ira/wip.git
10 years agoMerge branch 'master' of /home/tridge/samba/git/combined
Andrew Tridgell [Thu, 17 Sep 2009 22:24:20 +0000 (15:24 -0700)]
Merge branch 'master' of /home/tridge/samba/git/combined

10 years agoidl: regenerate IDL for NDR64 changes
Andrew Tridgell [Thu, 17 Sep 2009 22:16:21 +0000 (15:16 -0700)]
idl: regenerate IDL for NDR64 changes

10 years agos4-sam: add a note about the solaris client
Andrew Tridgell [Thu, 17 Sep 2009 22:15:58 +0000 (15:15 -0700)]
s4-sam: add a note about the solaris client

10 years agondrdump: fixed help
Andrew Tridgell [Thu, 17 Sep 2009 22:15:36 +0000 (15:15 -0700)]
ndrdump: fixed help

10 years agondr: num_auths is an array size, thus a uint3264
Andrew Tridgell [Thu, 17 Sep 2009 22:15:24 +0000 (15:15 -0700)]
ndr: num_auths is an array size, thus a uint3264

10 years agondr: split out ndr enum functions
Andrew Tridgell [Thu, 17 Sep 2009 22:14:31 +0000 (15:14 -0700)]
ndr: split out ndr enum functions

This allows for easier implementation of the NDR32/NDR64 split

10 years agoidl: recompile our IDL
Andrew Tridgell [Thu, 17 Sep 2009 16:59:50 +0000 (09:59 -0700)]
idl: recompile our IDL

The NDR64 change affects every IDL file

10 years agos4-pidl: add support for NDR64
Andrew Tridgell [Thu, 17 Sep 2009 16:10:21 +0000 (09:10 -0700)]
s4-pidl: add support for NDR64

Added support for NDR64 to the samba4 pidl generator

10 years agos4-rpc: added NDR64 support
Andrew Tridgell [Thu, 17 Sep 2009 16:09:48 +0000 (09:09 -0700)]
s4-rpc: added NDR64 support

This adds support for the nd464 binding string option

10 years agondr: added support for NDR64
Andrew Tridgell [Thu, 17 Sep 2009 16:08:47 +0000 (09:08 -0700)]
ndr: added support for NDR64

This adds NDR64 support for the push functions in libndr

10 years agoutil: use likely/unlikely for NT_STATUS_* macros
Andrew Tridgell [Thu, 17 Sep 2009 16:07:44 +0000 (09:07 -0700)]
util: use likely/unlikely for NT_STATUS_* macros

10 years agolibreplace: added likely()/unlikely() macros for gcc
Andrew Tridgell [Thu, 17 Sep 2009 16:07:17 +0000 (09:07 -0700)]
libreplace: added likely()/unlikely() macros for gcc

These macros allow the compile to better optimise code that has a lot
of if statements. I particularly want to use this for our low level
generated NDR code.

10 years agos3: build with pam support when possible
Björn Jacke [Thu, 17 Sep 2009 19:04:52 +0000 (21:04 +0200)]
s3: build with pam support when possible

10 years agoRe-generated idl files with () for enum values.
Jeremy Allison [Thu, 17 Sep 2009 18:28:37 +0000 (11:28 -0700)]
Re-generated idl files with () for enum values.
Jeremy.

10 years agoEnsure we enclose the enum values in brackets to make the cast work.
Jeremy Allison [Thu, 17 Sep 2009 18:27:51 +0000 (11:27 -0700)]
Ensure we enclose the enum values in brackets to make the cast work.
Jeremy.

10 years agoCheck in modified generated files for systems where pidl can't run.
Jeremy Allison [Thu, 17 Sep 2009 18:16:54 +0000 (11:16 -0700)]
Check in modified generated files for systems where pidl can't run.
Jeremy.

10 years agoMerge branch 'master' of ssh://jra@git.samba.org/data/git/samba
Jeremy Allison [Thu, 17 Sep 2009 18:11:23 +0000 (11:11 -0700)]
Merge branch 'master' of ssh://jra@git.samba.org/data/git/samba

10 years agoFix the problem with pidl generating invalid C for enums. According
Jeremy Allison [Thu, 17 Sep 2009 18:08:42 +0000 (11:08 -0700)]
Fix the problem with pidl generating invalid C for enums. According
to the C standard an enum is guarenteed to be an (int), which means
for 4 byte ints specifying a type of 0x80000000 is an invalid value.
The Solaris compiler complains about this. Fix by adding an (int)
cast in front of the value generation.
Jeremy.

