kamenim/samba.git
13 years agos3: Remove smbd_server_conn from reply_tcon_and_X
Volker Lendecke [Sat, 12 Jun 2010 09:41:00 +0000 (11:41 +0200)]
s3: Remove smbd_server_conn from reply_tcon_and_X

13 years agos3: Remove smbd_server_conn from reply_tcon
Volker Lendecke [Sat, 12 Jun 2010 09:40:37 +0000 (11:40 +0200)]
s3: Remove smbd_server_conn from reply_tcon

13 years agos3: Remove smbd_server_conn from make_connection_snum
Volker Lendecke [Sat, 12 Jun 2010 09:39:03 +0000 (11:39 +0200)]
s3: Remove smbd_server_conn from make_connection_snum

13 years agos3: Remove smbd_server_conn from change_to_user
Volker Lendecke [Sat, 12 Jun 2010 09:34:38 +0000 (11:34 +0200)]
s3: Remove smbd_server_conn from change_to_user

13 years agos3: Remove smbd_server_conn from reply_ntcancel
Volker Lendecke [Sat, 12 Jun 2010 09:29:26 +0000 (11:29 +0200)]
s3: Remove smbd_server_conn from reply_ntcancel

13 years agos3: Remove smbd_server_conn from send_nt_replies
Volker Lendecke [Sat, 12 Jun 2010 09:29:12 +0000 (11:29 +0200)]
s3: Remove smbd_server_conn from send_nt_replies

13 years agos3: Remove smbd_server_conn from reply_smb2002
Volker Lendecke [Sat, 12 Jun 2010 09:26:51 +0000 (11:26 +0200)]
s3: Remove smbd_server_conn from reply_smb2002

13 years agos3: Explicitly pass sconn to negprot_spnego()
Volker Lendecke [Sat, 12 Jun 2010 09:25:41 +0000 (11:25 +0200)]
s3: Explicitly pass sconn to negprot_spnego()

13 years agos3: Remove some global references to smbd_server_conn
Volker Lendecke [Sat, 12 Jun 2010 09:19:21 +0000 (11:19 +0200)]
s3: Remove some global references to smbd_server_conn

13 years agos3: Remove sconn arg from smbd_notify_cancel_by_smbreq
Volker Lendecke [Sat, 12 Jun 2010 11:08:31 +0000 (13:08 +0200)]
s3: Remove sconn arg from smbd_notify_cancel_by_smbreq

13 years agos3: Initialize smb_request->sconn in smbd_smb2_fake_smb_request
Volker Lendecke [Sat, 12 Jun 2010 10:54:02 +0000 (12:54 +0200)]
s3: Initialize smb_request->sconn in smbd_smb2_fake_smb_request

13 years agos3: Use "sconn" argument in smbd_[un]lock_socket_internal
Volker Lendecke [Sat, 12 Jun 2010 09:06:24 +0000 (11:06 +0200)]
s3: Use "sconn" argument in smbd_[un]lock_socket_internal

Metze, please check!

13 years agos3: Pass sconn to init_smb_request()
Volker Lendecke [Sat, 12 Jun 2010 09:01:27 +0000 (11:01 +0200)]
s3: Pass sconn to init_smb_request()

13 years agos3: Add "smbd_server_connection" to smb_request
Volker Lendecke [Sat, 12 Jun 2010 09:01:10 +0000 (11:01 +0200)]
s3: Add "smbd_server_connection" to smb_request

13 years agos3: Fix some nonempty lines
Volker Lendecke [Sat, 12 Jun 2010 08:28:42 +0000 (10:28 +0200)]
s3: Fix some nonempty lines

14 years agos4:dsdb Allow calling dsdb_convert_object_ex() directly
Andrew Bartlett [Sat, 12 Jun 2010 01:03:36 +0000 (11:03 +1000)]
s4:dsdb Allow calling dsdb_convert_object_ex() directly

This will allow the libnet_vampire code to manually convert individual
schema objects.

