testsuite: Remove build_farm testsuites

This test code is not run (and has not been run for about a decade).
Let's remove it - it's there in the git history if we ever want to try
to repurpose it again.

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Tue Mar 12 02:56:05 UTC 2019 on sn-devel-144

selftest: Remove RUN_FROM_BUILD_FARM env variable

The build farm is no longer used, so we can remove checks for this
environment variable in the selftest code.

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4/scripting: Remove legacy selftest-vars.sh script

This script was added in 2010 and has only been sporadically kept
up-to-date since. It doesn't appear to work (I think that selftest
and the testenvs have perhaps grown in complexity since 2010 and it's no
longer possible to try to access a testenv from a different
process-space, due to how we use the cwrap libraries).

There's now an alternative (export_envvars_to_file()) in the selftest
code to regenerate a similar file, if anyone actually needs it.

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

selftest: Add helper function to create exports file for a testenv

This writes out the environment variables that are normally setup in the
testenv xterm to a file. This allows them to be sourced later.

This function is currently unused. However, it provides an alternative
replacement for the selftest-vars.sh script (which is really out of
date). I do plan to make use of this function in a subsequent patch-set.

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

selftest: Move @exported_envvars into Samba.pm

This brings it closer to the code that actually sets these variables
(i.e. Samba3/4.pm).

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

selftest: Remove unused SELFTEST_INTERFACES

This doesn't appear to be used anywhere and dates back to 2008.

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

samba-tool domain provision: Fix --interactive module in python3

The prompts were not being printed to the screen because the stream
was not being flushed.

As reported on the samba mailing list by Adam Xu:
https://lists.samba.org/archive/samba/2019-March/221753.html

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13828

Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Tim Beale <timbeale@catalyst.net.nz>

gpo: Backup a policy folder containing GPE.INI

The GPE.INI file does not appear to be documented anywhere in the
protocol specifications and seems to be due to legacy code. It appears
that it used to be how the gPCUserExtensionNames and
gPCMachineExtensionNames were maintained without the requirement for
LDAP.

All we do is ignore the parsing of this file and copy it over as binary.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13825

Signed-off-by: Garming Sam <garming@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Tue Mar 12 01:41:32 UTC 2019 on sn-devel-144

gpo: During restore clobber GPT.INI with a blank version

Generally speaking, there is not much value to keeping this file. The
display name does not ever seem to be used and the version only applies
to the original domain or DC it was on. The command line option to
revert this behaviour is mostly for the tests or for having a straight
1:1 backup-restore for pure restoration.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13806

Signed-off-by: Garming Sam <garming@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

gpo: Parse GPT.INI with Latin-1

For some reason the French version of RSAT turns accents into ISO-8859-1.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13806

Signed-off-by: Garming Sam <garming@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

gpo: Add test for saving gPCXXXExtensionNames in backup

The test only checks that the LDAP values are equal before and after.
The format and specifics should be up to the implementation -- and be
amenable to changes later. As long as the extension doesn't match
.SAMBABACKUP or .xml, it won't be detected by the tools.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13627

Signed-off-by: Garming Sam <garming@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

gpo: Restore gPCMachineExtensionNames and gPCUserExtensionNames

After creating a backup and calling 'gpo restore', this makes it so that
restoring a GPO will instantly enable it for use.

There might be some cases where we might not want to do this, but for now just do it.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13627

Signed-off-by: Garming Sam <garming@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

py3: io.open takes a numeric buffering argument at index 2

It's unsure why this doesn't fail generically, but it fails on my machine sometimes...

Signed-off-by: Garming Sam <garming@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

gpo: Add tests to make sure that an empty gplink works

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13564

Signed-off-by: Garming Sam <garming@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

gpo: Ensure that gplink works when empty

It appears that RSAT can leave a space in the gPLink field, which we need to handle.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13564

Signed-off-by: Garming Sam <garming@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

modules: Add dependency on tirpc to vfs_nfs4acl_xattr

This is done as a new subsystem (either filled or empty) rather than via string
manipulation.

This will fix compile error on fedora.

Signed-off-by: Joe Guo <joeg@catalyst.net.nz>
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>

libsmb: Use sid_parse()

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

lib: Remove "struct sid_parse_ret" again

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

lib: Make sid_parse return the parsed length

Use a temporary struct as a return value to make the compiler catch all
callers. If we just changed bool->ssize_t, this would just generate a
warning. struct sid_parse_ret will go away in the next commit

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

Avoid NULL pointer dereference in SMBsendend handler

The "reply_sendend" function wouldn't check whether the connection had
any pending message state. A client sending an out-of-order SMBsendend
message would trigger a NULL pointer dereference.