10 years agospnego: Support ASN.1 BIT STRING and use it in SPNEGO.
Kouhei Sutou [Thu, 13 Aug 2009 06:12:01 +0000 (15:12 +0900)]
spnego: Support ASN.1 BIT STRING and use it in SPNEGO.

Signed-off-by: Günther Deschner <gd@samba.org>
10 years agos4:descriptor module - Revert and const fixups
Matthias Dieter Wallnöfer [Thu, 17 Sep 2009 16:37:46 +0000 (18:37 +0200)]
s4:descriptor module - Revert and const fixups

- Revert a change introduced by me since I didn't understood the meaning of the
  version check
- Added some "const" to suppress compiler warnings

10 years agos4:descriptor - cosmetic
Matthias Dieter Wallnöfer [Thu, 17 Sep 2009 16:17:55 +0000 (18:17 +0200)]
s4:descriptor - cosmetic

10 years agos4:libnet_become_dc - Fix some uninitialised variables
Matthias Dieter Wallnöfer [Thu, 17 Sep 2009 16:05:58 +0000 (18:05 +0200)]
s4:libnet_become_dc - Fix some uninitialised variables

10 years agos4:provision - Some rework
Matthias Dieter Wallnöfer [Thu, 17 Sep 2009 14:00:55 +0000 (16:00 +0200)]
s4:provision - Some rework

- Add/change "wellKnownObjects" attributes
- Order entries in "provision_basedn_modify.ldif"
- Add/change "delete entries" object under BASEDN and CONFIGDN
- Fix default version number of "Default domain policy" group policy
- Add "domain updates" objects for interoperability with MS AD maintaining tools
- Show version number in the "oEMInformation" attribute (suggested by ekacnet)
- Smaller fixups

10 years agovoid functions can't return a value. Found by the Solaris compiler.
Jeremy Allison [Thu, 17 Sep 2009 17:04:19 +0000 (10:04 -0700)]
void functions can't return a value. Found by the Solaris compiler.
Jeremy.

10 years agovoid functions can't return a value. Found by the Solaris compiler.
Jeremy Allison [Thu, 17 Sep 2009 17:03:14 +0000 (10:03 -0700)]
void functions can't return a value. Found by the Solaris compiler.
Jeremy.

10 years agovoid functions can't return a value. Found by the Solaris compiler.
Jeremy Allison [Thu, 17 Sep 2009 17:00:31 +0000 (10:00 -0700)]
void functions can't return a value. Found by the Solaris compiler.
Jeremy.

10 years agoWe now pass the Microsoft SMB2 fileio test with EA's and streams...
Jeremy Allison [Thu, 17 Sep 2009 16:29:07 +0000 (09:29 -0700)]
We now pass the Microsoft SMB2 fileio test with EA's and streams...
Jeremy.

10 years agoRemove ununsed variable warning.
Jeremy Allison [Thu, 17 Sep 2009 16:25:25 +0000 (09:25 -0700)]
Remove ununsed variable warning.
Jeremy.

10 years agos4/domain behaviour flags: Fix them up in various locations
Matthias Dieter Wallnöfer [Sun, 13 Sep 2009 10:37:41 +0000 (12:37 +0200)]
s4/domain behaviour flags: Fix them up in various locations

Additional notes:
- Bump the level to Windows Server 2008 R2 (we should support always the latest
  version - if we provision ourself)
- In "descriptor.c" the check for the "domainFunctionality" level shouldn't be
  needed: ACL owner groups (not owner user) are supported since Windows 2000
  Server (first AD edition)
  - I took the argument from: http://support.microsoft.com/kb/329194

10 years agos4/python: flags
Matthias Dieter Wallnöfer [Wed, 9 Sep 2009 19:24:34 +0000 (21:24 +0200)]
s4/python: flags

- Introduce the "userAccountControl", "groupType" and "sAMAccountType" flags
- Corrects the "domain/forestFunctionality" and "domainControllerFunctionality" flags

10 years agolibds/common/flags: various
Matthias Dieter Wallnöfer [Sun, 13 Sep 2009 09:01:44 +0000 (11:01 +0200)]
libds/common/flags: various