Andrew Bartlett

14 years agos4:dsdb Add debug
Andrew Bartlett [Thu, 10 Jun 2010 11:34:48 +0000 (21:34 +1000)]
s4:dsdb Add debug

14 years agos4:dsdb Simplfy match of objectclass in dsdb_schema_set_el_from_ldb_msg
Andrew Bartlett [Thu, 10 Jun 2010 11:29:57 +0000 (21:29 +1000)]
s4:dsdb Simplfy match of objectclass in dsdb_schema_set_el_from_ldb_msg

There is no need to do a full ldb_match_msg() for a simple case
insensitive string.

Andrew Bartlett

14 years agos4:provision Allow both additional and override prefixmaps in Schema
Andrew Bartlett [Wed, 9 Jun 2010 22:58:44 +0000 (08:58 +1000)]
s4:provision Allow both additional and override prefixmaps in Schema

The idea here is to allow some callers to specify a new prefixMap that
will override the values loaded from the prefixMap.txt.

Andrew Bartlett

14 years agos4:dsdb Allow a binary prefix map to be specified in the LDIF
Andrew Bartlett [Wed, 9 Jun 2010 22:51:30 +0000 (08:51 +1000)]
s4:dsdb Allow a binary prefix map to be specified in the LDIF

This allows it to be specified in either binary or as a string.

Andrew Bartlett

14 years agos4:dsdb Provide a function to convert from DRS prefix maps to the LDB prefixmap
Andrew Bartlett [Wed, 9 Jun 2010 22:27:59 +0000 (08:27 +1000)]
s4:dsdb Provide a function to convert from DRS prefix maps to the LDB prefixmap

This allows us to push a prefixmap directly into the schema we
generate in the provision code.

Andrew Bartlett

14 years agos4:dsdb Add more debugs to help track down failures to parse the prefixmap
Andrew Bartlett [Wed, 9 Jun 2010 22:25:49 +0000 (08:25 +1000)]
s4:dsdb Add more debugs to help track down failures to parse the prefixmap

14 years agos4:dsdb Put back the reference and set_attributes in dsdb_reference_schema
Andrew Bartlett [Wed, 9 Jun 2010 10:21:19 +0000 (20:21 +1000)]
s4:dsdb Put back the reference and set_attributes in dsdb_reference_schema

I'm not sure why I removed these in fe3e1af901c970f738bee92baac5d7d4f5736e17

Andrew Bartlett

14 years agocheck if LD_AS_NEEDED breaks linking with libreadline fixes #7209
Olaf Flebbe [Tue, 25 May 2010 11:22:56 +0000 (13:22 +0200)]
check if LD_AS_NEEDED breaks linking with libreadline fixes #7209

Signed-off-by: Simo Sorce <idra@samba.org>
14 years agos4/test-libnet: Re-align code a bit - more readable this way IMO
Kamen Mazdrashki [Fri, 11 Jun 2010 13:02:10 +0000 (16:02 +0300)]
s4/test-libnet: Re-align code a bit - more readable this way IMO

14 years agos4/test-libnet: Refactor test_group_create() a little to:
Kamen Mazdrashki [Fri, 11 Jun 2010 12:44:34 +0000 (15:44 +0300)]
s4/test-libnet: Refactor test_group_create() a little to:

- Close unused policy handle for created group
- re-align code - it is more readable this way IMHO

14 years agos4/test-libnet: Refactor test_user_create() little bit to:
Kamen Mazdrashki [Fri, 11 Jun 2010 12:37:12 +0000 (15:37 +0300)]
s4/test-libnet: Refactor test_user_create() little bit to:

- fix a flaw that user RID is not returned in case
  user is recreated
- close unused policy handle to created user
- do not take into account the result from
  test_samr_close_handle - we don't case

14 years agos4/test-libnet: add few comments for utility functions
Kamen Mazdrashki [Fri, 11 Jun 2010 12:30:09 +0000 (15:30 +0300)]
s4/test-libnet: add few comments for utility functions

14 years agos4/test-libnet: Move LSA close handle function to a common location
Kamen Mazdrashki [Fri, 11 Jun 2010 12:11:01 +0000 (15:11 +0300)]
s4/test-libnet: Move LSA close handle function to a common location

