ira/wip.git
8 years agos4-test: added dbcheck run to test suite
Andrew Tridgell [Wed, 13 Jul 2011 00:50:24 +0000 (10:50 +1000)]
s4-test: added dbcheck run to test suite

This should catch corruption that happens during a test run

Pair-Programmed-With: Andrew Bartlett <abartlet@samba.org>

8 years agosamba-tool: nicer error in passwordsettings with no settings
Andrew Tridgell [Tue, 19 Jul 2011 01:39:52 +0000 (11:39 +1000)]
samba-tool: nicer error in passwordsettings with no settings

Pair-Programmed-With: Andrew Bartlett <abartlet@samba.org>

8 years agosamba-tool: testparm doesn't take any credentials
Andrew Tridgell [Tue, 19 Jul 2011 01:19:59 +0000 (11:19 +1000)]
samba-tool: testparm doesn't take any credentials

Pair-Programmed-With: Andrew Bartlett <abartlet@samba.org>

8 years agosamba-tool: use 'exportkeytab' instead of 'dumpkeys'
Andrew Tridgell [Tue, 19 Jul 2011 01:03:44 +0000 (11:03 +1000)]
samba-tool: use 'exportkeytab' instead of 'dumpkeys'

a 'keytab' is a particular format known to administrators, whereas
'keys' is a bit too vague

Pair-Programmed-With: Andrew Bartlett <abartlet@samba.org>

8 years agosamba-tool: Fix __doc_ in base classes
Giampaolo Lauria [Mon, 18 Jul 2011 22:34:45 +0000 (18:34 -0400)]
samba-tool: Fix __doc_ in base classes

Changed prog to samba-tool as prog is only meaningful in Parser

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: removed synopsis code in base class
Giampaolo Lauria [Mon, 18 Jul 2011 21:46:02 +0000 (17:46 -0400)]
samba-tool: removed synopsis code in base class

As it is not always possible to determine the usage of a command solely based on the list of required and optional args, it is best
to have the subclasses always define it, rather than displaying an incorrect usage statement. Currently, all commands are subclassing the synopsis.

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: Fixed bugs to determine min and max # of allowed arguments
Giampaolo Lauria [Mon, 18 Jul 2011 20:48:03 +0000 (16:48 -0400)]
samba-tool: Fixed bugs to determine min and max # of allowed arguments

Fixed the bugs in the code to determine both the min and the max # of allowed arguments
Changed the argument suffix convention from "*" to "+" to represent one or more arguments as:
 1. It follows the Regular expression convention ("*" means 0 or more)
 2. It is what was missing in terms of functionality
NB Currently, no command is using the "*/+", but it is a good thing to have to help out the validation of the args if/when in the future
we have such need

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: Improved --help functionality
Giampaolo Lauria [Mon, 18 Jul 2011 19:45:39 +0000 (15:45 -0400)]
samba-tool: Improved --help functionality

Added a new --help msg
Return an error when no subcommand is specified

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: fixed __doc__ in base classes
Giampaolo Lauria [Mon, 18 Jul 2011 15:50:48 +0000 (11:50 -0400)]
samba-tool: fixed __doc__ in base classes

Replaced the "net" word with %prog in all instances

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: Fix error handling in SuperCommand class
Giampaolo Lauria [Mon, 18 Jul 2011 15:30:23 +0000 (11:30 -0400)]
samba-tool: Fix error handling in SuperCommand class

Created show_command_error method to handle errors in SuperCommand
Removed statement in SuperCommand to raise exception

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: improved Option list for all user commands
Giampaolo Lauria [Mon, 18 Jul 2011 14:03:16 +0000 (10:03 -0400)]
samba-tool: improved Option list for all user commands

Added metavar values for -H and added some default values for other options

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: added error handling for the user command
Giampaolo Lauria [Fri, 15 Jul 2011 19:23:36 +0000 (15:23 -0400)]
samba-tool: added error handling for the user command

Caught exception whenever possible, added new check for newpassword to make sure it contains some chars

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: fixed drs commands synopsis
Giampaolo Lauria [Fri, 15 Jul 2011 18:20:03 +0000 (14:20 -0400)]
samba-tool: fixed drs commands synopsis

