridgell [Fri, 25 Nov 2005 11:05:21 +0000 (11:05 +0000)]
r11902: added smb2_logoff() testing
Andrew Tridgell [Fri, 25 Nov 2005 11:04:42 +0000 (11:04 +0000)]
r11901: added smb2_logoff() support (metze correctly guessed opcode 2 was
logoff)
Stefan Metzmacher [Fri, 25 Nov 2005 10:46:38 +0000 (10:46 +0000)]
r11900: - make sure address and registered_by are replaced when they're not present
- make sure we don't add active records with 0 addresses
metze
Stefan Metzmacher [Fri, 25 Nov 2005 10:44:03 +0000 (10:44 +0000)]
r11899: add some usefull debug messages
metze
Stefan Metzmacher [Fri, 25 Nov 2005 10:11:01 +0000 (10:11 +0000)]
r11897: add 2 more sgroup vs. sgroup tests with the replica having no addresses
metze
Stefan Metzmacher [Fri, 25 Nov 2005 10:08:31 +0000 (10:08 +0000)]
r11896: max_version of 0 means unlimited
metze
Stefan Metzmacher [Fri, 25 Nov 2005 08:24:36 +0000 (08:24 +0000)]
r11895: - reorder some code to make it easier to follow, how the fields appear on the wire
- add some comments to the header file, to represent the wire format
metze
Andrew Tridgell [Fri, 25 Nov 2005 06:50:29 +0000 (06:50 +0000)]
r11894: fixed SMB2 trans code for pipe_flags
Andrew Tridgell [Fri, 25 Nov 2005 06:48:12 +0000 (06:48 +0000)]
r11893: fixed a dependency problem
Andrew Tridgell [Fri, 25 Nov 2005 05:56:16 +0000 (05:56 +0000)]
r11892: forgot to commit these changes
Andrew Tridgell [Fri, 25 Nov 2005 05:46:46 +0000 (05:46 +0000)]
r11891: - added pipe_flags field in smb2_trans
- while running dcerpc over SMB2, the server will occasionally send us
a oh-so-useful STATUS_PENDING result meaning "I don't have a result
for you yet, but I'm working on it". These can be discarded :-)
Andrew Tridgell [Fri, 25 Nov 2005 05:26:12 +0000 (05:26 +0000)]
r11890: added tests for the last few fields in SMB2 find requests
Andrew Tridgell [Fri, 25 Nov 2005 05:25:37 +0000 (05:25 +0000)]
r11889: added support for dcerpc ncacn_np over SMB2. You use it by giving the
flag 'smb2' in the dcerpc binding string. This gives a pretty good
test to the new SMB2 trans call.
Andrew Tridgell [Fri, 25 Nov 2005 05:23:55 +0000 (05:23 +0000)]
r11888: - added SMB2 trans support
- added session key to SMB2
- renamed 'unknown2' in create to 'impersonation'
Stefan Metzmacher [Thu, 24 Nov 2005 09:22:38 +0000 (09:22 +0000)]
r11887: reorder some tests
metze
Volker Lendecke [Wed, 23 Nov 2005 18:49:30 +0000 (18:49 +0000)]
r11885: Add forgotten files
Volker Lendecke [Wed, 23 Nov 2005 17:21:08 +0000 (17:21 +0000)]
r11884: Download ntconfig.pol
Volker Lendecke [Wed, 23 Nov 2005 15:52:23 +0000 (15:52 +0000)]
r11883: Also look up the membership in the domain local groups.