14 years agos4/test-libnet: remove unused code
Kamen Mazdrashki [Fri, 11 Jun 2010 12:01:07 +0000 (15:01 +0300)]
s4/test-libnet: remove unused code

Pre-connecting SAMR and LSA pipes is implemented in
test_libnet_context_init() function

14 years agos4/test-libnet: Fix compile-time warning
Kamen Mazdrashki [Fri, 11 Jun 2010 11:59:57 +0000 (14:59 +0300)]
s4/test-libnet: Fix compile-time warning

14 years agoAdjusted ACL tests to use encripted connection.
Nadezhda Ivanova [Fri, 11 Jun 2010 14:22:21 +0000 (17:22 +0300)]
Adjusted ACL tests to use encripted connection.

This way we get rid of the conditional if.Samba checks, because users are
successfully created and enabled.

14 years agos3-auth: Fix valgrind warning (unitialized var) in samu_to_SamInfo3().
Günther Deschner [Fri, 11 Jun 2010 13:05:36 +0000 (15:05 +0200)]
s3-auth: Fix valgrind warning (unitialized var) in samu_to_SamInfo3().

Guenther

s3:auth do not fail if there are 0 group sids

Signed-off-by: Günther Deschner <gd@samba.org>
14 years agoconfigure.in prevent generation of file "shlib.so"
Olaf Flebbe [Fri, 11 Jun 2010 11:52:23 +0000 (13:52 +0200)]
configure.in prevent generation of file "shlib.so"

14 years agos3:auth fix samu->info3 conversion
Simo Sorce [Fri, 11 Jun 2010 12:25:27 +0000 (08:25 -0400)]
s3:auth fix samu->info3 conversion

Some pdb_get_ functions where missing because of previous mis-patching

14 years agos3: Fix starving the echo responder
Volker Lendecke [Tue, 8 Jun 2010 12:33:05 +0000 (14:33 +0200)]
s3: Fix starving the echo responder

When both the echo responder and the 445 socket want to send stuff to the
worker smbd, the select loop is not fair. It always chooses the smaller file
descriptor to work on. This can mean that on a busy system the echo responder
never gets around to feed its stuff to the parent.

This fix chooses the async echo responder socket when both the 445 and the echo
responder socket are readable.

Yes, it is a very hackish fix which is required *now* I think. The proper fix
would be to either assign priorities to fd's in tevent, or the from my point of
view better fix would be to make tevent kindof round-robin.

Round-robin would mean that whenever a fd has been dealt with, it is taken off
the list of interested sockets, and only if no other socket is active, all of
the ones waiting are put back. This is a bit like EPOLL_ONESHOT, which I would
like to use for this in the epoll case. Although, I need to do some research if
maybe epoll already guarantees round-robin, I did not find anything in the docs
yet.

Volker

14 years agos4:rootdse LDB module - use LDB result constants
Matthias Dieter Wallnöfer [Fri, 11 Jun 2010 08:19:19 +0000 (10:19 +0200)]
s4:rootdse LDB module - use LDB result constants

14 years agos4:passwords.py - set the "dSHeuristics"
Matthias Dieter Wallnöfer [Fri, 11 Jun 2010 08:04:50 +0000 (10:04 +0200)]
s4:passwords.py - set the "dSHeuristics"

As per Nadia's request and abartlet's suggestion the test now also sets the
"dSHeuristics" attribute properly to be able to perform the tests also against
Windows without further configuration.

The code has the neat feature that it undoes the change and resets the
behaviour as it was before.