Added [options] as needed

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: fixed samba-tool user syntax
Andrew Tridgell [Thu, 21 Jul 2011 00:29:40 +0000 (10:29 +1000)]
samba-tool: fixed samba-tool user syntax

Pair-Programmed-With: Andrew Bartlett <abartlet@samba.org>

8 years agosamba-tool: fixed synopsis on user commands
Giampaolo Lauria [Fri, 15 Jul 2011 17:58:00 +0000 (13:58 -0400)]
samba-tool: fixed synopsis on user commands

Fixed all synopsis to contain [options], filter, and username

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: fixed synopsis on all "user" commands
Giampaolo Lauria [Fri, 15 Jul 2011 17:46:27 +0000 (13:46 -0400)]
samba-tool: fixed synopsis on all "user" commands

Added [options] where needed, fixed others where filter or username was needed, renamed name to username

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: moved takes_optiongroups definition to Command base class
Giampaolo Lauria [Fri, 15 Jul 2011 16:07:03 +0000 (12:07 -0400)]
samba-tool: moved takes_optiongroups definition to Command base class

The option groups should be defined at the Command base class level as they are in common across all samba-tool commands.
Major move advantages:
1. more OOP approach
2. enforcing consistency across commands
3. avoiding the need of declaring for every new command

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: removed the assignment to parser.prog
Giampaolo Lauria [Fri, 15 Jul 2011 13:47:51 +0000 (09:47 -0400)]
samba-tool: removed the assignment to parser.prog

The prog should only be set if we want it to be different than the name of the program that executed it. I think that for better portability we should not set it and let it default to samba-tool.

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: fixed prog name in samba-tool
Giampaolo Lauria [Wed, 6 Jul 2011 18:13:39 +0000 (14:13 -0400)]
samba-tool: fixed prog name in samba-tool

Changed the prog name from net to samba-tool so that the usage statement is now correct

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: update vampire.py message
Giampaolo Lauria [Tue, 28 Jun 2011 19:14:16 +0000 (15:14 -0400)]
samba-tool: update vampire.py message

The update was necessary to reflect the move from join to domain join as part of the object-action work

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: removed join as it has been replaced by domain join
Giampaolo Lauria [Tue, 28 Jun 2011 19:09:04 +0000 (15:09 -0400)]
samba-tool: removed join as it has been replaced by domain join

This is part of the samba-tool work to fit the object-action model

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: moved join to domain join
Giampaolo Lauria [Tue, 28 Jun 2011 19:06:41 +0000 (15:06 -0400)]
samba-tool: moved join to domain join

This is part of the samba-tool work to fit the object-action model

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: updated test suite for the new domain dumpkeys option
Giampaolo Lauria [Tue, 28 Jun 2011 16:29:09 +0000 (12:29 -0400)]
samba-tool: updated test suite for the new domain dumpkeys option

The test suite has been changed to reflect the move from export to "domain dumpkeys" to reflect the object-action model

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: removed export as it has been moved to domain dumpkeys
Giampaolo Lauria [Tue, 28 Jun 2011 16:22:39 +0000 (12:22 -0400)]
samba-tool: removed export as it has been moved to domain dumpkeys

The functionality of export has been moved to domain dumpkeys to fit the object-action model

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: moved export to domain dumpkeys
Giampaolo Lauria [Tue, 28 Jun 2011 16:19:08 +0000 (12:19 -0400)]
samba-tool: moved export to domain dumpkeys

This is part of the samba-tool work to fit the object-action model

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: updated test suite to reflect the move from domainlevel to domain level
Giampaolo Lauria [Tue, 28 Jun 2011 15:53:15 +0000 (11:53 -0400)]
samba-tool: updated test suite to reflect the move from domainlevel to domain level

The test suite needs to reflect the change from domailevel to "domain level" to fit the object-action model

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: removed domainlevel as it has been moved to domain level
Giampaolo Lauria [Tue, 28 Jun 2011 15:48:13 +0000 (11:48 -0400)]
samba-tool: removed domainlevel as it has been moved to domain level

The functionality of domainlevel has been moved the "domain level" to fit the object-action model