Volker
Stefan Metzmacher [Wed, 23 Nov 2005 12:30:57 +0000 (12:30 +0000)]
r11882: - use some better names
- we now pass the same_owner tests so test them with make test
metze
Stefan Metzmacher [Wed, 23 Nov 2005 12:29:37 +0000 (12:29 +0000)]
r11881: fix debug messages
metze
Stefan Metzmacher [Wed, 23 Nov 2005 12:24:12 +0000 (12:24 +0000)]
r11880: - we finally start to apply replicas to our database
- currently only the ADD, REPLACE and NOT REPLACE cases are handled complete
- other cases are just skiped for now
metze
Stefan Metzmacher [Wed, 23 Nov 2005 12:19:38 +0000 (12:19 +0000)]
r11879: some fixes for the new ldb api
metze
Stefan Metzmacher [Wed, 23 Nov 2005 11:37:54 +0000 (11:37 +0000)]
r11878: don't get only newer records than the last one we got
metze
Stefan Metzmacher [Wed, 23 Nov 2005 11:34:46 +0000 (11:34 +0000)]
r11877: - give winsdb_add/modify/delete() ldb_context as first argument
and add a flags argument to add and modify,the current flags are:
WINSDB_FLAG_ALLOC_VERSION to allocate a new version id for the record
WINSDB_FLAG_TAKE_OWNERSHIP to take the become the wins owner of the record
- fix handling of records with no addresses, this is valid for MHOMED and SGROUP
records when they're not in ACTIVE state
metze
Stefan Metzmacher [Wed, 23 Nov 2005 09:38:44 +0000 (09:38 +0000)]
r11873: add some interval vars to the service context
metze
Andrew Tridgell [Wed, 23 Nov 2005 07:33:09 +0000 (07:33 +0000)]
r11872: another attempt at fixing the data_blob_free in async dcerpc bind,
after volker gently pointed out that I should be wearing a paper bag
over my head for the last attempt :-)
Andrew Tridgell [Wed, 23 Nov 2005 00:51:24 +0000 (00:51 +0000)]
r11871: fixed a problem volker found with the async bind code, and a callback
destroying the dcerpc pipe
Andrew Tridgell [Wed, 23 Nov 2005 00:30:58 +0000 (00:30 +0000)]
r11870: fixed the problem volker reported with the RPX-XPLOGIN test. The
problem was caused by a callback destroying the packet processing
context while that context was being used in packet_recv()
This is the first time we have used the ability of talloc destructors
to 'refuse' a free request. It works well in this case as it makes the
composite API simpler to use for other code, and isolates the
complexity of having callbacks destroying the packet context to the
packet.c code.
Volker Lendecke [Tue, 22 Nov 2005 21:45:05 +0000 (21:45 +0000)]
r11869: talloc has been ported :-)
Volker Lendecke [Tue, 22 Nov 2005 20:55:59 +0000 (20:55 +0000)]
r11868: Get the wks and user accounts from a file.
Abartlet, please don't break this again, it's too darn useful for
near-realistic load tests on an RPC infrastructure and can be quite easily
expanded to more weird things a workstation might do during a login.
Yes, I promise I will document this test, but this must wait until the
weekend.
I might add simulating a profile download quite soon, we have the information
available from the info3.
Thanks,
Volker
Andrew Tridgell [Tue, 22 Nov 2005 13:14:04 +0000 (13:14 +0000)]
r11850: added a test suite for the SMB2 find calls
Andrew Tridgell [Tue, 22 Nov 2005 13:13:17 +0000 (13:13 +0000)]
r11849: added mapping between SMB2 and SMB find/search levels
Andrew Tridgell [Tue, 22 Nov 2005 13:12:22 +0000 (13:12 +0000)]
r11848: separate out the info levels common to SMB and SMB2 for raw_search
Andrew Tridgell [Tue, 22 Nov 2005 06:05:50 +0000 (06:05 +0000)]
r11843: fixed a valgrind error in the RPC-SAMLOGON test
Andrew Tridgell [Tue, 22 Nov 2005 03:21:42 +0000 (03:21 +0000)]
r11838: lower the default numops in smbtorture. When people want lots of
repeated operations they should manually specify --num-ops=XX
this should speed up a few of the build farm tests
Volker Lendecke [Mon, 21 Nov 2005 21:13:01 +0000 (21:13 +0000)]
r11837: Add a timeout before starting a torture test. It's really fun to watch 50
simulated logins doing stuff simultaneously. This is where the async
architecture *really* pays off.
Volker
Volker Lendecke [Mon, 21 Nov 2005 20:42:18 +0000 (20:42 +0000)]
r11836: More work in progress
Volker Lendecke [Mon, 21 Nov 2005 19:43:06 +0000 (19:43 +0000)]
r11835: Restore comments
Volker Lendecke [Mon, 21 Nov 2005 17:16:50 +0000 (17:16 +0000)]
r11834: More steps towards XP login
Stefan Metzmacher [Mon, 21 Nov 2005 14:21:56 +0000 (14:21 +0000)]
r11832: - some updates for the conflict resolving logic
- use better function names
metze
Stefan Metzmacher [Mon, 21 Nov 2005 14:14:58 +0000 (14:14 +0000)]
r11831: add some SGROUP not active vs. SGROUP any tests
metze
Stefan Metzmacher [Mon, 21 Nov 2005 12:21:52 +0000 (12:21 +0000)]
r11829: remove unused #if 1
metze
Stefan Metzmacher [Mon, 21 Nov 2005 12:16:45 +0000 (12:16 +0000)]
r11828: split out the async helper function into a new subsystem WB_HELPER
to use it in torture tests too
metze
Stefan Metzmacher [Mon, 21 Nov 2005 12:05:46 +0000 (12:05 +0000)]
r11827: add more sgroup vs sgroup tests
metze
Volker Lendecke [Mon, 21 Nov 2005 11:35:15 +0000 (11:35 +0000)]
r11826: A few steps down the login
Volker Lendecke [Mon, 21 Nov 2005 10:28:54 +0000 (10:28 +0000)]
r11825: Fix a debug msg
Andrew Tridgell [Mon, 21 Nov 2005 09:54:02 +0000 (09:54 +0000)]
r11824: fixed a valgrind error in the dcerpc_smb code
Andrew Tridgell [Mon, 21 Nov 2005 09:31:36 +0000 (09:31 +0000)]
r11823: make the socket_connect_send() context a child of the local state
this fixes a valgrind error
Andrew Tridgell [Mon, 21 Nov 2005 09:20:32 +0000 (09:20 +0000)]
r11822: reworked the socket_connect_multi() code so it is built on top of
socket_connect_send() rather than the lower level socket code. Also
simplified the state structures a fair bit, and added name resolution,
fixing a bug where the multi-port connect code did a separate name
resolution for each port being tried.
