samba.git
7 years agolibcli/smb: add SMB2_DIALECT_REVISION_224
Stefan Metzmacher [Thu, 22 Dec 2011 12:07:48 +0000 (13:07 +0100)]
libcli/smb: add SMB2_DIALECT_REVISION_224

This is specified in the new [MS-SMB2] preview document.

metze

7 years agos3-auth Restore shortcut for guest security token
Andrew Bartlett [Wed, 21 Dec 2011 06:26:23 +0000 (17:26 +1100)]
s3-auth Restore shortcut for guest security token

This was lost when the server_info and session_info structures were split.

This helps avoid doing lookups for the guest account to determine the
uid/gid and SID values.

Andrew Bartlett

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Autobuild-User: Stefan Metzmacher <metze@samba.org>
Autobuild-Date: Thu Dec 22 15:51:09 CET 2011 on sn-devel-104

7 years agoruncmd: use set_close_on_exec()
Andrew Tridgell [Wed, 30 Nov 2011 04:18:29 +0000 (15:18 +1100)]
runcmd: use set_close_on_exec()

this prevents a fd leak to child processes

Autobuild-User: Andrew Bartlett <abartlet@samba.org>
Autobuild-Date: Thu Dec 22 14:00:06 CET 2011 on sn-devel-104

7 years agoruncmd: use a pipe for stdin to child processes
Andrew Tridgell [Wed, 30 Nov 2011 03:07:51 +0000 (14:07 +1100)]
runcmd: use a pipe for stdin to child processes

this allows child processes to detect the exit of the parent by
looking for EOF on stdin

7 years agoChange the signature of pthreadpool_finished_job() to return 0
Jeremy Allison [Thu, 22 Dec 2011 04:38:32 +0000 (20:38 -0800)]
Change the signature of pthreadpool_finished_job() to return 0
on success, errno on fail and return the jobid in a separate variable.

I need this fix for my vfs_aio_pthread.c module.

Autobuild-User: Jeremy Allison <jra@samba.org>
Autobuild-Date: Thu Dec 22 12:12:33 CET 2011 on sn-devel-104

7 years agoparam: domain_logons and domain_master are of type enum_bool_auto
Amitay Isaacs [Thu, 22 Dec 2011 05:40:10 +0000 (16:40 +1100)]
param: domain_logons and domain_master are of type enum_bool_auto

These parameters should be defined as int and not bool. This fixes
the test failures on big endian machines.

Autobuild-User: Amitay Isaacs <amitay@samba.org>
Autobuild-Date: Thu Dec 22 10:37:42 CET 2011 on sn-devel-104

7 years agos3-rpc_server: Add my copyright for my previous work here
Andrew Bartlett [Fri, 16 Dec 2011 02:19:06 +0000 (13:19 +1100)]
s3-rpc_server: Add my copyright for my previous work here

Autobuild-User: Andrew Bartlett <abartlet@samba.org>
Autobuild-Date: Thu Dec 22 09:02:57 CET 2011 on sn-devel-104

7 years agos3-rpc_server: Remove old comment
Andrew Bartlett [Fri, 16 Dec 2011 02:18:50 +0000 (13:18 +1100)]
s3-rpc_server: Remove old comment

7 years agos3-rpc: added "rpc_server:default" config option
Andrew Tridgell [Tue, 6 Dec 2011 01:15:40 +0000 (12:15 +1100)]
s3-rpc: added "rpc_server:default" config option

this allows the config to specify a default behaviour (embedded,
external or disabled) for unknown pipes. This is needed to allow the
s3 smbd server to redirect unknown pipes to the s4 RPC server when
using s3 smbd as a file server for a s4 DC. If rpc_server:default is
not specified then this change preserves the old behaviour

7 years agos3-loadparm: mark a few parameters as having P_BYTES type
Andrew Tridgell [Wed, 30 Nov 2011 06:29:10 +0000 (17:29 +1100)]
s3-loadparm: mark a few parameters as having P_BYTES type

this fixes a problem with "max xmit = 32K" resulting in a maximum SMB
transmit size of 32 bytes, which makes for some very interesting
network traces!

This is what was causing smbtorture to consume gigabytes of memory in
the rpc.schannel test

7 years agoparam: Fix the data type for bAvailable
Amitay Isaacs [Wed, 21 Dec 2011 03:08:02 +0000 (14:08 +1100)]
param: Fix the data type for bAvailable

This causes the copy_service() to not copy bAvailable boolean on
big endian machines causing tests to fail.