Reviewed-by: Garming Sam <garming@catalyst.net.nz>
Signed-off-by: Michael Hanselmann <public@hansmi.ch>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4/scripting/autoidl: p3 exception syntax

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

test/blackbox: py3 compatible print in documentation.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4/scripting: MORE py3 compatible print functions

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

selftest/filter-subunit: use py3 print

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

selftest/format-subunit-json: remove useless py2 print

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

pidl/s4/python: call export "export" in py exceptions

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

libsmb: Use tevent_req_simple_finish_ntstatus

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Fri Mar  8 19:16:18 UTC 2019 on sn-devel-144

libsmb: Add "in_cblobs" to cli_smb2_unlink

This reveals the fact that unlink is an open/close in smb2 through the
API. This is not nice, but it's an internal API with currently only
one user. And it enables posix semantics for the open easily.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

libsmb: Add "in_cblobs" to cli_smb2_rmdir

This reveals the fact that rmdir is an open/close in smb2 through the
API. This is not nice, but it's an internal API with currently only
one user. And it enables posix semantics for the open easily.

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

libsmb: Make cli_smb2_unlink async

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

libsmb: Simplify cli_smb2_mxac

smb2_create_blob_find() can search for a create blob for us

Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

lib:util: Move debug message for mkdir failing to log level 1

If you connnect to a host with smbclient this gets always printed.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13823

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Fri Mar  8 01:41:27 UTC 2019 on sn-devel-144

samba-o3: fix -Werror=maybe-uninitialized in lib/mscat/mscat_pks7.c

samba-o3 test failed in ubuntu:1804 image with:

    ../../lib/mscat/mscat_pkcs7.c: In function ‘mscat_pkcs7_import_catfile’:
    ../../lib/mscat/mscat_pkcs7.c:143:18: error: ‘blob.length’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
      mscat_data.size = blob.length;
      ~~~~~~~~~~~~~~~~^~~~~~~~~~~~~
    ../../lib/mscat/mscat_pkcs7.c:142:18: error: ‘blob.data’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
      mscat_data.data = blob.data;
      ~~~~~~~~~~~~~~~~^~~~~~~~~~~
    ../../lib/mscat/mscat_pkcs7.c: In function ‘mscat_pkcs7_verify’:
    ../../lib/mscat/mscat_pkcs7.c:225:16: error: ‘blob.length’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
       ca_data.size = blob.length;
       ~~~~~~~~~~~~~^~~~~~~~~~~~~
    ../../lib/mscat/mscat_pkcs7.c:224:16: error: ‘blob.data’ may be used uninitialized in this function [-Werror=maybe-uninitialized]
       ca_data.data = blob.data;
       ~~~~~~~~~~~~~^~~~~~~~~~~
    cc1: all warnings being treated as errors

Since in `mscat_read_file`, it may still return rc = 0 while goto error,
ends up with blob uninitialized.

Signed-off-by: Joe Guo <joeg@catalyst.net.nz>
Reviewed-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

samba-o3: fix -Werror=strict-overflow error in s4/torture/raw/eas module

samba-o3 test failed in ubuntu:16.04 docker container:

    ==> /home/samba/samba/samba-o3.stderr <==
    ../../source4/torture/raw/eas.c: In function ‘test_max_eas’:
    ../../source4/torture/raw/eas.c:286:12: error: assuming signed overflow does not occur when simplifying conditional to constant [-Werror=strict-overflow]
     static bool test_max_eas(struct smbcli_state *cli, struct torture_context *tctx)
                ^
    cc1: all warnings being treated as errors

`total += j` may overflow. Change total type to `size_t` to mute error.

Signed-off-by: Joe Guo <joeg@catalyst.net.nz>
Reviewed-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

samba-o3: fix -Werror=strict-overflow error in lib/ldb-samba/ldb_ildap module

samba-o3 test failed in ubuntu:16.04 docker container:

    ==> /home/samba/samba/samba-o3.stderr <==
    ../../lib/ldb-samba/ldb_ildap.c: In function ‘ildb_handle_request’:
    ../../lib/ldb-samba/ldb_ildap.c:535:2: error: assuming signed overflow does not occur when simplifying conditional to constant [-Werror=strict-overflow]
      for (i = 0; i < n; i++) {
      ^
    ../../lib/ldb-samba/ldb_ildap.c:579:2: error: assuming signed overflow does not occur when simplifying conditional to constant [-Werror=strict-overflow]
      for (i = 0; i < n; i++) {
      ^
    cc1: all warnings being treated as errors

Change type to mute errors.