Andrew Tridgell [Mon, 21 Nov 2005 09:17:00 +0000 (09:17 +0000)]
r11821: got rid of two more unnecessary variables and made the variable names
a bit more consistent
Andrew Tridgell [Mon, 21 Nov 2005 07:23:36 +0000 (07:23 +0000)]
r11820: fixed some problems with the socket socket.c code.
- removed the duplicate calls to socket_connect(), instead creating a
common function socket_send_connect() used by both code paths
- fixed some NULL ptr checks (probably was cut-and-paste bugs)
- ensure we use the result of the name resolution
- added a few comments
- use 'fde' for the file description event. The variable name
'connect_ev' immediately made me think of an event context, not a
fde. Using common variable name conventions makes code a bit easier
to read
Andrew Tridgell [Mon, 21 Nov 2005 06:01:02 +0000 (06:01 +0000)]
r11819: simplified the async rpc bind code a little.
- removed the struct dcerpc_request_state as all the state
information is already available on the dcerpc_pipe structure, so just
use that
- added a single dcerpc_recv_data() handler for receiving packets
from the transport layer. This then does the initial decoding of
the dcerpc packet, and then looks at the packet type in order to
work out who to dispatch it to. This should allow in-flight async
rpc requests to still work while a new bind or alter context is
happening
- ensure that if the transport indicates the connection is dead that
any in-flight bind or alter context requests are given an error
- removed full_request_private and instead use separate bind_private
and alter_private pointers
- added a few comments for some bits I found hard to understand
Andrew Tridgell [Mon, 21 Nov 2005 04:18:43 +0000 (04:18 +0000)]
r11818: - changed the option torture:echo_TestSleep=yes/no to the more generic
option torture:quick=yes/no. This should be used in all slow tests to
enable a quick mode
- enabled the test_rpc_quick.sh tests in 'make quicktest'
Andrew Tridgell [Mon, 21 Nov 2005 04:09:36 +0000 (04:09 +0000)]
r11817: fixed the problem with the RPC join tests. The problem was that
revision 11809 had removed the change to the session key function
after authentication succeeds.
Andrew Tridgell [Mon, 21 Nov 2005 02:52:33 +0000 (02:52 +0000)]
r11816: this fixes some of the problems with the recent async rpc changes and
ncacn_ip_tcp/ncalrpc. The problem was that svn revision 11809 removed
the logic that forced the CONNECT auth type for authenticated binds
which don't have an explicit SIGN or SEAL flag set.
Rafal Szczesniak [Sun, 20 Nov 2005 22:00:47 +0000 (22:00 +0000)]
r11815: A bit more comments and spaces for better readability.
rafal
Rafal Szczesniak [Sun, 20 Nov 2005 21:40:41 +0000 (21:40 +0000)]
r11814: 1) Set name resolution methods according to smb.conf
2) Set credentials workstation name, otherwise rpc bind function
segfaults on auth stage
rafal
Rafal Szczesniak [Sun, 20 Nov 2005 21:33:30 +0000 (21:33 +0000)]
r11813: Const-ify name resolution method list and use string list
utilities to set the context field.
rafal
Volker Lendecke [Sun, 20 Nov 2005 17:34:56 +0000 (17:34 +0000)]
r11812: Convert winbind to the async bind routines. Also remove tridge's hack for the
winbind "bug" :-)
Volker
Volker Lendecke [Sun, 20 Nov 2005 17:33:08 +0000 (17:33 +0000)]
r11811: Revert 11808, this broke the selftest. No idea why though. Why do we to the
resolve_name if it's not used?
