ldb-samba: Handle generic mdb:// url scheme in ldb_relative_path()

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

ldb: Release ldb 1.4.0

* New LMDB backend
* Comprehensive tests for index behaviour
* Enforce transactions for writes
* Enforce read lock use for all reads

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

ldb: Add MDB support to ldb://

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Signed-off-by: Garming Sam <garming@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

TODO: ldb_mdb/tests: add tests for multiple opens across forks

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
TODO: DO we want this for tdb too???

ldb_mdb/tests: test large index key value

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

ldb_mdb: Remove implicit read lock and remove transaction counter

The way to know if we are in a transaction is if there is a non-NULL
transaction handle.

Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>

ldb_mdb: prevent MDB_env reuse across forks

MDB_env's may not be reused accross forks.  Check the pid that the lmdb
structure was created by, and return an error if it is being used by a
different process.

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

ldb_mdb: handle EBADE from mdb_env_open

Under some circumstances mdb_env_open returns EBADE, we treat this as
indicating the file is not a valid lmdb format file.

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

ldb_mdb: Tests for wrap open

Tests to ensure that the mdb_env wrapping code correctly handles
multiple ldb's point to the same physical database file.

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

ldb_mdb: Wrap mdb_env_open

Wrap mdb_env_open to ensure that we only have one MDB_env opened per
database in each process

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

TODO COMMIT MESSAGE ldb_mdb: Apply LMDB key length restrictions at key-value layer

We need to enforce the GUID index mode so end-users do not get a supprise
in mid-operation and we enforce a max key length of 511 so that the
index key trunctation is done correctly.

Otherwise the DB will appear to work until a very long key (DN or index)
is used, after which it will be sad.

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

sq ldb_mdb/tests: Run api and index test also on lmdb LIKE THIS? SQ??? WHY not the same dict for all tests?  ldb: Allow tests to operate on ldb_mdb after new restrictions

SQ??? ldb_mdb/tests: Run api and index test also on lmdb WHY not the same dict for all tests?  ldb: Allow tests to operate on ldb_mdb after new restrictions

The tests were working fine on ldb_mdb but only in the mode without
the GUID index.  This mode is not OK for production so has been banned
but this means we need to rework the tests to set an objectGUID on
each record.

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

ldb_mdb/tests: Run api and index test also on lmdb

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

TODO: ldb_mdb/tests: Add tests to check for max key length and DB size

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
TODO: is "# we don't want to run ldb_lmdb_size_test" correct?

ldb_mdb: Don't allow modify operations on a read only db

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

ldb_mdb: Store pid to change destructor on fork

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

ldb_mdb: Enable LDB_FLG_NOSYNC in ldb_mdb

This is used in selftest with 'ldb:nosync = true'.

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

lib/ldb/tools/ldbdump.c #include <lmdb.h>

sq lib/ldb/ldb_mdb/ldb_mdb.[ch] include <lmdb.h>

sq ldb_mdb: Implement the lmdb backend for ldb

SQ ldb_mdb: make backends expose if there is an active transaction

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Signed-off-by: Andrew Bartlett <abartlet@samba.org>

METZE01 ldb_mdb: Implement the lmdb backend for ldb

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

FAILS: TODO: ldb/tests: add tests for transaction_start/lock_read across forks

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
TODO: this should work also with tdb, we need to have the same external
behavior of tdb and mdb, everything else will cause problems!

TODO: I think we should also protect against a fork between
transaction_start and transaction_[prepare]_commit.

TODO commit message ldb: Allow tests to operate on ldb_mdb after new restrictions

The tests were working fine on ldb_mdb but only in the mode without
the GUID index.  This mode is not OK for production so has been banned
but this means we need to rework the tests to set an objectGUID on
each record.