Autobuild-User: Amitay Isaacs <amitay@samba.org>
Autobuild-Date: Thu Dec 22 05:30:49 CET 2011 on sn-devel-104

7 years agos3: There's no reason not to at least build winbind on darwin
Volker Lendecke [Wed, 21 Dec 2011 21:38:00 +0000 (22:38 +0100)]
s3: There's no reason not to at least build winbind on darwin

It does not necessarily do nsswitch services, but as a NETLOGON proxy
it should work fine

Autobuild-User: Volker Lendecke <vlendec@samba.org>
Autobuild-Date: Thu Dec 22 00:16:44 CET 2011 on sn-devel-104

7 years agos3-dns: prevent from potentially doing wrong SRV DNS lookups.
Günther Deschner [Wed, 21 Dec 2011 14:47:35 +0000 (15:47 +0100)]
s3-dns: prevent from potentially doing wrong SRV DNS lookups.

With an empty sitename we asked for e.g.
_ldap._tcp.._sites.dc._msdcs.AD.EXAMPLE.COM

Guenther

Autobuild-User: Günther Deschner <gd@samba.org>
Autobuild-Date: Wed Dec 21 17:23:25 CET 2011 on sn-devel-104

7 years agoidl: Avoid c++ style comments
Volker Lendecke [Wed, 21 Dec 2011 10:29:38 +0000 (11:29 +0100)]
idl: Avoid c++ style comments

Autobuild-User: Volker Lendecke <vlendec@samba.org>
Autobuild-Date: Wed Dec 21 13:36:01 CET 2011 on sn-devel-104

7 years agotdb: don't free old recovery area when expanding if already at EOF.
Rusty Russell [Wed, 21 Dec 2011 03:47:25 +0000 (14:17 +1030)]
tdb: don't free old recovery area when expanding if already at EOF.

We allocate a new recovery area by expanding the file.  But if the
recovery area is already at the end of file (as shown in at least one
client case), we can simply expand the record, rather than freeing it
and creating a new one.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Autobuild-User: Rusty Russell <rusty@rustcorp.com.au>
Autobuild-Date: Wed Dec 21 06:25:40 CET 2011 on sn-devel-104

7 years agotdb: use same expansion factor logic when expanding for new recovery area.
Rusty Russell [Wed, 21 Dec 2011 03:47:16 +0000 (14:17 +1030)]
tdb: use same expansion factor logic when expanding for new recovery area.

If we're expanding because the current recovery area is too small, we
expand only the amount we need.  This can quickly lead to exponential
growth when we have a slowly-expanding record (hence a
slowly-expanding transaction size).

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
7 years agoAllow an object to be deleted from a directory if the caller has DELETE_CHILD access
Jeremy Allison [Tue, 20 Dec 2011 19:38:37 +0000 (11:38 -0800)]
Allow an object to be deleted from a directory if the caller has DELETE_CHILD access
even if we don't have access to read the ACL on the object. Fixes bug #8673 - NT ACL issue.
Different fix needed for 3.6.x.

Autobuild-User: Jeremy Allison <jra@samba.org>
Autobuild-Date: Tue Dec 20 22:13:51 CET 2011 on sn-devel-104

7 years agos3: Fix some False/NULL hickups
Volker Lendecke [Tue, 20 Dec 2011 09:25:05 +0000 (10:25 +0100)]
s3: Fix some False/NULL hickups

Autobuild-User: Volker Lendecke <vlendec@samba.org>
Autobuild-Date: Tue Dec 20 13:13:17 CET 2011 on sn-devel-104

7 years agosamba-tool: Add transaction wrapper for creating GPO
Amitay Isaacs [Wed, 14 Dec 2011 01:18:57 +0000 (12:18 +1100)]
samba-tool: Add transaction wrapper for creating GPO

This ensures that if the GPT files are not copied via SMB,
AD changes are rolled back.

Autobuild-User: Amitay Isaacs <amitay@samba.org>
Autobuild-Date: Tue Dec 20 09:12:41 CET 2011 on sn-devel-104

7 years agotdb2: Avoid a malloc/memcpy in _tdb1_store
Volker Lendecke [Thu, 15 Dec 2011 09:50:34 +0000 (10:50 +0100)]
tdb2: Avoid a malloc/memcpy in _tdb1_store

Autobuild-User: Volker Lendecke <vlendec@samba.org>
Autobuild-Date: Mon Dec 19 16:53:40 CET 2011 on sn-devel-104

7 years agotdb: Avoid a malloc/memcpy in _tdb_store
Volker Lendecke [Thu, 15 Dec 2011 09:50:34 +0000 (10:50 +0100)]
tdb: Avoid a malloc/memcpy in _tdb_store

7 years agos4-dsdb: Relax the conditions where we can't do a subtree delete
Matthieu Patou [Tue, 15 Nov 2011 10:32:33 +0000 (11:32 +0100)]
s4-dsdb: Relax the conditions where we can't do a subtree delete

If the parent object is a SAM object (as defined in 3.1.1.5.2.3
Special Classes and Attributes of MS-ADTS) then we can use the subtree
delete control even if the object is a critical one.