I know this is my code, but I don't understand why it breaks tests.
Volker
Volker Lendecke [Sun, 20 Nov 2005 17:02:28 +0000 (17:02 +0000)]
r11810: Fix talloc hierarchy
Volker Lendecke [Sun, 20 Nov 2005 16:28:39 +0000 (16:28 +0000)]
r11809: Make dcerpc_bind_auth async.
This also removes dcerpc_bind_auth_password, the only user of
dcerpc_bind_auth. And this was not only passwords anyway.
Andrew Bartlett, as usual: Please take a close look.
Thanks,
Volker
Volker Lendecke [Sun, 20 Nov 2005 16:27:41 +0000 (16:27 +0000)]
r11808: Fix socket_connect_ev
Volker Lendecke [Sun, 20 Nov 2005 10:18:33 +0000 (10:18 +0000)]
r11807: Make dcerpc_alter_context async. Same thing: Please look at this
closely. Survives RPC-ALTERCONTEXT.
Volker
Volker Lendecke [Sat, 19 Nov 2005 22:36:19 +0000 (22:36 +0000)]
r11806: Minor cleanup
Volker Lendecke [Sat, 19 Nov 2005 22:34:46 +0000 (22:34 +0000)]
r11805: dcerpc_bind_auth is only used in dcerpc_bind_auth_password
Volker Lendecke [Sat, 19 Nov 2005 22:31:26 +0000 (22:31 +0000)]
r11804: Asyncify the anonymous bind, convert the calls in xplogin.c.
Tridge et al, please take a close look at this. It survives my basic rpc-login
test as well as rpc-lsa, but this is critical I think.
Volker
Volker Lendecke [Sat, 19 Nov 2005 21:15:37 +0000 (21:15 +0000)]
r11803: New start for RPC-LOGIN load test, this time async. It's gonna be fun to see
hundreds of logins from a single process.... :-)
Does not do all that is necessary yet. But as the old one was #ifdef'ed out,
this does not hurt much I think.
Volker
Stefan Metzmacher [Sat, 19 Nov 2005 10:40:31 +0000 (10:40 +0000)]
r11802: - add copyright
- fix crash bug when running as non root
- add extra comments in the output
metze
Andrew Tridgell [Sat, 19 Nov 2005 06:39:12 +0000 (06:39 +0000)]
r11801: - added basic SMB2 find support
- added SMB2-SCANFIND test
- cleaned up continue flags in EAs and find requests
Andrew Tridgell [Sat, 19 Nov 2005 05:55:08 +0000 (05:55 +0000)]
r11800: - filled in unknown fields in SMB2 all_info level
- allow setting of the ALL_EAS flags bits in SMB2 getinfo
Jeremy Allison [Sat, 19 Nov 2005 01:12:10 +0000 (01:12 +0000)]
r11797: Added OpenSSH fix for "%.*s" format crash. From Darren Tucker
<dtucker@zip.com.au>
Jeremy.
Volker Lendecke [Fri, 18 Nov 2005 23:48:51 +0000 (23:48 +0000)]
r11796: Two more uses of !composite_is_ok
Andrew Tridgell [Fri, 18 Nov 2005 23:34:47 +0000 (23:34 +0000)]
r11795: used a couple more of volkers composite helper functions. They
certainly make the code more compact.
Andrew Tridgell [Fri, 18 Nov 2005 23:27:58 +0000 (23:27 +0000)]
r11794: - fixed a valgrind error in libnet, caused by using a stack variable
after the function has returned (the *address variable was assigned
into the state).
- changed libnet to use event_context_find() instead of
event_context_init(), so it works as a child of existing code that
uses a event context
Andrew Tridgell [Fri, 18 Nov 2005 23:15:32 +0000 (23:15 +0000)]
r11791: simplify the SMB2 connect code following some suggestions from volker
Stefan Metzmacher [Fri, 18 Nov 2005 14:13:49 +0000 (14:13 +0000)]
r11789: - add the start of a SMB2 server
- it does Negprot and SessionSetup yet
the rest returns NT_STATUS_NOT_IMPLEMENTED
- it's off by default, enable with:
smbsrv:enable smb2 = yes
- negotition in the SMB Negprot isn't supported yet
- it's only tested with smbtorture SMB2-CONNECT
not with vista as client
metze
Stefan Metzmacher [Fri, 18 Nov 2005 14:13:21 +0000 (14:13 +0000)]
r11788: remove unused header files
metze
Stefan Metzmacher [Fri, 18 Nov 2005 13:33:56 +0000 (13:33 +0000)]
r11787: add missing file
metze