14 years agos4:passwords.py - use "self" before "enable_account" to make really sure which LDB...
Matthias Dieter Wallnöfer [Fri, 11 Jun 2010 08:04:19 +0000 (10:04 +0200)]
s4:passwords.py - use "self" before "enable_account" to make really sure which LDB connection should be used

14 years agos4:ldap.py - check how the directory server behaves when the 'distinguishedName'...
Matthias Dieter Wallnöfer [Fri, 11 Jun 2010 07:19:02 +0000 (09:19 +0200)]
s4:ldap.py - check how the directory server behaves when the 'distinguishedName' attribute wasn't set correctly

(On add operations the server does completely ignore it and uses the "header"
DN)

14 years agos4:ldap.py - check if the "name" of an entry was really ignored
Matthias Dieter Wallnöfer [Fri, 11 Jun 2010 07:17:08 +0000 (09:17 +0200)]
s4:ldap.py - check if the "name" of an entry was really ignored

14 years agoldb:ldb_sqlite3 backend - remove a warning which has already been resolved
Matthias Dieter Wallnöfer [Fri, 11 Jun 2010 06:52:12 +0000 (08:52 +0200)]
ldb:ldb_sqlite3 backend - remove a warning which has already been resolved

14 years agoldb:ldb_sqlite3 backend - remove checks which are now done elsewhere
Matthias Dieter Wallnöfer [Fri, 11 Jun 2010 06:50:36 +0000 (08:50 +0200)]
ldb:ldb_sqlite3 backend - remove checks which are now done elsewhere

(In SAMBA LDB modules)

14 years agoAdd missing stub functions.
Jelmer Vernooij [Thu, 10 Jun 2010 23:47:45 +0000 (01:47 +0200)]
Add missing stub functions.

14 years agos4-python: Fix formatting.
Jelmer Vernooij [Thu, 10 Jun 2010 21:12:53 +0000 (23:12 +0200)]
s4-python: Fix formatting.

14 years agos3:rpc improve handles memory hierarchy and use better name.
Simo Sorce [Mon, 7 Jun 2010 21:38:01 +0000 (17:38 -0400)]
s3:rpc improve handles memory hierarchy and use better name.

While there also cleanup the code a bit.

Signed-off-by: Günther Deschner <gd@samba.org>
14 years agos4-smbtorture: rework internals of RPC-SPOOLSS-ACCESS testsuite.
Günther Deschner [Tue, 8 Jun 2010 10:26:15 +0000 (12:26 +0200)]
s4-smbtorture: rework internals of RPC-SPOOLSS-ACCESS testsuite.

Guenther

14 years agos4/test: use test_libnet_context_init() function to create libnet_context
Kamen Mazdrashki [Thu, 10 Jun 2010 15:23:06 +0000 (18:23 +0300)]
s4/test: use test_libnet_context_init() function to create libnet_context

This way we ensure that LIBNET call will be issued against
the DC we are testing against

14 years agos4/test: fix SEGFAULT in test_group_create when rid is NULL
Kamen Mazdrashki [Thu, 10 Jun 2010 15:21:44 +0000 (18:21 +0300)]
s4/test: fix SEGFAULT in test_group_create when rid is NULL

Caller of this function may not need group RID and
in such case it passes NULL for rid parameter

14 years agos4/test-libnet: Helper func to initialize libnet_context
Kamen Mazdrashki [Thu, 10 Jun 2010 10:12:08 +0000 (13:12 +0300)]
s4/test-libnet: Helper func to initialize libnet_context

and create rpc connections to SAMR and LSA pipes.
This function should be used in cases where we need
'deterministic' connection to the DC we are testing against.

The problem is that most of the test follow the pattern:
1. Connect to server supplied on command line
2. Create user/group through connection in 1.
3. Use 'fresh' libnet_context to query/alter the object in
   step 2.

In singe-DC environment this work well. But in multiple-DC
environment libnet may connect to another DC (step 3), not
the one we use in step 1 and 2. This leads to false error -
object created in step 2 is not found on DC we are querying
in step 3.

14 years agos4/test: Call test_user_cleanup with original RDN for the user
Kamen Mazdrashki [Wed, 9 Jun 2010 23:02:06 +0000 (02:02 +0300)]
s4/test: Call test_user_cleanup with original RDN for the user

'name' var is changed during the test execution and should
contain samAccountName for the user, while test_user_cleanup()
expects account RDN to be passed

14 years agos4/test: Fix incorrect NTSTATUS comparison
Kamen Mazdrashki [Wed, 9 Jun 2010 22:59:51 +0000 (01:59 +0300)]
s4/test: Fix incorrect NTSTATUS comparison

In case group already exists, we should get NT_STATUS_GROUP_EXISTS.