Autobuild-User: Matthieu Patou <mat@samba.org>
Autobuild-Date: Mon Dec 19 14:32:19 CET 2011 on sn-devel-104

7 years agos4:drsuapi/getncchanges: return WERR_NOMEM if talloc_array() fails
Stefan Metzmacher [Tue, 15 Nov 2011 10:47:42 +0000 (11:47 +0100)]
s4:drsuapi/getncchanges: return WERR_NOMEM if talloc_array() fails

metze

7 years agos4-drs: introduce a timeout in the getncchanges processing to always return something...
Matthieu Patou [Thu, 10 Nov 2011 10:23:40 +0000 (11:23 +0100)]
s4-drs: introduce a timeout in the getncchanges processing to always return something in less than x seconds

Signed-off-by: Andrew Tridgell <tridge@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
7 years agos4-drs: avoid calling unecesserly ldb_msg_find_attr_as_* as this call in unefficient
Matthieu Patou [Mon, 14 Nov 2011 17:53:30 +0000 (18:53 +0100)]
s4-drs: avoid calling unecesserly ldb_msg_find_attr_as_* as this call in unefficient

Current implementation of ldb_msg_find_attr_as_* iterate on the list of
attributes returned by the search and make a string comparison. As we
sorting the array of messages / guids we tend to call this function many
times. By storing the GUID and the USN in a separate structure we are
sure to call this function only once per attribute and object.

Signed-off-by: Andrew Tridgell <tridge@samba.org>
7 years agos4-becomedc: replicate first with DRS_CRITICAL_ONLY and DRS_GET_ANC objects for the...
Matthieu Patou [Fri, 4 Nov 2011 23:22:47 +0000 (00:22 +0100)]
s4-becomedc: replicate first with DRS_CRITICAL_ONLY and DRS_GET_ANC objects for the base dn partition

Windows dcpromo do the same: getncchanges with DRS_GET_ANC and
DRS_CRITICAL_ONLY, then it does a getncchanges without those flags for
the rest.

Signed-off-by: Andrew Tridgell <tridge@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
7 years agos4-join: replicate first with DRS_CRITICAL_ONLY and DRS_GET_ANC objects for the base...
Matthieu Patou [Tue, 25 Oct 2011 18:13:00 +0000 (20:13 +0200)]
s4-join: replicate first with DRS_CRITICAL_ONLY and DRS_GET_ANC objects for the base dn partition

Windows dcpromo do the same: getncchanges with DRS_GET_ANC and
DRS_CRITICAL_ONLY, then it does a getncchanges without those flags for
the rest.

Signed-off-by: Andrew Tridgell <tridge@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
7 years agotdb: be more careful on 4G files.
Rusty Russell [Mon, 19 Dec 2011 05:17:50 +0000 (15:47 +1030)]
tdb: be more careful on 4G files.

I came across a tdb which had wrapped to 4G + 4K, and the contents had been
destroyed by processes which thought it only 4k long.  Fix this by checking
on open, and making tdb_oob() check for wrap itself.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
Autobuild-User: Rusty Russell <rusty@rustcorp.com.au>
Autobuild-Date: Mon Dec 19 07:52:01 CET 2011 on sn-devel-104

7 years agopatch tdb-oob-fix.patch
Rusty Russell [Mon, 19 Dec 2011 05:04:39 +0000 (15:34 +1030)]
patch tdb-oob-fix.patch

7 years agos3-net: Fix the return codes. 0 on success, -1 on failure
Amitay Isaacs [Tue, 6 Dec 2011 07:11:11 +0000 (18:11 +1100)]
s3-net: Fix the return codes. 0 on success, -1 on failure

Autobuild-User: Amitay Isaacs <amitay@samba.org>
Autobuild-Date: Mon Dec 19 01:57:24 CET 2011 on sn-devel-104

7 years agos3: Fix the clustering build
Volker Lendecke [Sat, 17 Dec 2011 16:04:09 +0000 (17:04 +0100)]
s3: Fix the clustering build

ctdb_private.h already defines set_close_on_exec.