- Reorders the header file to have the order "userAccountFlags", "groupType",
  "sAMAccountType" (matches the order in the flag_mapping.c and samldb module)
- Fixes the group account flags properly up
- Fixes the flags for "domain/forestFunctionality" and "domainControllerFunctionality"
  up

10 years agoMerge branch 'master' of /home/tridge/samba/git/combined
Andrew Tridgell [Wed, 16 Sep 2009 10:58:21 +0000 (03:58 -0700)]
Merge branch 'master' of /home/tridge/samba/git/combined

10 years agos4-repl: raise a debug level
Andrew Tridgell [Wed, 16 Sep 2009 10:57:56 +0000 (03:57 -0700)]
s4-repl: raise a debug level

10 years agos4-dsdb: treat uSNHighest as 0 if @REPLCHANGED doesn't exist
Andrew Tridgell [Wed, 16 Sep 2009 10:43:37 +0000 (03:43 -0700)]
s4-dsdb: treat uSNHighest as 0 if @REPLCHANGED doesn't exist

When a partition is first created it still needs a uSNHighest value

10 years agoMerge branch 'master' of /home/tridge/samba/git/combined
Andrew Tridgell [Wed, 16 Sep 2009 10:58:13 +0000 (03:58 -0700)]
Merge branch 'master' of /home/tridge/samba/git/combined

10 years agos4-repl: raise a debug level
Andrew Tridgell [Wed, 16 Sep 2009 10:57:56 +0000 (03:57 -0700)]
s4-repl: raise a debug level

10 years agoMerge branch 'master' of ssh://git.samba.org/data/git/samba
Andrew Tridgell [Wed, 16 Sep 2009 10:57:09 +0000 (03:57 -0700)]
Merge branch 'master' of ssh://git.samba.org/data/git/samba

10 years agoMerge branch 'master' of /home/tridge/samba/git/combined
Andrew Tridgell [Wed, 16 Sep 2009 10:56:07 +0000 (03:56 -0700)]
Merge branch 'master' of /home/tridge/samba/git/combined

10 years agos4-dsdb: treat uSNHighest as 0 if @REPLCHANGED doesn't exist
Andrew Tridgell [Wed, 16 Sep 2009 10:43:37 +0000 (03:43 -0700)]
s4-dsdb: treat uSNHighest as 0 if @REPLCHANGED doesn't exist

When a partition is first created it still needs a uSNHighest value

10 years agolibcli/auth: rewrite schannel sign/seal code to be more generic
Stefan Metzmacher [Wed, 16 Sep 2009 00:03:46 +0000 (02:03 +0200)]
libcli/auth: rewrite schannel sign/seal code to be more generic

This prepares support for HMAC-SHA256/AES.

metze

10 years agolib/crypto: include aes.h into crypto.h
Stefan Metzmacher [Wed, 16 Sep 2009 00:36:49 +0000 (02:36 +0200)]
lib/crypto: include aes.h into crypto.h

metze

10 years agoMerge branch 'master' of /home/tridge/samba/git/combined
Andrew Tridgell [Wed, 16 Sep 2009 03:51:30 +0000 (20:51 -0700)]
Merge branch 'master' of /home/tridge/samba/git/combined

10 years agos4-repl: take advantage of async RPC forwarding
Andrew Tridgell [Wed, 16 Sep 2009 03:51:10 +0000 (20:51 -0700)]
s4-repl: take advantage of async RPC forwarding

This uses async RPC forwarding for the DsReplicaSync call

10 years agos4-rpc: added a module for forwarding RPC requests
Andrew Tridgell [Wed, 16 Sep 2009 03:50:30 +0000 (20:50 -0700)]
s4-rpc: added a module for forwarding RPC requests

dcesrv_irpc_forward_rpc_call() can be used to forward an arbitrary RPC
request to another task in Samba4, with the return being handled
asynchronously.