14 years agos4/test: remove create/delete group implementation in API-CREATEGROUP test
Kamen Mazdrashki [Tue, 8 Jun 2010 15:03:00 +0000 (18:03 +0300)]
s4/test: remove create/delete group implementation in API-CREATEGROUP test

We already have common implementation for create/delete
group in libnet/utils.c

14 years agos4/test: Remove duplicated test_opendomain() functionality
Kamen Mazdrashki [Tue, 8 Jun 2010 12:44:00 +0000 (15:44 +0300)]
s4/test: Remove duplicated test_opendomain() functionality

We already have a nice common test_domain_open()
function in libnet/util.c

14 years agos4/test: libnet_user tests should use common test_user_cleanup
Kamen Mazdrashki [Mon, 7 Jun 2010 15:44:14 +0000 (18:44 +0300)]
s4/test: libnet_user tests should use common test_user_cleanup

implementation instead of not-so-informative test_cleanup()

14 years agos4/test: Refactor test_user_cleanup to be more robust
Kamen Mazdrashki [Mon, 7 Jun 2010 15:33:35 +0000 (18:33 +0300)]
s4/test: Refactor test_user_cleanup to be more robust

We always call this function with user's account RDN
while samr_LookupNames() searches for samAccountName.

14 years agos4/test: remove duplicated comment
Kamen Mazdrashki [Mon, 7 Jun 2010 14:51:26 +0000 (17:51 +0300)]
s4/test: remove duplicated comment

14 years agos3:misc make use of server_[event/messaging]_context directly
Andreas Schneider [Thu, 10 Jun 2010 15:48:39 +0000 (11:48 -0400)]
s3:misc make use of server_[event/messaging]_context directly

Untangle these functions from smbd specific dependencies so they can be freely
used in multiple servers.

14 years agos3:winbindd use common server context functions
Simo Sorce [Thu, 10 Jun 2010 15:55:27 +0000 (11:55 -0400)]
s3:winbindd use common server context functions

14 years agos3:lib make server contexts generic
Simo Sorce [Thu, 10 Jun 2010 15:54:00 +0000 (11:54 -0400)]
s3:lib make server contexts generic

Pair-programmed-with: Andreas Schneider <asn@samba.org>

14 years agoFix the build in the non WITH_AIO case (sorry).
Jeremy Allison [Thu, 10 Jun 2010 21:28:55 +0000 (14:28 -0700)]
Fix the build in the non WITH_AIO case (sorry).

14 years agos3: readd h_errno struct member but rename it
Björn Jacke [Thu, 10 Jun 2010 21:15:19 +0000 (23:15 +0200)]
s3: readd h_errno struct member but rename it

as pointed out by metze this is a structure of fixed size, which should not be
changed.

14 years agoSince idtree assigns sequentially, it rarely reaches high numbers.
Rusty Russell [Thu, 10 Jun 2010 20:27:51 +0000 (13:27 -0700)]
Since idtree assigns sequentially, it rarely reaches high numbers.
But such numbers can be forced with idr_get_new_above(), and that
reveals two bugs:

1) Crash in sub_remove() caused by pa array being too short.
2) Shift by more than 32 in _idr_find(), which is undefined, causing
   the "outside the current tree" optimization to misfire and return NULL.

14 years agoImplement AIO in SMB2. Doesn't allow cancel calls yet (to be added).
Jeremy Allison [Thu, 10 Jun 2010 20:20:37 +0000 (13:20 -0700)]
Implement AIO in SMB2. Doesn't allow cancel calls yet (to be added).

Jeremy.

14 years agoEnsure we correctly setup req->subreq before testing for async. Stops crashes when
Jeremy Allison [Thu, 10 Jun 2010 20:19:25 +0000 (13:19 -0700)]
Ensure we correctly setup req->subreq before testing for async. Stops crashes when
XX_send functions set async to stop interim replies being sent.