Signed-off-by: Joe Guo <joeg@catalyst.net.nz>
Reviewed-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

pygpo: take ownership of password pointer

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13822
Signed-off-by: Kristján Valur Jónsson <kristjan@rvx.is>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Noel Power <npower@samba.org>
Autobuild-User(master): Noel Power <npower@samba.org>
Autobuild-Date(master): Thu Mar  7 15:08:19 UTC 2019 on sn-devel-144

pygpo: Safer handling of memory for ads_ptr.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13822
Signed-off-by: Kristján Valur Jónsson <kristjan@rvx.is>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Noel Power <npower@samba.org>

pygpo: Fix module initialization.

* Add reference count to type.

* Add error checking.

* Remove unnecessary tp_new method.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13822
Signed-off-by: Kristján Valur Jónsson <kristjan@rvx.is>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Noel Power <npower@samba.org>

pygpo: keep a reference to python credentials in the ADS struct to keep the internal pointer valid.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13822
Signed-off-by: Kristján Valur Jónsson <kristjan@rvx.is>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Noel Power <npower@samba.org>

pygpo: More python exception cleanup.

* Don't override existing exceptions.

* Careful with talloc contexts.

* Return NULL on error.

* Add more information to exception messages from internal functions.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13822
Signed-off-by: Kristján Valur Jónsson <kristjan@rvx.is>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Noel Power <npower@samba.org>

pygpo: Fix error handing when getting gpo unix path.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13822
Signed-off-by: Kristján Valur Jónsson <kristjan@rvx.is>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Noel Power <npower@samba.org>

pygpo: Proper exception exit in py_ads_connect().

connect() now succeeds or raises an exception.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13822
Signed-off-by: Kristján Valur Jónsson <kristjan@rvx.is>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Noel Power <npower@samba.org>

pygpo: Replace the use of SystemError with RuntimeError.

SystemError is reserved for internal errors in the interpreter.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13822
Signed-off-by: Kristján Valur Jónsson <kristjan@rvx.is>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Noel Power <npower@samba.org>

subunit/run.py: change shebang to python3

always use explicit python version at current stage.

Signed-off-by: Joe Guo <joeg@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Noel Power <npower@samba.org>
Autobuild-User(master): Noel Power <npower@samba.org>
Autobuild-Date(master): Thu Mar  7 13:03:56 UTC 2019 on sn-devel-144

tests/auto_log_pass_change.py: only care about the last expected message other than exact messages count

The messages count could be different because of racing condition.
And we should only care about the last expected one.

Signed-off-by: Joe Guo <joeg@catalyst.net.nz>
Reviewed-by: Andrew Bartlett abartlet@samba.org
Reviewed-by: Noel Power npower@samba.org

subunit/run.py: make iso8601 UTC usage python 2/3 compatible

In `iso8601/iso8601.py`:

    if sys.version_info >= (3, 2, 0):
        UTC = datetime.timezone.utc
        ...
    else:
        class Utc(datetime.tzinfo):
            ...

        UTC = Utc()

The class `Utc` is only available for python < 3.2.0.
Use `UTC` instance instead, which is python 2/3 compatible.

Signed-off-by: Joe Guo <joeg@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Noel Power <npower@samba.org>

s4-server: Open and close a transaction on sam.ldb at startup

This fixes upgrading from 4.7 and earlier releases, and makes the DB
reindexing more transparent. It should also make it easier to handle
future normalisation rule changes, e.g. if we change the pack-format
of integer indexes in a future release.

Without this change, the  should have still handled reindexing the
database. We don't know why exactly this wasn't happening correctly,
but opening a transaction early in the samba process startup should
now guarantee that the DB is correctly reindexed by the time the main
samba code runs.

An alternative fix would have been to open a transaction in the the
DSDB module stack every time we connect to the database. However, this
would add an extra write lock every time we open the DB, whereas
starting samba happens much more infrequently.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13760

Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Gary Lockyer <gary@catalyst.net.nz>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Thu Mar  7 04:58:42 UTC 2019 on sn-devel-144

dsdb: Provide better error strings in rootdse GUID attribute handling

Signed-off-by: Andrew Bartlett <abartlet@samba.org>

kcc: Give a better error message when samdb_ntds_objectGUID fails

Signed-off-by: Andrew Bartlett <abartlet@samba.org>

dsdb: Unify samdb_{get,set}_ntds_{objectGUID,invocation_id}

The new unified versions have better debugging and ensure
that both functions continue to have the same control flow.