This is useful for forwarding DRS requests to the repl or kcc tasks

10 years agoMerge branch 'master' of /home/tridge/samba/git/combined
Andrew Tridgell [Wed, 16 Sep 2009 02:53:05 +0000 (19:53 -0700)]
Merge branch 'master' of /home/tridge/samba/git/combined

10 years agos4-drs: lock down key DRS calls
Andrew Tridgell [Wed, 16 Sep 2009 02:26:33 +0000 (19:26 -0700)]
s4-drs: lock down key DRS calls

The key DRS calls should only be allowed by administrators or domain
controllers

10 years agos4-security: added a new security level SECURITY_DOMAIN_CONTROLLER
Andrew Tridgell [Wed, 16 Sep 2009 02:25:45 +0000 (19:25 -0700)]
s4-security: added a new security level SECURITY_DOMAIN_CONTROLLER

This will be used as a simple way to lock down DRS replication to
administrators and domain controllers

10 years agoMerge branch 'master' of /home/tridge/samba/git/combined
Andrew Tridgell [Wed, 16 Sep 2009 01:46:18 +0000 (18:46 -0700)]
Merge branch 'master' of /home/tridge/samba/git/combined

10 years agos4-ldb: ldap attribute names can contain a '.'
Andrew Tridgell [Tue, 15 Sep 2009 21:07:43 +0000 (14:07 -0700)]
s4-ldb: ldap attribute names can contain a '.'

When they are of the form of OIDs

10 years agos4-ldb: expose ldb_transaction_prepare_commit() in ldb
Andrew Tridgell [Tue, 15 Sep 2009 21:07:06 +0000 (14:07 -0700)]
s4-ldb: expose ldb_transaction_prepare_commit() in ldb

It is useful to be able to control the 2 phase commit from application
code (s4 replication uses it)

10 years agos4-repl: don't do double replication
Andrew Tridgell [Tue, 15 Sep 2009 21:06:07 +0000 (14:06 -0700)]
s4-repl: don't do double replication

When we replicate from a remote DC, we need to note the new uSN that
the local changes have resulted in, and modify the uSN that the notify
task uses to determine if it should send a ReplicaSync message back to
the remote DC. Otherwise we end up always triggering a ReplicaSync
every time we replicate from another DC

10 years agotdb: allow reads after prepare commit
Andrew Tridgell [Tue, 15 Sep 2009 21:04:22 +0000 (14:04 -0700)]
tdb: allow reads after prepare commit

We previously only allowed a commit to happen after a prepare
commit. It is in fact safe to allow reads between a prepare and a
commit, and the s4 replication code can make use of that, so allow it.

10 years agos4-drs: filter based on local_usn
Andrew Tridgell [Tue, 15 Sep 2009 18:47:42 +0000 (11:47 -0700)]
s4-drs: filter based on local_usn

The getncchanges uSN is in our local space, so we must compare it to
the local_usn in replPropertyMetaData

10 years agos4-repl: make sure we marshal the replPropertyMetaData after the last change
Andrew Tridgell [Tue, 15 Sep 2009 18:46:59 +0000 (11:46 -0700)]
s4-repl: make sure we marshal the replPropertyMetaData after the last change

we were setting local_usn after the marshall, so it wasn't going into
the object

10 years agos4-dsdb: use DLIST_ADD() not DLIST_ADD_END()
Andrew Tridgell [Tue, 15 Sep 2009 17:01:26 +0000 (10:01 -0700)]
s4-dsdb: use DLIST_ADD() not DLIST_ADD_END()

Using DLIST_ADD_END() to construct a long list is very inefficient (it
is O(n^2). These lists are not ordered, so using DLIST_ADD() is much
better.

10 years agos4-ldb: cope better with corruption of tdb records
Andrew Tridgell [Tue, 15 Sep 2009 17:00:24 +0000 (10:00 -0700)]
s4-ldb: cope better with corruption of tdb records

When doing an indexed search if we hit a corrupt record we abandoned
the indexed search and did a full search. The problem was that we
might have sent some records to the caller already, which means the
caller ended up with duplicate records. Fix this by returning a search
error if indexing returns an error and we have given any records to
the caller.

10 years agotalloc: when we enable NULL tracking, reparent the autofree context
Andrew Tridgell [Tue, 15 Sep 2009 16:43:27 +0000 (09:43 -0700)]
talloc: when we enable NULL tracking, reparent the autofree context

If NULL tracking is enabled after the autofree context is initialised
then autofree ends up separate from the null_context. This means that
talloc_report_full() doesn't report the autofree context. Fix this by
reparenting the autofree context when we create the null_context.