Jeremy.

14 years agoDon't use the autofree context for the globals. This causes child smbd's forked
Jeremy Allison [Thu, 10 Jun 2010 20:17:35 +0000 (13:17 -0700)]
Don't use the autofree context for the globals. This causes child smbd's forked
by modules to crash due to destructors being called (found when using the vfs_aio_fork
module with smb2).

Jeremy.

14 years agos3:cleaunp reformatting for readability
Simo Sorce [Thu, 10 Jun 2010 18:56:03 +0000 (14:56 -0400)]
s3:cleaunp reformatting for readability

14 years agos3:cleanup remove trailing spaces
Simo Sorce [Thu, 10 Jun 2010 18:36:19 +0000 (14:36 -0400)]
s3:cleanup remove trailing spaces

14 years agos3: fix build on HP-UX
Björn Jacke [Thu, 10 Jun 2010 15:19:16 +0000 (17:19 +0200)]
s3: fix build on HP-UX

this struct member h_errno is not used in the HP-UX code paths, it was just
there because Solaris has it, too. As h_errno is a function call macro on HP-UX
when thread support is enabled we run into trouble here. Just commenting it out
should be okay as we don't use it anyway.

14 years agos4:ldap.py - add testcase which demonstrates the reset of the "primaryGroupID"
Matthias Dieter Wallnöfer [Thu, 10 Jun 2010 14:18:41 +0000 (16:18 +0200)]
s4:ldap.py - add testcase which demonstrates the reset of the "primaryGroupID"

14 years agos4:samldb LDB module - fix up the case when the old and new "primaryGroupID" are...
Matthias Dieter Wallnöfer [Thu, 10 Jun 2010 14:15:10 +0000 (16:15 +0200)]
s4:samldb LDB module - fix up the case when the old and new "primaryGroupID" are the same

14 years agos4:samldb LDB module - don't create multiple "ac" module contexts on modify operations
Matthias Dieter Wallnöfer [Thu, 10 Jun 2010 14:08:23 +0000 (16:08 +0200)]
s4:samldb LDB module - don't create multiple "ac" module contexts on modify operations

Since we do now run sequentially through all checks we don't need multiple "ac"
contexts anymore.

14 years agos4:dcesrv_samr_Add/DeleteAliasMember - provide better NTSTATUS return codes when...
Matthias Dieter Wallnöfer [Thu, 10 Jun 2010 13:47:48 +0000 (15:47 +0200)]
s4:dcesrv_samr_Add/DeleteAliasMember - provide better NTSTATUS return codes when something didn't work

14 years agos4:dcesrv_samr_GetAliasMembership - fix type of counter variables
Matthias Dieter Wallnöfer [Thu, 10 Jun 2010 10:42:04 +0000 (12:42 +0200)]
s4:dcesrv_samr_GetAliasMembership - fix type of counter variables

14 years agos4:dcesrv_samr_DeleteAliasMember - add more braces to fit better the coding styles
Matthias Dieter Wallnöfer [Thu, 10 Jun 2010 10:19:04 +0000 (12:19 +0200)]
s4:dcesrv_samr_DeleteAliasMember - add more braces to fit better the coding styles

14 years agos4:dcesrv_samr_AddAliasMembership - Merge the two error blocks into one
Matthias Dieter Wallnöfer [Thu, 10 Jun 2010 10:17:48 +0000 (12:17 +0200)]
s4:dcesrv_samr_AddAliasMembership - Merge the two error blocks into one

14 years agos4:dcesrv_samr_Add/DelGroupMember - remove the account type check
Matthias Dieter Wallnöfer [Thu, 10 Jun 2010 09:50:12 +0000 (11:50 +0200)]
s4:dcesrv_samr_Add/DelGroupMember - remove the account type check

MS-SAMR 3.1.5.8 speaks from accounts which are not necessarely only users.