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: moved domainlevel to domain level
Giampaolo Lauria [Tue, 28 Jun 2011 15:41:39 +0000 (11:41 -0400)]
samba-tool: moved domainlevel to domain level

This is part of the samba-tool work to fit the object-action model

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: removed machinepw as it has been moved to domain machinepassword
Giampaolo Lauria [Mon, 27 Jun 2011 21:06:35 +0000 (17:06 -0400)]
samba-tool: removed machinepw as it has been moved to domain machinepassword

The functionality of machinepwd has been moved to "domain machinepassword" to fit the object-action model

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: moved machinepw to domain machinepassword
Giampaolo Lauria [Mon, 27 Jun 2011 21:04:10 +0000 (17:04 -0400)]
samba-tool: moved machinepw to domain machinepassword

This is part of the samba-tool work to fit the object-action model

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: update test suite for the new domain object
Giampaolo Lauria [Mon, 27 Jun 2011 20:49:43 +0000 (16:49 -0400)]
samba-tool: update test suite for the new domain object

Changed test suite to reflect the changes from setpassword to "domain setpassword" to fit the object-action model

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: removed pwsettings
Giampaolo Lauria [Mon, 27 Jun 2011 20:37:26 +0000 (16:37 -0400)]
samba-tool: removed pwsettings

pwsettings functionality has been moved to user passwordsettings to fit the object-action model

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: created domain object, moved pwsettings to user passwordsettings
Giampaolo Lauria [Mon, 27 Jun 2011 20:32:22 +0000 (16:32 -0400)]
samba-tool: created domain object, moved pwsettings to user passwordsettings

This is part of the samba-tool work to fit the object-action model

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: update test suite for add setpassword
Giampaolo Lauria [Mon, 27 Jun 2011 18:31:34 +0000 (14:31 -0400)]
samba-tool: update test suite for add setpassword

The test suite needs to change from setpassword to "user setpassword" to reflect the new cmd syntax

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: removed setpassword.py
Giampaolo Lauria [Mon, 27 Jun 2011 17:07:30 +0000 (13:07 -0400)]
samba-tool: removed setpassword.py

The functionality in setppasword has now been moved to "user setpassword" to fit the object-action model

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: added setpassword to user
Giampaolo Lauria [Mon, 27 Jun 2011 16:59:41 +0000 (12:59 -0400)]
samba-tool: added setpassword to user

This is part of the samba-tool work to fit the object-action model

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agosamba-tool: fix summary of the fsmo command to be clearer
Giampaolo Lauria [Wed, 15 Jun 2011 22:41:02 +0000 (18:41 -0400)]
samba-tool: fix summary of the fsmo command to be clearer

fsmo command is for general FSMO management

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-torture: run_simple_posix_open_test(): replace cli_read_old() with
Björn Baumbach [Tue, 19 Jul 2011 14:15:52 +0000 (16:15 +0200)]
s3-torture: run_simple_posix_open_test(): replace cli_read_old() with
cli_read()

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Autobuild-User: Stefan Metzmacher <metze@samba.org>
Autobuild-Date: Wed Jul 20 23:22:09 CEST 2011 on sn-devel-104

8 years agos3-torture: rw_torture2(): replace cli_read_old() with cli_read()
Björn Baumbach [Tue, 19 Jul 2011 14:01:42 +0000 (16:01 +0200)]
s3-torture: rw_torture2(): replace cli_read_old() with cli_read()

Signed-off-by: Stefan Metzmacher <metze@samba.org>
8 years agos3-torture: rw_torture3(): replace cli_read_old() with cli_read()
Björn Baumbach [Tue, 19 Jul 2011 13:49:03 +0000 (15:49 +0200)]
s3-torture: rw_torture3(): replace cli_read_old() with cli_read()

Signed-off-by: Stefan Metzmacher <metze@samba.org>
8 years agos3-torture: rw_torture(): replace cli_read_old() with cli_read()
Björn Baumbach [Tue, 19 Jul 2011 13:29:14 +0000 (15:29 +0200)]
s3-torture: rw_torture(): replace cli_read_old() with cli_read()

Signed-off-by: Stefan Metzmacher <metze@samba.org>
8 years agos3-libsmb: introduce new NTSTATUS cli_read()
Björn Baumbach [Tue, 19 Jul 2011 09:11:27 +0000 (11:11 +0200)]
s3-libsmb: introduce new NTSTATUS cli_read()

Replacement for cli_read_old()

Signed-off-by: Stefan Metzmacher <metze@samba.org>
8 years agos3-libsmb: replace cli_read() with cli_read_old()
Björn Baumbach [Tue, 19 Jul 2011 08:48:16 +0000 (10:48 +0200)]
s3-libsmb: replace cli_read() with cli_read_old()

Will introduce new cli_read() function.