Signed-off-by: Andrew Bartlett <abartlet@samba.org>

WHATSNEW: Add the removal of the web server

Signed-off-by: Garming Sam <garming@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Garming Sam <garming@samba.org>
Autobuild-Date(master): Thu Mar  7 03:17:52 UTC 2019 on sn-devel-144

paged_results: Remove C++ commment and unneeded TODO

Signed-off-by: Garming Sam <garming@catalyst.net.nz>
Reviewed-by: Tim Beale <timbeale@catalyst.net.nz>

passdb: Increase ABI version to 0.28.0

The change from c906153cc lib: Remove some unused code
removed functions, but only updated the minor version
of the ABI. Update the passdb version to 0.28.0
to reflect this change.
file.

Signed-off-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Thu Mar  7 01:30:49 UTC 2019 on sn-devel-144

web_server: Remove the web port smb.conf parameter

With the removal of the web server, there are not any users of this
parameter and so should just be removed.

Signed-off-by: Garming Sam <garming@catalyst.net.nz>
Reviewed-by: Jeremy Allison <jra@samba.org>

web_server: Remove the unused Python WSGI web server

SWAT was removed in Samba 4.1 and there isn't any reason to keep a web
server in our codebase. The web server was not turned on by default.

The web server plainly does not hold up to modern web server standards
and allows for resource exhaustion (and probably generally has bugs).
Credit goes to Michael Hanselmann for prompting us to remove this
service entirely.

Signed-off-by: Garming Sam <garming@catalyst.net.nz>
Reviewed-by: Jeremy Allison <jra@samba.org>

sam.c: allocate account_sid on tmp_ctx

Signed-off-by: Isaac Boukris <iboukris@gmail.com>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Wed Mar  6 04:30:22 UTC 2019 on sn-devel-144

sam.c: fix incorrect check of talloc_new() allocation

Signed-off-by: Isaac Boukris <iboukris@gmail.com>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>

ndr_spoolss_buf: fix out of scope use of stack variable in NDR_SPOOLSS_PUSH_ENUM_OUT()

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13818

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: David Disseldorp <ddiss@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Björn Jacke <bjacke@samba.org>
Reviewed-by: Günther Deschner <gd@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>

Enable make test even without lmdb

Bug: https://bugzilla.samba.org/show_bug.cgi?id=13630

Signed-off-by: Mathieu Parent <math.parent@gmail.com>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Garming Sam <garming@catalyst.net.nz>

lib/winbind_util: Add winbind_xid_to_sid for --without-winbind

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13813

Signed-off-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Wed Mar  6 01:53:16 UTC 2019 on sn-devel-144

lib/winbind_util: Remove winbind_[gu]id_to_sid

Commit c906153cc7 removed these functions, now also remove them for the
--without-winbind case.

Signed-off-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

lib/winbind_util: Move include out of ifdef

This fixes compile errors about missing prototypes with
--picky-developer and --without-winbind

Signed-off-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

dns_hub: Add some debug as to what DNS proxying is happening

This should make it clear at run-time how dns_hub is actually proxying
DNS requests, which will hopefully aid in debugging problems (i.e.
forgetting to add a mapping when adding a new DNS realm).

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Wed Mar  6 00:48:43 UTC 2019 on sn-devel-144

dns_hub: Minor variable rename

We've dropped the iface logic now - this dictionary maps from
realm-to-IP.

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

selftest: Map realm to IP address (instead of iface)

The code is more readable if the hashmap translates between realm and
DC-name, rather than realm-to-iface. We already have a function to map
between DC-name and iface (and since we're doing this, we might as well
map straight to IP address).

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

selftest: Pass realm-to-IP mapping to dns_hub as an argument

Instead of storing hashmaps in 2 different files, we can just convert a
perl hashmap into a string, pass it to dns_hub, and convert it back into
a python dictionary.

The main reason for doing this is the IP-to-testenv mapping now all
lives in a single file (Samba.pm). All this logic is right next to each
other rather than being split across multiple files. Hopefully this will
make it easier to keep it up to date as we add new testenvs.

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

selftest: Split out dns_hub's testenv realm-to-IP logic

Add a separate helper function, as the realm-to-IPv4-addr logic is
fairly self-contained.