14 years agos4:dcesrv_samr_AddGroupMember - also the error code "LDB_ERR_ATTRIBUTE_OR_VALUE_EXIST...
Matthias Dieter Wallnöfer [Thu, 10 Jun 2010 09:48:33 +0000 (11:48 +0200)]
s4:dcesrv_samr_AddGroupMember - also the error code "LDB_ERR_ATTRIBUTE_OR_VALUE_EXISTS" is allowed

This is returned when the group is the primary group of the specified entry.

14 years agos4:samba_dsdb LDB module - move the "objectclass_attrs" module back
Matthias Dieter Wallnöfer [Thu, 10 Jun 2010 09:05:43 +0000 (11:05 +0200)]
s4:samba_dsdb LDB module - move the "objectclass_attrs" module back

I think it should be lower in order to control also the "instanceType" module.

14 years agos4:ldap.py - add a test to demonstrate the 'instanceType' behaviour
Matthias Dieter Wallnöfer [Thu, 10 Jun 2010 08:52:31 +0000 (10:52 +0200)]
s4:ldap.py - add a test to demonstrate the 'instanceType' behaviour

14 years agos4:instancetype LDB module - prevent all types of "instanceType" manipulation
Matthias Dieter Wallnöfer [Thu, 10 Jun 2010 08:39:52 +0000 (10:39 +0200)]
s4:instancetype LDB module - prevent all types of "instanceType" manipulation

Also on Windows Server you aren't able to change it.

14 years agos3: Fix EnumDomainAliases when no aliases are in LDAP
Volker Lendecke [Thu, 10 Jun 2010 13:40:14 +0000 (15:40 +0200)]
s3: Fix EnumDomainAliases when no aliases are in LDAP

We used to return NT_STATUS_ACCESS_DENIED, now we just return 0 entries, just
like W2k8 does.

usrmgr.exe was pretty unhappy with the NT_STATUS_ACCESS_DENIED

14 years agos3:pdb_ldap: fix bug 7505 - init_sam_from_ldap stores group in sid2uid cache
Michael Adam [Thu, 10 Jun 2010 09:56:15 +0000 (11:56 +0200)]
s3:pdb_ldap: fix bug 7505 - init_sam_from_ldap stores group in sid2uid cache

14 years agos3: Add a dummy test to prove the maxfd calculation
Volker Lendecke [Thu, 10 Jun 2010 07:50:32 +0000 (09:50 +0200)]
s3: Add a dummy test to prove the maxfd calculation

To be run in a debugger, there's no way to inspect the internal tevent_select.c
maxfd calculation

14 years agotevent: Fix maxfd calculation in tevent_select
Volker Lendecke [Thu, 10 Jun 2010 07:41:11 +0000 (09:41 +0200)]
tevent: Fix maxfd calculation in tevent_select

When doing

        fd1 = tevent_add_fd(ev, ev, 2, 0, NULL, NULL);
        fd2 = tevent_add_fd(ev, ev, 3, 0, NULL, NULL);
        TALLOC_FREE(fd2);
        fd2 = tevent_add_fd(ev, ev, 1, 0, NULL, NULL);

we end up with select_ev->maxfd==1. This is wrong.

An alternative fix might be to make select_ev->maxfd an unsigned int and make
EVENT_INVALID_MAXFD==UINT_MAX. But in theory we might end up with an fd of
UINT_MAX.

std_event_add_fd() contains exactly the same piece of code, so I'm directly
pushing it.

Volker

14 years agos3-smbd: Fix typo in comment.
Karolin Seeger [Thu, 10 Jun 2010 08:35:56 +0000 (10:35 +0200)]
s3-smbd: Fix typo in comment.

Karolin

14 years agos3-passdb: Fix typo in comment.
Karolin Seeger [Thu, 10 Jun 2010 08:34:58 +0000 (10:34 +0200)]
s3-passdb: Fix typo in comment.

Karolin

14 years agoRename "allow_smb2" -> "using_smb2" and make the usage clearer.
Jeremy Allison [Thu, 10 Jun 2010 02:12:02 +0000 (19:12 -0700)]
Rename "allow_smb2" -> "using_smb2" and make the usage clearer.