10 years agos4-repl: add a debug to make it easier to monitor replication
Andrew Tridgell [Tue, 15 Sep 2009 16:23:14 +0000 (09:23 -0700)]
s4-repl: add a debug to make it easier to monitor replication

10 years agos3: Fix reading beyond the end of a named stream in xattr_streams
Volker Lendecke [Wed, 16 Sep 2009 01:20:49 +0000 (03:20 +0200)]
s3: Fix reading beyond the end of a named stream in xattr_streams

This was found thanks to a test by Sivani from Microsoft against Samba at the
SDC plugfest

10 years agos3: Add some debugs to streams_xattr
Volker Lendecke [Wed, 16 Sep 2009 01:15:53 +0000 (03:15 +0200)]
s3: Add some debugs to streams_xattr

10 years agoschannel: remove last traces of gensec.
Günther Deschner [Wed, 16 Sep 2009 01:23:05 +0000 (03:23 +0200)]
schannel: remove last traces of gensec.

Guenther

10 years agolib/crypto: link in AES crypto for s4 as well.
Günther Deschner [Wed, 16 Sep 2009 00:09:06 +0000 (02:09 +0200)]
lib/crypto: link in AES crypto for s4 as well.

Guenther

10 years agos3-schannel: remove unused schannel_decode/schannel_encode.
Günther Deschner [Tue, 15 Sep 2009 22:52:33 +0000 (00:52 +0200)]
s3-schannel: remove unused schannel_decode/schannel_encode.

Guenther

10 years agoschannel: fully share schannel sign/seal between s3 and 4.
Günther Deschner [Tue, 15 Sep 2009 22:26:17 +0000 (00:26 +0200)]
schannel: fully share schannel sign/seal between s3 and 4.

Guenther

10 years agoschannel: move schannel_sign to main directory.
Günther Deschner [Tue, 15 Sep 2009 16:29:10 +0000 (18:29 +0200)]
schannel: move schannel_sign to main directory.

Guenther

10 years agos4-schannel: try to fix the build.
Günther Deschner [Tue, 15 Sep 2009 23:07:26 +0000 (01:07 +0200)]
s4-schannel: try to fix the build.

Guenther

10 years agos4-schannel: first step of decoupling schannel from gensec.
Günther Deschner [Sun, 13 Sep 2009 16:42:45 +0000 (18:42 +0200)]
s4-schannel: first step of decoupling schannel from gensec.

Guenther

10 years agos4-schannel: strip trailing whitespace.
Günther Deschner [Sun, 13 Sep 2009 13:21:20 +0000 (15:21 +0200)]
s4-schannel: strip trailing whitespace.

Guenther

10 years agos3-schannel: fix blob length when pulling off a NL_AUTH_SIGNATURE in
Günther Deschner [Tue, 15 Sep 2009 20:13:12 +0000 (22:13 +0200)]
s3-schannel: fix blob length when pulling off a NL_AUTH_SIGNATURE in
cli_pipe_verify_schannel().

Guenther

10 years agolib/crypto: add aes encryption routines to main cryto lib.
Günther Deschner [Tue, 15 Sep 2009 21:52:20 +0000 (23:52 +0200)]
lib/crypto: add aes encryption routines to main cryto lib.

Guenther

10 years agoMerge branch 'master' of /home/tridge/samba/git/combined
Andrew Tridgell [Tue, 15 Sep 2009 21:53:02 +0000 (14:53 -0700)]
Merge branch 'master' of /home/tridge/samba/git/combined

10 years agos4-ldb: ldap attribute names can contain a '.'
Andrew Tridgell [Tue, 15 Sep 2009 21:07:43 +0000 (14:07 -0700)]
s4-ldb: ldap attribute names can contain a '.'

When they are of the form of OIDs

10 years agos4-ldb: expose ldb_transaction_prepare_commit() in ldb
Andrew Tridgell [Tue, 15 Sep 2009 21:07:06 +0000 (14:07 -0700)]
s4-ldb: expose ldb_transaction_prepare_commit() in ldb

It is useful to be able to control the 2 phase commit from application
code (s4 replication uses it)

10 years agos4-repl: don't do double replication
Andrew Tridgell [Tue, 15 Sep 2009 21:06:07 +0000 (14:06 -0700)]
s4-repl: don't do double replication

When we replicate from a remote DC, we need to note the new uSN that
the local changes have resulted in, and modify the uSN that the notify
task uses to determine if it should send a ReplicaSync message back to
the remote DC. Otherwise we end up always triggering a ReplicaSync
every time we replicate from another DC

10 years agotdb: allow reads after prepare commit
Andrew Tridgell [Tue, 15 Sep 2009 21:04:22 +0000 (14:04 -0700)]
tdb: allow reads after prepare commit

We previously only allowed a commit to happen after a prepare
commit. It is in fact safe to allow reads between a prepare and a
commit, and the s4 replication code can make use of that, so allow it.