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

selftest: Try to tie dns_hub IP mapping to Samba.pm better

dns_hub.py maps the testenv realm to an IP and Samba.pm maps the testenv
NetBIOS name to an IP. We need to keep the two places consistent, as we
add or remove testenvs.

This patch changes dns_hub.py so that it uses a similar hashmap to
Samba.pm. We now have a hashmap with the same name in 2 different
places, so hopefully that's easier to tie them together and keep them in
sync.

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

selftest: Cleanup Samba.pm iface mapping

It looks a bit cleaner if we declare the hash-map in one go, rather than
adding each entry one at a time. Also added a comment explaining what
the hash-map is for, and fixed up tab vs spaces inconsistencies.

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

selftest: Avoid hard-coding client IP address

We implicitly assume the client IP used by selftest is always
127.0.0.11. Add an iface entry for the client to make this a little more
explicit.

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

selftest: dns_hub doesn't need to store $swiface

dns_hub doesn't need to store $ctx->{swiface}. Other testenvs store this
and export it as SOCKET_WRAPPER_DEFAULT_IFACE (i.e. for the tests to
use), but dns_hub doesn't need to do this.

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

selftest: Add helper functions to get IP addresses

Let's centralize these assumptions in one place.

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

tests: Reduce likelihood of auth_log test locking up during CI

We would sometimes see the auth_log test hang during a CI run. The CI
job would eventually fail after consuming a costly 10 hours of CI
runtime.

We believe the problem is around the test creating multiple instances of
the Messaging() context. This is a similar race condition to what was
seen in 19f34b2161dee26.

Currently a new Messaging() context is created for every test case. By
using classmethods instead, the Messaging context is only created once
per python test file execution (i.e. creation of the python class,
rather than initialization of the python object, which happens for every
test-case).

This means the test will only create one Messaging() context, which
should avoid any race conditions.

Changes:
+ removed msg_ctxs - this wasn't actually used for anything.
+ use classmethods to setup and tear-down the Messaging() context (and
tweak lp initialization accordingly).
+ fix discardMessages() - the loop wasn't actually discarding any
messages previously (this may also have been the cause of the test
hanging).

Signed-off-by: Aaron Haslett <aaronhaslett@catalyst.net.nz>
Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Noel Power <npower@samba.org>
Autobuild-User(master): Noel Power <npower@samba.org>
Autobuild-Date(master): Tue Mar  5 13:10:43 UTC 2019 on sn-devel-144

pidl/Python: initialise a datablob

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Mon Mar  4 22:41:01 UTC 2019 on sn-devel-144

dsdb pytsts: reduce scale of subtree rename speed test

The speed test, when it was introduced a few patches ago, was
deliberately slow so that we could see how much better the changes
were. It used 500 users, 50 groups, and 27 computers.

Before the changes, it took this long:

rename ou took 64.373s
rename group took 0.160s
rename user took 0.004s
rename computer took 0.123s

After using the sorted links, it took this long:

rename ou took 12.984s
rename group took 0.161s
rename user took 0.004s
rename computer took 0.122s

And with the final patch to stop the linear search early on success:

rename ou took 11.680s
rename group took 0.089s
rename user took 0.004s
rename computer took 0.128s

"rename ou" is the one we were aiming at. Now that we have done that,
we reduce the size of the test so as not to slow down everyone's
autobuilds.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

dsdb/linked_attributes: shortcut exit for backlink fix

In most cases there can only be one link for each GUID. If we assume
that is true, we can skip half the search, on average.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

dsdb/linked_attributes: improve formatting in some places

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

dsdb/linked_attributes: initialise more pointers to NULL

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

dsdb linked attributes: fix forward links faster

Rename operations can be very slow in large database with many group
memberships, because the linked attributes need to be found and
rewritten for each moved object and the way we did that was naive.

For a while now Samba has kept forward links in sorted order, so
finding group memberships can be an O(log n) rather than O(n)
operation. This patch makes use of that.

The backlinks are not sorted, nor are forward links in old databases,
so we have to use a linear search in those cases.

There is a little bit of extra work to handle the few kinds of forward
links (e.g. msDS-RevealedUsers) that have DN+Binary values.

Tim and Garming came up with the basic idea and a prototype.

Pair-programmed-with: Tim Beale <timbeale@catalyst.net.nz>
Pair-programmed-with: Garming Sam <garming@catalyst.net.nz>
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

dsdb: linked_attributes module knows about sorted links

Until now the linked attrbutes module has allocated its private data
on a per transaction basis, but we prefer to check the sorted links
feature less often than that. So the private data struct is given
module life time and a transaction member to carry out the old role.