Autobuild-User: Volker Lendecke <vlendec@samba.org>
Autobuild-Date: Sat Dec 17 18:41:39 CET 2011 on sn-devel-104

7 years agolibcli/util: add NT_STATUS_NETWORK_SESSION_EXPIRED and NT_STATUS_ALL_SIDS_FILTERED
Stefan Metzmacher [Sat, 17 Dec 2011 10:20:50 +0000 (11:20 +0100)]
libcli/util: add NT_STATUS_NETWORK_SESSION_EXPIRED and NT_STATUS_ALL_SIDS_FILTERED

metze

Autobuild-User: Stefan Metzmacher <metze@samba.org>
Autobuild-Date: Sat Dec 17 14:24:40 CET 2011 on sn-devel-104

7 years agos4 dns: Allow updates based on smb.conf setting
Kai Blin [Fri, 16 Dec 2011 13:25:57 +0000 (14:25 +0100)]
s4 dns: Allow updates based on smb.conf setting

Autobuild-User: Kai Blin <kai@samba.org>
Autobuild-Date: Sat Dec 17 04:19:40 CET 2011 on sn-devel-104

7 years agos4 dns: Add some debug output
Kai Blin [Fri, 16 Dec 2011 13:22:51 +0000 (14:22 +0100)]
s4 dns: Add some debug output

7 years agos4 dns: Actually handle the update request
Kai Blin [Fri, 16 Dec 2011 13:21:52 +0000 (14:21 +0100)]
s4 dns: Actually handle the update request

7 years agos4 dns: Implement RFC-compatible update prescan
Kai Blin [Fri, 16 Dec 2011 13:07:11 +0000 (14:07 +0100)]
s4 dns: Implement RFC-compatible update prescan

7 years agos4 dns: Update prerequisite checking conforming to RFC
Kai Blin [Fri, 16 Dec 2011 12:45:22 +0000 (13:45 +0100)]
s4 dns: Update prerequisite checking conforming to RFC

7 years agos4 dns: Give better test output on RCODE mismatch
Kai Blin [Fri, 16 Dec 2011 10:36:47 +0000 (11:36 +0100)]
s4 dns: Give better test output on RCODE mismatch

7 years agos4 dns: Move record lookup to dns_utils.c
Kai Blin [Fri, 16 Dec 2011 08:43:47 +0000 (09:43 +0100)]
s4 dns: Move record lookup to dns_utils.c

7 years agos4 dns: Implement dns record comparison function
Kai Blin [Fri, 16 Dec 2011 08:24:18 +0000 (09:24 +0100)]
s4 dns: Implement dns record comparison function

7 years agos4 dns: Implement dns name equality check
Kai Blin [Fri, 16 Dec 2011 08:21:51 +0000 (09:21 +0100)]
s4 dns: Implement dns name equality check

7 years agos4 dns: Fix some typos in comments and a whitespace
Kai Blin [Fri, 16 Dec 2011 08:13:31 +0000 (09:13 +0100)]
s4 dns: Fix some typos in comments and a whitespace

7 years agocheck_name() not needed in mkdir.
Jeremy Allison [Fri, 16 Dec 2011 23:48:47 +0000 (15:48 -0800)]
check_name() not needed in mkdir.

Autobuild-User: Jeremy Allison <jra@samba.org>
Autobuild-Date: Sat Dec 17 02:44:55 CET 2011 on sn-devel-104

7 years agoEnsure VALID_STAT before accessing fields.
Jeremy Allison [Fri, 16 Dec 2011 23:48:03 +0000 (15:48 -0800)]
Ensure VALID_STAT before accessing fields.

7 years agos4:torture/rap/printing: check smbcli_rap_netprintqgetinfo() WERR status
Stefan Metzmacher [Fri, 16 Dec 2011 17:34:57 +0000 (18:34 +0100)]
s4:torture/rap/printing: check smbcli_rap_netprintqgetinfo() WERR status

metze

Autobuild-User: Stefan Metzmacher <metze@samba.org>
Autobuild-Date: Sat Dec 17 01:09:32 CET 2011 on sn-devel-104

7 years agoFix bug #8664 - Renaming a symlink fails if the symlink target is outside of the...
Jeremy Allison [Fri, 16 Dec 2011 20:26:45 +0000 (12:26 -0800)]
Fix bug #8664 - Renaming a symlink fails if the symlink target is outside of the share.

Autobuild-User: Jeremy Allison <jra@samba.org>
Autobuild-Date: Fri Dec 16 23:32:26 CET 2011 on sn-devel-104

7 years agoRestrict UCF_UNIX_NAME_LOOKUP (symlink allowed) to restricted set of calls.
Jeremy Allison [Fri, 16 Dec 2011 19:26:41 +0000 (11:26 -0800)]
Restrict UCF_UNIX_NAME_LOOKUP (symlink allowed) to restricted set of calls.

7 years agowafsamba: Fix undefined variable name in error message.
Jelmer Vernooij [Fri, 16 Dec 2011 16:53:01 +0000 (17:53 +0100)]
wafsamba: Fix undefined variable name in error message.