14 years agoSplit out the "finished write processing" code into a function so it can be called
Jeremy Allison [Thu, 10 Jun 2010 00:09:11 +0000 (17:09 -0700)]
Split out the "finished write processing" code into a function so it can be called
by both sync and async code.

Jeremy.

14 years agoEnsure we don't send SMB1 keepalives on an SMB2 connection.
Jeremy Allison [Thu, 10 Jun 2010 00:08:41 +0000 (17:08 -0700)]
Ensure we don't send SMB1 keepalives on an SMB2 connection.

Jeremy.

14 years agoSplit out the "finished read processing" code into a function so it can be called
Jeremy Allison [Wed, 9 Jun 2010 17:48:12 +0000 (10:48 -0700)]
Split out the "finished read processing" code into a function so it can be called
by both sync and async code.

Jeremy.

14 years agos3-winbind: Fixed setting default sequence number.
Andreas Schneider [Wed, 9 Jun 2010 14:15:28 +0000 (16:15 +0200)]
s3-winbind: Fixed setting default sequence number.

14 years agos3: fix calculation of st_blocks in streams_xattr
Björn Jacke [Wed, 9 Jun 2010 13:24:26 +0000 (15:24 +0200)]
s3: fix calculation of st_blocks in streams_xattr

Thanks to Joachim Schmitz for finding that miscalculation.

14 years agos3: fix calculation of st_blocks in init_stat_ex_from_stat
Björn Jacke [Wed, 9 Jun 2010 13:21:24 +0000 (15:21 +0200)]
s3: fix calculation of st_blocks in init_stat_ex_from_stat

this is just relevant for the very rare systems that don't have the st_blocks
struct member. This is a fixup of the fix for bug 7474.

Thanks to Joachim Schmitz for spotting this!

14 years agoAdded guids for the validated writes.
Nadezhda Ivanova [Wed, 9 Jun 2010 09:03:32 +0000 (12:03 +0300)]
Added guids for the validated writes.

14 years agos3: Fix bug 7253
Volker Lendecke [Tue, 16 Mar 2010 20:03:34 +0000 (21:03 +0100)]
s3: Fix bug 7253

acct_ctrl is 32 bit in LOGIN_CACHE, but "w" as a format specifier for
tdb_unpack only writes 16 bits. Okay on x86, not okay on Solaris.

Thanks to Vladimir.Marek@Sun.COM!

Volker

14 years agoFix a valgrind error found by SMB2-COMPOUND test.
Jeremy Allison [Wed, 9 Jun 2010 04:20:07 +0000 (21:20 -0700)]
Fix a valgrind error found by SMB2-COMPOUND test.

If a file is closed we must also NULL out all chained_fsp
pointers when the fsp is freed to prevent invalid pointer
access.

Jeremy.

14 years agoSecond part of fix for 7501 - SMB2: CREATE request replies getting mangled.
Jeremy Allison [Wed, 9 Jun 2010 00:44:05 +0000 (17:44 -0700)]
Second part of fix for 7501 - SMB2: CREATE request replies getting mangled.

Based on code from Ira Cooper <samba@ira.wakeful.net>, and also
advice on refactoring the patch into a function call. outbuf vectors
can be reallocated by smb2 processing code, so when returning interim
responses we must not make assumptions about vector size.

Jeremy

14 years agoFix for bug 7501 - SMB2: CREATE request replies getting mangled.
Ira Cooper [Tue, 8 Jun 2010 18:20:23 +0000 (11:20 -0700)]
Fix for bug 7501 -  SMB2: CREATE request replies getting mangled.

Code for dup_smb2_req() was duplicating the wrong vector (i, instead
of i+2) when returning a non-minimal SMB2 response.

14 years agos3-selftest: disable domain creds smbclient_s3 tests for now.
Günther Deschner [Tue, 8 Jun 2010 12:34:20 +0000 (14:34 +0200)]
s3-selftest: disable domain creds smbclient_s3 tests for now.

Not sure they were ever working before.

With this patch, make selftest is down to 4 failures from 259 tests.

(two times LOCK9 from s3 smbtorture, one spoolss notify test and
samba3.posix_s3.unix.info2).

Guenther