In coming patches, the sorted links flag will be used.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

dsdb:replmd: add compatible feature helper function

repl_meta_data.c uses the compatible features attribute of the
"@SAMBA_DSDB" special object to record that linked attributes are
being stored in the database in a sorted order. Soon the
linked_attributes module is going to want to know the same thing, and
in time other modules will want to know about other compatible
features, so we introduce a helper function.

Error checking is slightly improved.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

dsdb/pytests: sanity checks for links under subtree renames

These tests will ensure that linked attributes continue to be handled
correctly under forthcoming changes. The la_move_ou_tree_big() test
will show that the changes make this much faster, after which it can
perhaps be removed.

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

replmd/la: disambiguate error messages a bit

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

dsdb/group_audit: use common get_parsed_dns_trusted()

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

dsdb:util_links: count el->values with unsigned int

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

tests/rodc_rwdc: p.communicate() gives bytes, not str

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

dns_hub: use python 3 shebang

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4/auth/krb: fix spelling of entries

Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:torture: Make sure we do not create a shadow 'struct params'

Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

tests: Work auth_log CLIENT_IP out from config instead of env var

Instead of passing the CLIENT_IP to the auth_log tests, we can just
work out the source-IP that the client will use from its smb.conf file.

This only works for auth_log_pass_change, but not auth_log.py - the
latter still needs to be run on the :local testenv for other reasons, so
it doesn't use the client.conf. However, we can still update the base
code to use the client.conf IP, as auth_log.py overrides
self.remoteAddress anyway.

The main advantage of this change is it avoids having hardcoded IP
addresses in the selftest framework.

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

tests: Work audit_log CLIENT_IP out from config instead of env var

Instead of passing the CLIENT_IP to the audit_log tests, we can just
work out the source-IP that the client will use from its smb.conf file.
Because the audit_log tests are all run on the non-local testenv,
they'll already use the client.conf and the 127.0.0.11 address.

The main advantage of this change is it avoids having hardcoded IP
addresses in the selftest framework.

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

tests: Remove explicit SOCKET_WRAPPER usage from auth_log tests

The auth-logging tests are an odd combination of server and client
behaviour. On the one hand we want a IRPC connection to see the auth
events being logged on the server. On the other hand, we want the auth
events to appear to be happening on a client. Currently we hardcode in
the use of a SOCKET_WRAPPER interface to make this happen.

We can avoid this explicit socket wrapper usage by using the server
smb.conf instead in the one place we actually want to act like the
server (creating the IRPC connection). Then we can switch from using
the 'ad_dc*:local' testenvs to use 'ad_dc*', in order to act like a
client by default. The SERVERCONFFILE environment variable has already
been added for the few cases where a test needs explicit access to the
server's smb.conf.

However, for samba.tests.auth_log, the samlogon test cases are still
reliant on being run on the :local testenv, and so we can't switch them
over just yet. This is because the samlogon is using the DC's machine
creds underneath, which will fail on the non-local testenv. We could
create separate machine creds for the client and use those, but this is
a non-trivial rework of the test code.

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:tests: Remove unused DC_ENV variable

I believe this was a leftover remnant from an earlier patch revision -
it's now been replaced by the DC_SERVERCONFFILE variable.

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:tests: Move duplicated test cases into loop

This is more consistent with how we run tests elsewhere.

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

s4:tests: Avoid passing unnecessary env variables to auth_log tests

These tests all use the ncalrpc connection, so they're always testing a
connection that's local to the server-side. Therefore passing in the
CLIENT_IP and SOCKET_WRAPPER_DEFAULT_IFACE variables (in order to try to
simulate a client connecting) is unnecessary.

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

tests: Remove redundant credentials from auth_log tests

The LDB connection in these tests is to the direct sam.ldb file on disk,
so the credentials are not actually needed (and in fact, weren't event
initialized correctly). These tests always need to run on the DC itself
(i.e. :local testenv) because they use ncalrpc connections.

Signed-off-by: Tim Beale <timbeale@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

tests: add a simple test for smbcacls -x

Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Mon Mar  4 19:11:06 UTC 2019 on sn-devel-144

smbcacls: add -x argument, prints maximum access

Signed-off-by: Ralph Boehme <slow@samba.org>

s3:libsmb: add cli_query_mxac()

Works only for SMB2.

Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

s3:libsmb: add cli_smb2_query_mxac()

Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>