Autobuild-User: Jelmer Vernooij <jelmer@samba.org>
Autobuild-Date: Fri Dec 16 21:53:57 CET 2011 on sn-devel-104

7 years agos3:idmap_autorid: use less transactions
Christian Ambach [Fri, 16 Dec 2011 17:41:54 +0000 (18:41 +0100)]
s3:idmap_autorid: use less transactions

reduce the amount of transactions that are mostly unnecessary because no
updates were done, only reads

Autobuild-User: Christian Ambach <ambi@samba.org>
Autobuild-Date: Fri Dec 16 20:18:37 CET 2011 on sn-devel-104

7 years agos3:dbwrap fix return value of db_tdb_parse
Christian Ambach [Fri, 16 Dec 2011 13:51:46 +0000 (14:51 +0100)]
s3:dbwrap fix return value of db_tdb_parse

use the TDB ecode to determine the NTSTATUS return value
and not the return code that is just -1

7 years agos3:printing/spoolssd: avoid using server_messaging_context()
Stefan Metzmacher [Thu, 15 Dec 2011 10:57:24 +0000 (11:57 +0100)]
s3:printing/spoolssd: avoid using server_messaging_context()

metze

Autobuild-User: Stefan Metzmacher <metze@samba.org>
Autobuild-Date: Fri Dec 16 16:29:41 CET 2011 on sn-devel-104

7 years agos3:smbd: remove unused sconn_server_id()
Stefan Metzmacher [Thu, 15 Dec 2011 10:54:01 +0000 (11:54 +0100)]
s3:smbd: remove unused sconn_server_id()

metze

7 years agos3:smbd: avoid using sconn_server_id()
Stefan Metzmacher [Thu, 15 Dec 2011 10:51:20 +0000 (11:51 +0100)]
s3:smbd: avoid using sconn_server_id()

metze

7 years agos3:locking: avoid using sconn_server_id()
Stefan Metzmacher [Thu, 15 Dec 2011 10:50:43 +0000 (11:50 +0100)]
s3:locking: avoid using sconn_server_id()

metze

7 years agos3:lib/conn_tdb: avoid using sconn_server_id()
Stefan Metzmacher [Thu, 15 Dec 2011 10:50:01 +0000 (11:50 +0100)]
s3:lib/conn_tdb: avoid using sconn_server_id()

metze

7 years agos4:libcli/rap: the caller should get talloc children from call->ndr_pull_{param,data}
Stefan Metzmacher [Fri, 16 Dec 2011 10:26:54 +0000 (11:26 +0100)]
s4:libcli/rap: the caller should get talloc children from call->ndr_pull_{param,data}

This should fix the flakey 'rap.printing.rap_printq_getinfo' test,
see https://git.samba.org/autobuild.flakey/2011-12-16-0022/samba4.stdout

metze

Autobuild-User: Stefan Metzmacher <metze@samba.org>
Autobuild-Date: Fri Dec 16 13:11:42 CET 2011 on sn-devel-104

8 years agos4:libcli/rap: call->ndr_push_{param,data} can be talloc childs of 'call'
Stefan Metzmacher [Fri, 16 Dec 2011 10:25:37 +0000 (11:25 +0100)]
s4:libcli/rap: call->ndr_push_{param,data} can be talloc childs of 'call'

metze

8 years agos4:libcli/rap: use talloc_zero() and initialize everything
Stefan Metzmacher [Fri, 16 Dec 2011 10:18:04 +0000 (11:18 +0100)]
s4:libcli/rap: use talloc_zero() and initialize everything

metze

8 years agos4:libcli/rap: remove unused members from struct rap_call
Stefan Metzmacher [Fri, 16 Dec 2011 08:17:45 +0000 (09:17 +0100)]
s4:libcli/rap: remove unused members from struct rap_call

metze

Autobuild-User: Stefan Metzmacher <metze@samba.org>
Autobuild-Date: Fri Dec 16 10:53:42 CET 2011 on sn-devel-104

8 years agos4:libcli/rap: initialize call->rcv_datalen
Stefan Metzmacher [Fri, 16 Dec 2011 08:16:40 +0000 (09:16 +0100)]
s4:libcli/rap: initialize call->rcv_datalen

metze

8 years agos3: Add some const to find_oplock_types
Volker Lendecke [Thu, 15 Dec 2011 09:12:42 +0000 (10:12 +0100)]
s3: Add some const to find_oplock_types

Autobuild-User: Volker Lendecke <vlendec@samba.org>
Autobuild-Date: Fri Dec 16 09:10:07 CET 2011 on sn-devel-104

8 years agos3: Make open_file_ntcreate a bit easier to read
Volker Lendecke [Thu, 15 Dec 2011 15:30:08 +0000 (16:30 +0100)]
s3: Make open_file_ntcreate a bit easier to read