10 years agos4-drs: filter based on local_usn
Andrew Tridgell [Tue, 15 Sep 2009 18:47:42 +0000 (11:47 -0700)]
s4-drs: filter based on local_usn

The getncchanges uSN is in our local space, so we must compare it to
the local_usn in replPropertyMetaData

10 years agos4-repl: make sure we marshal the replPropertyMetaData after the last change
Andrew Tridgell [Tue, 15 Sep 2009 18:46:59 +0000 (11:46 -0700)]
s4-repl: make sure we marshal the replPropertyMetaData after the last change

we were setting local_usn after the marshall, so it wasn't going into
the object

10 years agos4-dsdb: use DLIST_ADD() not DLIST_ADD_END()
Andrew Tridgell [Tue, 15 Sep 2009 17:01:26 +0000 (10:01 -0700)]
s4-dsdb: use DLIST_ADD() not DLIST_ADD_END()

Using DLIST_ADD_END() to construct a long list is very inefficient (it
is O(n^2). These lists are not ordered, so using DLIST_ADD() is much
better.

10 years agos4-ldb: cope better with corruption of tdb records
Andrew Tridgell [Tue, 15 Sep 2009 17:00:24 +0000 (10:00 -0700)]
s4-ldb: cope better with corruption of tdb records

When doing an indexed search if we hit a corrupt record we abandoned
the indexed search and did a full search. The problem was that we
might have sent some records to the caller already, which means the
caller ended up with duplicate records. Fix this by returning a search
error if indexing returns an error and we have given any records to
the caller.

10 years agotalloc: when we enable NULL tracking, reparent the autofree context
Andrew Tridgell [Tue, 15 Sep 2009 16:43:27 +0000 (09:43 -0700)]
talloc: when we enable NULL tracking, reparent the autofree context

If NULL tracking is enabled after the autofree context is initialised
then autofree ends up separate from the null_context. This means that
talloc_report_full() doesn't report the autofree context. Fix this by
reparenting the autofree context when we create the null_context.

10 years agos4-repl: add a debug to make it easier to monitor replication
Andrew Tridgell [Tue, 15 Sep 2009 16:23:14 +0000 (09:23 -0700)]
s4-repl: add a debug to make it easier to monitor replication

10 years agoMerge branch 'master' of /home/tridge/samba/git/combined
Andrew Tridgell [Tue, 15 Sep 2009 21:42:46 +0000 (14:42 -0700)]
Merge branch 'master' of /home/tridge/samba/git/combined

10 years agos4-libnet: use updated dsdb commit function
Andrew Tridgell [Tue, 15 Sep 2009 21:42:26 +0000 (14:42 -0700)]
s4-libnet: use updated dsdb commit function

10 years agoMerge branch 'master' of /home/tridge/samba/git/combined
Andrew Tridgell [Tue, 15 Sep 2009 21:08:21 +0000 (14:08 -0700)]
Merge branch 'master' of /home/tridge/samba/git/combined

10 years agos4-ldb: ldap attribute names can contain a '.'
Andrew Tridgell [Tue, 15 Sep 2009 21:07:43 +0000 (14:07 -0700)]
s4-ldb: ldap attribute names can contain a '.'

When they are of the form of OIDs

10 years agos4-ldb: expose ldb_transaction_prepare_commit() in ldb
Andrew Tridgell [Tue, 15 Sep 2009 21:07:06 +0000 (14:07 -0700)]
s4-ldb: expose ldb_transaction_prepare_commit() in ldb

It is useful to be able to control the 2 phase commit from application
code (s4 replication uses it)

10 years agos4-repl: don't do double replication
Andrew Tridgell [Tue, 15 Sep 2009 21:06:07 +0000 (14:06 -0700)]
s4-repl: don't do double replication

When we replicate from a remote DC, we need to note the new uSN that
the local changes have resulted in, and modify the uSN that the notify
task uses to determine if it should send a ReplicaSync message back to
the remote DC. Otherwise we end up always triggering a ReplicaSync
every time we replicate from another DC

10 years agotdb: allow reads after prepare commit
Andrew Tridgell [Tue, 15 Sep 2009 21:04:22 +0000 (14:04 -0700)]
tdb: allow reads after prepare commit

We previously only allowed a commit to happen after a prepare
commit. It is in fact safe to allow reads between a prepare and a
commit, and the s4 replication code can make use of that, so allow it.