Signed-off-by: Stefan Metzmacher <metze@samba.org>
8 years agos3:libsmb: move cli->cnum to cli->smb1.tid and hide it behind cli_state_[g|s]et_tid()
Stefan Metzmacher [Tue, 19 Jul 2011 14:25:52 +0000 (16:25 +0200)]
s3:libsmb: move cli->cnum to cli->smb1.tid and hide it behind cli_state_[g|s]et_tid()

metze

8 years agos3:libsmb: move cli->pid to cli->smb1.pid and hide it behind cli_[g|s]etpid()
Stefan Metzmacher [Tue, 19 Jul 2011 10:56:51 +0000 (12:56 +0200)]
s3:libsmb: move cli->pid to cli->smb1.pid and hide it behind cli_[g|s]etpid()

metze

8 years agos3:libsmb: add cli->smb1.vc_num and hide it behind cli_state_get_vc_num()
Stefan Metzmacher [Tue, 19 Jul 2011 10:52:10 +0000 (12:52 +0200)]
s3:libsmb: add cli->smb1.vc_num and hide it behind cli_state_get_vc_num()

This makes it clearer, why we send the pid value in the session setup.

metze

8 years agos3:libsmb: move cli->mid to cli->smb1.mid
Stefan Metzmacher [Tue, 19 Jul 2011 09:34:35 +0000 (11:34 +0200)]
s3:libsmb: move cli->mid to cli->smb1.mid

metze

8 years agos3:libsmb: smb_bytes_talloc_string() doesn't need a cli_state
Stefan Metzmacher [Tue, 19 Jul 2011 14:02:31 +0000 (16:02 +0200)]
s3:libsmb: smb_bytes_talloc_string() doesn't need a cli_state

metze

8 years agos4:kdc: restore the behavior before the last heimdal import
Stefan Metzmacher [Wed, 20 Jul 2011 08:22:10 +0000 (10:22 +0200)]
s4:kdc: restore the behavior before the last heimdal import

metze

Autobuild-User: Stefan Metzmacher <metze@samba.org>
Autobuild-Date: Wed Jul 20 12:12:38 CEST 2011 on sn-devel-104

8 years agos3-gse Work around the MIT 1.9 gss_krb5_import_cred
Andrew Bartlett [Wed, 20 Jul 2011 02:06:30 +0000 (12:06 +1000)]
s3-gse Work around the MIT 1.9 gss_krb5_import_cred

We detect this function at configure time, but it currently fails to
operate the way we need - that is, when the principal is not
specified, it gives this error.  When the principal is specified we
get 'wrong principal in request' in the GSS acceptor, so for now the
best option is to fall back to the alternate approach.

Andrew Bartlett

Autobuild-User: Andrew Bartlett <abartlet@samba.org>
Autobuild-Date: Wed Jul 20 06:35:05 CEST 2011 on sn-devel-104

8 years agos3-gse Allow printing the partial error string
Andrew Bartlett [Wed, 20 Jul 2011 02:04:45 +0000 (12:04 +1000)]
s3-gse Allow printing the partial error string

We may not be able to obtain the full error string, so print what we can get.

This is required when the error is the the GSSAPI layer, not the mechanism.