Move the declaration of "id" closer to its use

8 years agos3: Make open_file_ntcreate a bit easier to read
Volker Lendecke [Thu, 15 Dec 2011 15:30:08 +0000 (16:30 +0100)]
s3: Make open_file_ntcreate a bit easier to read

Move the calculation of "new_file_created" closer to its use

8 years agos3: Make open_file_ntcreate a bit easier to read
Volker Lendecke [Thu, 15 Dec 2011 15:30:08 +0000 (16:30 +0100)]
s3: Make open_file_ntcreate a bit easier to read

Remove a negation where it's not necessary

8 years agoAdd support for TCP_DEFER_ACCEPT
Volker Lendecke [Thu, 15 Dec 2011 15:12:37 +0000 (16:12 +0100)]
Add support for TCP_DEFER_ACCEPT

"man tcp" on Linux says:

TCP_DEFER_ACCEPT

Allows a listener to be awakened only when data arrives on the  socket.  Takes
an integer value (seconds), this can bound the maximum number of attempts TCP
will make to complete the connection. This option should not be used in code
intended to be portable.

This might reduce the 139/445 forks a bit on high-load servers

8 years agos3: Slightly simplify delay_for_exclusive_oplocks
Volker Lendecke [Thu, 15 Dec 2011 13:35:07 +0000 (14:35 +0100)]
s3: Slightly simplify delay_for_exclusive_oplocks

8 years agos3: Slightly simplify delay_for_batch_oplocks
Volker Lendecke [Thu, 15 Dec 2011 13:33:06 +0000 (14:33 +0100)]
s3: Slightly simplify delay_for_batch_oplocks

8 years agoThird part of fix for bug #8663 - deleting a symlink fails if the symlink target...
Jeremy Allison [Fri, 16 Dec 2011 00:55:32 +0000 (16:55 -0800)]
Third part of fix for bug #8663 - deleting a symlink fails if the symlink target is outside of the share.

smbd_check_access_rights() - we can always delete a symlink.

Autobuild-User: Jeremy Allison <jra@samba.org>
Autobuild-Date: Fri Dec 16 03:32:15 CET 2011 on sn-devel-104

8 years agoSecond part of fix for bug #8663 - deleting a symlink fails if the symlink target...
Jeremy Allison [Fri, 16 Dec 2011 00:33:37 +0000 (16:33 -0800)]
Second part of fix for bug #8663 - deleting a symlink fails if the symlink target is outside of the share.

Ensure we use UCF_UNIX_NAME_LOOKUP flags on filename_convert()
when doing a UNIX infolevel in trans2setfilepathinfo().

8 years agoFirst part of fix for bug #8663 - deleting a symlink fails if the symlink target...
Jeremy Allison [Fri, 16 Dec 2011 00:31:51 +0000 (16:31 -0800)]
First part of fix for bug #8663 - deleting a symlink fails if the symlink target is outside of the share.

Remove two unneeded check_name() calls. They have already been done
in order to get here.

8 years agolib/param: Set s4 "host msdfs = true" by default
Andrew Bartlett [Thu, 15 Dec 2011 22:26:04 +0000 (09:26 +1100)]
lib/param: Set s4 "host msdfs = true" by default

This matches the s3 loadparm, and makes this feature available
by default for our users in a DC environment.  (This is needed
for the correct operation of GPOs).

Andrew Bartlett

Autobuild-User: Andrew Bartlett <abartlet@samba.org>
Autobuild-Date: Fri Dec 16 01:08:34 CET 2011 on sn-devel-104

8 years agoloadparm: fixed service list handling
Andrew Tridgell [Tue, 6 Dec 2011 00:30:24 +0000 (11:30 +1100)]
loadparm: fixed service list handling

when you have:

  server services = +smb -s3fs

and 'smb' is already in the list, then this should not be an
error. This ensures that a config that specifically sets the services
it wants doesn't generate an error if the service list being set
happens to be the default

8 years agopidfile: use set_close_on_exec()
Andrew Tridgell [Wed, 30 Nov 2011 04:18:54 +0000 (15:18 +1100)]
pidfile: use set_close_on_exec()

this prevents a fd leak to child processes

8 years agos4-socket: use set_close_on_exec()
Andrew Tridgell [Wed, 30 Nov 2011 04:18:45 +0000 (15:18 +1100)]
s4-socket: use set_close_on_exec()

this prevents a fd leak to child processes

8 years agogenrand: use set_close_on_exec()
Andrew Tridgell [Wed, 30 Nov 2011 04:18:08 +0000 (15:18 +1100)]
genrand: use set_close_on_exec()

this prevents a fd leak to child processes

8 years agolib/util: added set_close_on_exec()
Andrew Tridgell [Wed, 30 Nov 2011 04:17:47 +0000 (15:17 +1100)]
lib/util: added set_close_on_exec()

this was already in tevent_util.c, but library layering prevented us
from using it in some other libraries