10 years agoMerge branch 'master' of /home/tridge/samba/git/combined
Andrew Tridgell [Tue, 15 Sep 2009 18:48:23 +0000 (11:48 -0700)]
Merge branch 'master' of /home/tridge/samba/git/combined

10 years agos4-drs: filter based on local_usn
Andrew Tridgell [Tue, 15 Sep 2009 18:47:42 +0000 (11:47 -0700)]
s4-drs: filter based on local_usn

The getncchanges uSN is in our local space, so we must compare it to
the local_usn in replPropertyMetaData

10 years agos4-repl: make sure we marshal the replPropertyMetaData after the last change
Andrew Tridgell [Tue, 15 Sep 2009 18:46:59 +0000 (11:46 -0700)]
s4-repl: make sure we marshal the replPropertyMetaData after the last change

we were setting local_usn after the marshall, so it wasn't going into
the object

10 years agos4-dsdb: use DLIST_ADD() not DLIST_ADD_END()
Andrew Tridgell [Tue, 15 Sep 2009 17:01:26 +0000 (10:01 -0700)]
s4-dsdb: use DLIST_ADD() not DLIST_ADD_END()

Using DLIST_ADD_END() to construct a long list is very inefficient (it
is O(n^2). These lists are not ordered, so using DLIST_ADD() is much
better.

10 years agos4-ldb: cope better with corruption of tdb records
Andrew Tridgell [Tue, 15 Sep 2009 17:00:24 +0000 (10:00 -0700)]
s4-ldb: cope better with corruption of tdb records

When doing an indexed search if we hit a corrupt record we abandoned
the indexed search and did a full search. The problem was that we
might have sent some records to the caller already, which means the
caller ended up with duplicate records. Fix this by returning a search
error if indexing returns an error and we have given any records to
the caller.

10 years agotalloc: when we enable NULL tracking, reparent the autofree context
Andrew Tridgell [Tue, 15 Sep 2009 16:43:27 +0000 (09:43 -0700)]
talloc: when we enable NULL tracking, reparent the autofree context

If NULL tracking is enabled after the autofree context is initialised
then autofree ends up separate from the null_context. This means that
talloc_report_full() doesn't report the autofree context. Fix this by
reparenting the autofree context when we create the null_context.

10 years agos4-repl: add a debug to make it easier to monitor replication
Andrew Tridgell [Tue, 15 Sep 2009 16:23:14 +0000 (09:23 -0700)]
s4-repl: add a debug to make it easier to monitor replication

10 years agolibreplace: white space cleanups
Björn Jacke [Tue, 15 Sep 2009 18:26:24 +0000 (20:26 +0200)]
libreplace: white space cleanups

10 years agos3: ignore cups-config to tidy up library dependencies
Björn Jacke [Tue, 15 Sep 2009 17:41:58 +0000 (19:41 +0200)]
s3: ignore cups-config to tidy up library dependencies

contrary to krb5-config for example, which outputs useful things, cups-config
--libs does not output libs we have to link against. It outputs libs that cups
linked against. We just have to link against cups.

10 years agolibcli:nbt put util_net.c protos in new header file
Andrew Bartlett [Tue, 15 Sep 2009 17:11:45 +0000 (10:11 -0700)]
libcli:nbt put util_net.c protos in new header file

This fixed a very odd build problem due to util.h importing
system/network.h being imported before the uid_wapper code.

Andrew Bartlett

10 years agos4:schema Add code to provide an index into the subClass tree
Andrew Bartlett [Tue, 15 Sep 2009 15:14:54 +0000 (08:14 -0700)]
s4:schema Add code to provide an index into the subClass tree

In time, this should avoid the astounding (order) complexity of the
objectclass sorting in objectclass.c eventually.

Andrew Bartlett

10 years agos3-dcerpc: really fix remaining old auth level constants. sorry...
Günther Deschner [Tue, 15 Sep 2009 17:32:39 +0000 (19:32 +0200)]
s3-dcerpc: really fix remaining old auth level constants. sorry...

Guenther