Andrew Bartlett

8 years agos3-auth fix dummy function in the not-with-kerberos case
Andrew Bartlett [Wed, 20 Jul 2011 01:40:02 +0000 (11:40 +1000)]
s3-auth fix dummy function in the not-with-kerberos case

8 years agos3-auth Replace False with false in auth_util.c
Andrew Bartlett [Tue, 19 Jul 2011 10:15:45 +0000 (20:15 +1000)]
s3-auth Replace False with false in auth_util.c

Autobuild-User: Andrew Bartlett <abartlet@samba.org>
Autobuild-Date: Wed Jul 20 02:31:15 CEST 2011 on sn-devel-104

8 years agos3-auth Replace True with true in auth_util.c
Andrew Bartlett [Tue, 19 Jul 2011 10:15:12 +0000 (20:15 +1000)]
s3-auth Replace True with true in auth_util.c

8 years agos3-auth Fix spelling
Andrew Bartlett [Tue, 19 Jul 2011 10:11:22 +0000 (20:11 +1000)]
s3-auth Fix spelling

8 years agos3-auth Remove pointless destructor in make_server_info
Andrew Bartlett [Tue, 19 Jul 2011 06:05:32 +0000 (16:05 +1000)]
s3-auth Remove pointless destructor in make_server_info

All the callers allocate ->info3 as a talloc child already.

As regardes the TALLOC_ZERO(), I added this originally out of parinoia
many years ago.  We do not consistantly zero session keys in memory,
and for NTLMv2 and Kerberos they are random for each sesssion, so
breaking into smbd far enough to read an old session key isn't a
particularly interesting attack, compared with (say) reading the
keytab or the password database.  (NTLM and LM session keys are fixed
derivitives of the passwords however).

Andrew Bartlett

8 years agos3-auth inline make_auth_session_info into only caller
Andrew Bartlett [Tue, 19 Jul 2011 05:58:20 +0000 (15:58 +1000)]
s3-auth inline make_auth_session_info into only caller

8 years agosecurity.idl: Use gid_t for gid in security_unix_token
Andrew Bartlett [Tue, 19 Jul 2011 05:50:49 +0000 (15:50 +1000)]
security.idl: Use gid_t for gid in security_unix_token

8 years agos3-auth Remove seperate guest boolean
Andrew Bartlett [Tue, 19 Jul 2011 01:57:05 +0000 (11:57 +1000)]
s3-auth Remove seperate guest boolean

Instead, we base our guest calculations on the presence or absense of the
authenticated users group in the token, ensuring that we have only
one canonical source of this important piece of authorization data

Andrew Bartlett

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agoauth: Set NETLOGON_GUEST and use it to determine guest status
Andrew Bartlett [Tue, 19 Jul 2011 00:51:08 +0000 (10:51 +1000)]
auth: Set NETLOGON_GUEST and use it to determine guest status

These additional measures should help ensure we do not accidentily upgrade
a guest to an authenticated user in the future.

Andrew Bartlett

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agoselftest: Add tests to verify that the named pipe proxy works.
Andrew Bartlett [Mon, 18 Jul 2011 10:26:26 +0000 (20:26 +1000)]
selftest: Add tests to verify that the named pipe proxy works.

This verifies that for NTLM authenticated connections, named pipe
forwarding works as expected, including the session keys.

Andrew Bartlett

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agoselftest: Pass lsass and epmapper across the named pipe proxy to the AD server
Andrew Bartlett [Mon, 18 Jul 2011 09:56:17 +0000 (19:56 +1000)]
selftest: Pass lsass and epmapper across the named pipe proxy to the AD server

Eventually we will have just one end point mapper, but for now we need
to use the source4 one for the AD tests.

Andrew Bartlett

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agoauth: remove now unused auth3_session_info from auth.idl
Andrew Bartlett [Mon, 18 Jul 2011 08:29:47 +0000 (18:29 +1000)]
auth: remove now unused auth3_session_info from auth.idl

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agoauth: Move make_user_info_SamBaseInfo() to talloc_strdup and out of memory checking
Andrew Bartlett [Mon, 18 Jul 2011 04:00:14 +0000 (14:00 +1000)]
auth: Move make_user_info_SamBaseInfo() to talloc_strdup and out of memory checking

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agoauth: Split out make_user_info_SamBaseInfo and add authenticated argument
Andrew Bartlett [Mon, 18 Jul 2011 03:55:20 +0000 (13:55 +1000)]
auth: Split out make_user_info_SamBaseInfo and add authenticated argument

This will allow the source3 auth code to call this without needing to
double-parse the SIDs

Andrew Bartlett

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-rpc_server remove per-element copies of auth_session_info
Andrew Bartlett [Mon, 18 Jul 2011 03:10:30 +0000 (13:10 +1000)]
s3-rpc_server remove per-element copies of auth_session_info

This is not required any more now that they are the same structure,
and shows the value in having a common structure across the codebase.