8 years agodynconfig: added SBINDIR and BINDIR as updated dynconfig variables
Andrew Tridgell [Tue, 29 Nov 2011 23:07:30 +0000 (10:07 +1100)]
dynconfig: added SBINDIR and BINDIR as updated dynconfig variables

this allows these to work correctly in a build environment, pointing
at bin/

8 years agos3: Fix a typo
Volker Lendecke [Thu, 15 Dec 2011 14:48:33 +0000 (15:48 +0100)]
s3: Fix a typo

Autobuild-User: Volker Lendecke <vlendec@samba.org>
Autobuild-Date: Thu Dec 15 23:35:28 CET 2011 on sn-devel-104

8 years agos3:vfs fix up vfs_gpfs after function pointer renames
Christian Ambach [Thu, 15 Dec 2011 13:31:03 +0000 (14:31 +0100)]
s3:vfs fix up vfs_gpfs after function pointer renames

In 422494a8e630e2ca89386344eaa5346388698a32, vfs_gpfs.c was forgotten

Autobuild-User: Christian Ambach <ambi@samba.org>
Autobuild-Date: Thu Dec 15 21:08:02 CET 2011 on sn-devel-104

8 years agos3-libsmb: Don't duplicate kerberos service tickets.
Andreas Schneider [Mon, 21 Nov 2011 17:19:43 +0000 (18:19 +0100)]
s3-libsmb: Don't duplicate kerberos service tickets.

This fixes bug #8628.

Each time we do a client connection. Each time we call to function to
get the service ticket from the cache we duplicate it. So with each
connection we end up with one or three duplicated tickets.

Autobuild-User: Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date: Thu Dec 15 19:30:42 CET 2011 on sn-devel-104

8 years agos3-dbwrap: All relevant backends provide parse_record(). Remove the fallback.
Volker Lendecke [Fri, 9 Dec 2011 10:22:50 +0000 (11:22 +0100)]
s3-dbwrap: All relevant backends provide parse_record(). Remove the fallback.

Signed-off-by: Michael Adam <obnox@samba.org>
Autobuild-User: Michael Adam <obnox@samba.org>
Autobuild-Date: Thu Dec 15 17:41:53 CET 2011 on sn-devel-104

8 years agos3-dbwrap: Remove the "fetch" db_context callback
Volker Lendecke [Thu, 8 Dec 2011 15:46:09 +0000 (16:46 +0100)]
s3-dbwrap: Remove the "fetch" db_context callback

Signed-off-by: Michael Adam <obnox@samba.org>
8 years agos3-dbwrap: Rewrite dbwrap_fetch in terms of dbwrap_parse_record
Volker Lendecke [Thu, 8 Dec 2011 15:43:40 +0000 (16:43 +0100)]
s3-dbwrap: Rewrite dbwrap_fetch in terms of dbwrap_parse_record

Signed-off-by: Michael Adam <obnox@samba.org>
8 years agos3-dbwrap: Rewrite dbwrap_fallback_parse_record based on dbwrap_fetch_locked
Volker Lendecke [Thu, 8 Dec 2011 15:37:40 +0000 (16:37 +0100)]
s3-dbwrap: Rewrite dbwrap_fallback_parse_record based on dbwrap_fetch_locked

This is in preparation to remove the db_context->fetch function pointer

Signed-off-by: Michael Adam <obnox@samba.org>
8 years agos3-dbwrap: For nostalgic reasons, make dbwrap_file.c compile at least
Volker Lendecke [Thu, 8 Dec 2011 15:05:08 +0000 (16:05 +0100)]
s3-dbwrap: For nostalgic reasons, make dbwrap_file.c compile at least

Signed-off-by: Michael Adam <obnox@samba.org>
8 years agos3-dbwrap: Fix some blank line endings
Volker Lendecke [Thu, 8 Dec 2011 15:04:19 +0000 (16:04 +0100)]
s3-dbwrap: Fix some blank line endings

Signed-off-by: Michael Adam <obnox@samba.org>
8 years agos3-dbwrap: Implement db_ctdb_parse_record in terms of db_ctdb_fetch
Volker Lendecke [Thu, 8 Dec 2011 14:56:35 +0000 (15:56 +0100)]
s3-dbwrap: Implement db_ctdb_parse_record in terms of db_ctdb_fetch

Signed-off-by: Michael Adam <obnox@samba.org>
8 years agos3-dbwrap: Make dbwrap_parse_record return NTSTATUS
Volker Lendecke [Thu, 8 Dec 2011 14:50:33 +0000 (15:50 +0100)]
s3-dbwrap: Make dbwrap_parse_record return NTSTATUS

Also, the parser now returns void. The parser is called if and only if
dbwrap_parse_record returns NT_STATUS_OK.

Signed-off-by: Michael Adam <obnox@samba.org>
8 years agos3:smbd/globals: remove unused msg_ctx_to_sconn()
Stefan Metzmacher [Wed, 14 Dec 2011 12:26:36 +0000 (13:26 +0100)]
s3:smbd/globals: remove unused msg_ctx_to_sconn()

metze

Autobuild-User: Stefan Metzmacher <metze@samba.org>
Autobuild-Date: Thu Dec 15 14:24:59 CET 2011 on sn-devel-104

8 years agos3:smbd: pass smbd_server_connection and a snumused function pointer to reload_services()
Stefan Metzmacher [Wed, 14 Dec 2011 12:25:20 +0000 (13:25 +0100)]
s3:smbd: pass smbd_server_connection and a snumused function pointer to reload_services()

metze

8 years agos3:printing/spoolssd: pass event and messaging context to check_updater_child()
Stefan Metzmacher [Wed, 14 Dec 2011 10:12:49 +0000 (11:12 +0100)]
s3:printing/spoolssd: pass event and messaging context to check_updater_child()

metze

8 years agos3:rpc_server/spoolss: remove reload_services check from delete_printer_hook()
Stefan Metzmacher [Thu, 15 Dec 2011 10:03:29 +0000 (11:03 +0100)]
s3:rpc_server/spoolss: remove reload_services check from delete_printer_hook()

As the spoolss code can run embedded or external relative to the
smbd file server process, it's very tricky to verify if a share
is still in use.

Checking the result of the "deleteprinter command" command should
be enough to check for success. We should not return WERR_ACCESS_DENIED
if the share is still in use, by the current client, as the primary
printer definition is already deleted.

metze

8 years agos3:smbd: split smb_conf_updated into parent and child versions
Stefan Metzmacher [Wed, 14 Dec 2011 11:39:36 +0000 (12:39 +0100)]
s3:smbd: split smb_conf_updated into parent and child versions

metze

8 years agos3:smbd: split ID_CACHE_* message handling into parent and child parts
Stefan Metzmacher [Wed, 14 Dec 2011 09:23:30 +0000 (10:23 +0100)]
s3:smbd: split ID_CACHE_* message handling into parent and child parts

metze

8 years agolib/ccan: Fix some typos in libccan.m4
Martin Schwenke [Thu, 15 Dec 2011 04:31:47 +0000 (15:31 +1100)]
lib/ccan: Fix some typos in libccan.m4

Signed-off-by: Martin Schwenke <martin@meltin.net>
Autobuild-User: Rusty Russell <rusty@rustcorp.com.au>
Autobuild-Date: Thu Dec 15 07:40:33 CET 2011 on sn-devel-104

8 years agolib/ccan: fix m4 tests for builtin_expect and compound literals.
Rusty Russell [Thu, 15 Dec 2011 05:04:55 +0000 (15:34 +1030)]
lib/ccan: fix m4 tests for builtin_expect and compound literals.

Signed-off-by: Rusty Russell <rusty@rustcorp.com.au>
8 years agos3-includes: remove PRIMARY_* SID_INDEX defines defined elsewhere already.
Günther Deschner [Wed, 14 Dec 2011 18:36:52 +0000 (19:36 +0100)]
s3-includes: remove PRIMARY_* SID_INDEX defines defined elsewhere already.

They really do not belong to smb.h.

Guenther

Autobuild-User: Günther Deschner <gd@samba.org>
Autobuild-Date: Wed Dec 14 21:15:39 CET 2011 on sn-devel-104

8 years agos3: Fix raw.mux after UNUSED_SHARE_MODE_ENTRY was removed
Volker Lendecke [Wed, 14 Dec 2011 15:54:07 +0000 (16:54 +0100)]
s3: Fix raw.mux after UNUSED_SHARE_MODE_ENTRY was removed

See the large comment in notify_deferred_opens for an explanation

Autobuild-User: Volker Lendecke <vlendec@samba.org>
Autobuild-Date: Wed Dec 14 19:08:30 CET 2011 on sn-devel-104

8 years agos3: Slightly simplify the logic in defer_open
Volker Lendecke [Wed, 14 Dec 2011 09:43:36 +0000 (10:43 +0100)]
s3: Slightly simplify the logic in defer_open

Doing 3 &&ed conditions in a single if-statement is easier to understand to me
than continuing out separately.

8 years agos3: Remove UNUSED_SHARE_MODE_ENTRY
Volker Lendecke [Wed, 14 Dec 2011 09:35:56 +0000 (10:35 +0100)]
s3: Remove UNUSED_SHARE_MODE_ENTRY