In particular, now any additional state that needs to be added to the
auth_session_info will be transparently available across the named
pipe proxy, without a need to modify the mapping layer.

Andrew Bartlett

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-auth Use the common auth_session_info
Andrew Bartlett [Mon, 18 Jul 2011 03:06:47 +0000 (13:06 +1000)]
s3-auth Use the common auth_session_info

This patch finally has the same structure being used to describe the
authorization data of a user across the whole codebase.

This will allow of our session handling to be accomplished with common code.

Andrew Bartlett

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-auth use auth_user_info not netr_SamInfo3 in auth3_session_info
Andrew Bartlett [Mon, 18 Jul 2011 02:58:25 +0000 (12:58 +1000)]
s3-auth use auth_user_info not netr_SamInfo3 in auth3_session_info

This makes auth3_session_info identical to auth_session_info

The logic to convert the info3 to a struct auth_user_info is
essentially moved up the stack from the named pipe proxy in
source3/rpc_server to create_local_token().

Andrew Bartlett

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-rpc_server read and write the unix_token and unix_info across named_pipe_auth
Andrew Bartlett [Mon, 18 Jul 2011 02:23:04 +0000 (12:23 +1000)]
s3-rpc_server read and write the unix_token and unix_info across named_pipe_auth

This ensures that the exact same token is used on both sides of the
pipe, when a full token is passed (ie, source3 to source3, but not yet
source4 to to source3 as the unix info isn't calculated there yet).

If we do not have unix_token, we fall back to the old behaviour and go
via create_local_token().  (However, in this case the security_token
is now overwritten, as it is better to have it match the rest of the
session_info create_local_token() builds).

Andrew Bartlett

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-auth reimplement copy_session_info via NDR pull/push
Andrew Bartlett [Mon, 18 Jul 2011 02:38:05 +0000 (12:38 +1000)]
s3-auth reimplement copy_session_info via NDR pull/push

This ensures we do not miss elements.  Pattern copied from auth_netlogond.

Andrew Bartlett

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agoauth: use char * pointers in auth.idl
Andrew Bartlett [Mon, 18 Jul 2011 08:04:12 +0000 (18:04 +1000)]
auth: use char * pointers in auth.idl

We need to use this, and not utf8string because we need to
transport NULL pointers correctly.

Andrew Bartlett

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-auth Remove pointless destructor
Andrew Bartlett [Mon, 18 Jul 2011 02:29:50 +0000 (12:29 +1000)]
s3-auth Remove pointless destructor

All the users of this structure allocate info3 on the session_info

Andrew Bartlett

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-auth import auth3_session_info into IDL
Andrew Bartlett [Mon, 18 Jul 2011 02:28:50 +0000 (12:28 +1000)]
s3-auth import auth3_session_info into IDL

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-auth Avoid redundant copies in create_local_token()
Andrew Bartlett [Mon, 18 Jul 2011 01:40:36 +0000 (11:40 +1000)]
s3-auth Avoid redundant copies in create_local_token()

These values were not read before being overwritten again.

Andrew Bartlett

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-auth Add comments to copy_session_info_serverinfo_guest()
Andrew Bartlett [Mon, 18 Jul 2011 01:31:49 +0000 (11:31 +1000)]
s3-auth Add comments to copy_session_info_serverinfo_guest()

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-auth inline copy_serverinfo_session_info into only caller
Andrew Bartlett [Mon, 18 Jul 2011 01:30:55 +0000 (11:30 +1000)]
s3-auth inline copy_serverinfo_session_info into only caller

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-auth use a cached auth_serversupplied_info in make_server_info_guest()
Andrew Bartlett [Mon, 18 Jul 2011 01:22:50 +0000 (11:22 +1000)]
s3-auth use a cached auth_serversupplied_info in make_server_info_guest()

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-auth remove extra from auth3_session_info
Andrew Bartlett [Mon, 18 Jul 2011 00:20:25 +0000 (10:20 +1000)]
s3-auth remove extra from auth3_session_info

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-auth Clarify inputs and ouptuts by using elements from server_info
Andrew Bartlett [Mon, 18 Jul 2011 00:14:19 +0000 (10:14 +1000)]
s3-auth Clarify inputs and ouptuts by using elements from server_info

This allows us not the put all of these elements into the auth3_session_info
if they are only used as inputs to these functions.

Andrew Bartlett

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-auth assert that security_token is present in the copy, and explain why nss_token...
Andrew Bartlett [Fri, 15 Jul 2011 08:05:42 +0000 (18:05 +1000)]
s3-auth assert that security_token is present in the copy, and explain why nss_token can be skipped

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-auth Remove unused nss_token variable
Andrew Bartlett [Fri, 15 Jul 2011 07:57:55 +0000 (17:57 +1000)]
s3-auth Remove unused nss_token variable

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-auth: Remove unused lm_session_key from auth3_session_info
Andrew Bartlett [Fri, 15 Jul 2011 07:45:48 +0000 (17:45 +1000)]
s3-auth: Remove unused lm_session_key from auth3_session_info

The long term authorization state needs only the final, negotiated
session key, and not the original LM key that may possibly have been
an input.

The special case of the guest account simply needs both values filled
back in with the zeros to avoid changing behaviour in the cached
server_info.

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-auth remove unused copy_serverinfo
Andrew Bartlett [Fri, 15 Jul 2011 06:49:21 +0000 (16:49 +1000)]
s3-auth remove unused copy_serverinfo

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-auth Use system boolean in auth_user_info_unix
Andrew Bartlett [Fri, 15 Jul 2011 06:12:41 +0000 (16:12 +1000)]
s3-auth Use system boolean in auth_user_info_unix

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-auth Use guest boolean in auth_user_info_unix
Andrew Bartlett [Fri, 15 Jul 2011 06:09:52 +0000 (16:09 +1000)]
s3-auth Use guest boolean in auth_user_info_unix

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agoauth: Put 'guest' and 'system' booleans into auth_user_info_unix
Andrew Bartlett [Fri, 15 Jul 2011 05:10:03 +0000 (15:10 +1000)]
auth: Put 'guest' and 'system' booleans into auth_user_info_unix

This will allow a transformation of auth3_session_info into
auth_session_info by substitution.

Andrew Bartlett

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-auth Use struct auth_user_info_unix for unix_name and sanitized_username
Andrew Bartlett [Fri, 15 Jul 2011 05:55:31 +0000 (15:55 +1000)]
s3-auth Use struct auth_user_info_unix for unix_name and sanitized_username

This is closer to the layout of struct auth_session_info in auth.idl

Andrew Bartlett

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agoauth: include auth.idl structures into common_auth.h
Andrew Bartlett [Fri, 15 Jul 2011 05:22:41 +0000 (15:22 +1000)]
auth: include auth.idl structures into common_auth.h

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-auth Use *unix_token rather than utok in struct auth3_session_info
Andrew Bartlett [Fri, 15 Jul 2011 04:59:14 +0000 (14:59 +1000)]
s3-auth Use *unix_token rather than utok in struct auth3_session_info

This brings this structure one step closer to the struct auth_session_info.

A few SMB_ASSERT calls are added in some key places to ensure that
this pointer is initialised, to make tracing any bugs here easier in
future.

NOTE: Many of the users of this structure should be reviewed, as unix
and NT access checks are mixed in a way that should just be done using
the NT ACL.  This patch has not changed this behaviour however.

Andrew Bartlett

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-auth Use struct auth3_session_info outside the auth subsystem
Andrew Bartlett [Fri, 15 Jul 2011 02:45:17 +0000 (12:45 +1000)]
s3-auth Use struct auth3_session_info outside the auth subsystem

This seperation between the structure used inside the auth modules and
in the wider codebase allows for a gradual migration from struct
auth_serversupplied_info -> struct auth_session_info (from auth.idl)

The idea here is that we keep a clear seperation between the structure
before and after the local groups, local user lookup and the session
key modifications have been processed, as the lack of this seperation
has caused issues in the past.

Andrew Bartlett

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-auth Add struct auth3_session_info to aid transition to auth_session info
Andrew Bartlett [Fri, 15 Jul 2011 01:38:49 +0000 (11:38 +1000)]
s3-auth Add struct auth3_session_info to aid transition to auth_session info

This will allow a gradual conversion of the required elements from the
current struct auth_serversupplied_info.

This commit adds the structure definition and some helper functions to
copy between the two structures.

At this stage these structures and functions are IDENTICAL to the
existing code, and so show the past history of that code.  The plan is
to slowly modify them over the course of the patch series, so that the
changes being made a clear.

By using a seperate structure to auth_serversupplied_info we can
remove elements that are not needed after the authentication, and we
can choose a layout that best reflects the needs of runtime users,
rather than the internals of the authentication subsystem.

By eventually using the auth_session_info from auth.idl, we will gain
a single session authorization structure across the whole codebase,
allowing more code to be shared, and a much more transparent process
for forwarding authorization credentials over the named pipe proxy.

Andrew Bartlett

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-auth Add const to indicate input elements
Andrew Bartlett [Mon, 18 Jul 2011 00:19:27 +0000 (10:19 +1000)]
s3-auth Add const to indicate input elements

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agoauth: Preserve guest flag on transition via netr_SamInfo3
Andrew Bartlett [Tue, 19 Jul 2011 00:43:23 +0000 (10:43 +1000)]
auth: Preserve guest flag on transition via netr_SamInfo3

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos3-auth Restore nss_token behaviour by reading from server_info
Andrew Bartlett [Fri, 15 Jul 2011 07:57:35 +0000 (17:57 +1000)]
s3-auth Restore nss_token behaviour by reading from server_info

The implementation of copy_serverinfo(), used to copy server_info into
session_info never copied the nss_token variable, and so
17d8f0ad30847bb940f645ee1817d782ddaaee74 introduced this regression.

Andrew Bartlett

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agoAdd my copyright
Andrew Bartlett [Mon, 18 Jul 2011 12:26:31 +0000 (22:26 +1000)]
Add my copyright

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agos4-param Handle P_CHAR and P_BOOLREV in pyparam
Andrew Bartlett [Fri, 15 Jul 2011 01:37:39 +0000 (11:37 +1000)]
s4-param Handle P_CHAR and P_BOOLREV in pyparam

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agodebug: log early messages to stdout, and keep it open
Andrew Bartlett [Mon, 18 Jul 2011 07:07:25 +0000 (17:07 +1000)]
debug: log early messages to stdout, and keep it open

The --log-stdout option was compromised by the log file descriptors being
closed once the file process forked.

Andrew Bartlett

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agolib/util Change debug priority order: DEBUG_STDOUT now overrides DEBUG_FILE
Andrew Bartlett [Mon, 18 Jul 2011 06:00:16 +0000 (16:00 +1000)]
lib/util Change debug priority order: DEBUG_STDOUT now overrides DEBUG_FILE

Signed-off-by: Andrew Tridgell <tridge@samba.org>
8 years agoSecond part of fix for bug 8310 - toupper_ascii() is broken on big-endian systems.
Jeremy Allison [Tue, 19 Jul 2011 20:35:45 +0000 (13:35 -0700)]
Second part of fix for bug 8310 - toupper_ascii() is broken on big-endian systems.

Re-add:
smb_ucs2_t toupper_w(smb_ucs2_t v);

and ensure it is called whenever we are operating on smb_ucs2_t
variables. I'd like to make the definition of smb_ucs2_t incompatible
with int and codepoint_t so they can't be mixed, but that's a patch
for another time.

Autobuild-User: Jeremy Allison <jra@samba.org>
Autobuild-Date: Tue Jul 19 23:48:05 CEST 2011 on sn-devel-104

8 years agoFirst part of fix for bug 8310 - toupper_ascii() is broken on big-endian systems
Jeremy Allison [Tue, 19 Jul 2011 20:19:29 +0000 (13:19 -0700)]
First part of fix for bug 8310 - toupper_ascii() is broken on big-endian systems

Remove
int toupper_ascii(int c);
int tolower_ascii(int c);
int isupper_ascii(int c);
int islower_ascii(int c);

and replace with their _m equivalents, as they are identical.