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

tests/dlz_bind9: support for multiple db types by using ldb://

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

tests: Replace some references to tdb with ldb://

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

ldb-samba: Handle generic ldb:// url scheme in ldb_relative_path()

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

ldb: Introduce new generic ldb:// prefix to allow backend autodetection

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Signed-off-by: Garming Sam <garming@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

dsdb:schema_load: make use of ldb_relative_path() in partition_metadata_open()

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

dsdb:partition_metadata: make use of ldb_relative_path() in partition_metadata_open()

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

ldb tests: add cmocka tests of kv operation interactions with transactions

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

ldb tests: api ensure database correctly populated

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>

ldb_tdb: Disallow reads without a transaction or read lock

This will ensure we match LMDB behaviour and avoid a repeat of the per-record locking
issues (compared with full DB locking) we had before Samba 4.7.

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

ldb_tdb: Disallow TDB nested transactions and use tdb_transaction_active()

This avoids keeping a counter, which can be error-prone.

Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>

ldb_tdb: Do not make search or DB modifications without a lock

The ldb_cache startup code would previously not take a read lock
nor a sufficiently wide write transaction.

The new code takes a read lock, and if it needs to write takes a
write lock (transaction) and re-reads before continuing.

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Signed-off-by: Andrew Bartlett <abartlet@samba.org>

ldb: make key/value backends expose if there is an active transaction

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Signed-off-by: Andrew Bartlett <abartlet@samba.org>

ldb_tdb: Make sure max_key_length != 0 requires a GUID index mode

We need to enforce the GUID index mode so end-users do not get a supprise
in mid-operation and we enforce a max key length of 511 so that the
index key trunctation is done correctly.

Otherwise the DB will appear to work until a very long key (DN or index)
is used, after which it will be sad.

Signed-off-by: Gary Lockyer <gary@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>

ldb/tests: don't use TEVENT_DEPRECATED in ldb_kv_ops_test.c

Signed-off-by: Stefan Metzmacher <metze@samba.org>

torture: Test compound request request counters

This will send an unfixed smbd into the

SMB_ASSERT(op->request_count > 0);

in smbd_smb2_request_reply_update_counts

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13215
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Thu Apr 12 14:38:39 CEST 2018 on sn-devel-144

s3:smb2_server: correctly maintain request counters for compound requests

If a session expires during a compound request chain,
we exit smbd_smb2_request_dispatch() with
'return smbd_smb2_request_error(req, ...)' before
calling smbd_smb2_request_dispatch_update_counts().

As req->request_counters_updated was only reset
within smbd_smb2_request_dispatch_update_counts(),
smbd_smb2_request_reply_update_counts() was called
twice on the same request, which triggers
SMB_ASSERT(op->request_count > 0);

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13215

Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>

gpo: python chardet is not a dep of samba

Signed-off-by: David Mulder <dmulder@suse.com>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Douglas Bagnall <dbagnall@samba.org>
Autobuild-Date(master): Thu Apr 12 11:27:03 CEST 2018 on sn-devel-144

libgpo: gpo_copy_file() shouldn't explicitly call smb1

Don't call cli_openx directly to open a file this
calls smb1 code explicitly, which fails if we did
a multi-protocol negotiate and negotiated smb2+.
Use the higher level cli_open() instead.

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

param: Add python binding for lpcfg_cache_path

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

gpo: Create a gp_sec_ext module

Move the gp_sec_ext into it's own module,
which is how new gp_ext's will be created.

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

gpo: Move implementation from samba_gpoupdate

The implementation of group policy apply should
not be in the application script. One reason is
to implement user apply, we can call these
functions via the python c-api, (passing creds
via the command line will expose them via ps).
Another reason for this is if some overrides
the smb.conf "gpo update command" option, it
would be useful to have these functions.

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

gpo: Create base class gp_inf_ext

Abstract the process of reading from an ini file,
since other extensions will be reading gpos this
way.

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

gpo: Move the file parse function to gp_ext

A file will always be read from the sysvol the
same way, but the data will be read differently.
This patch moves the parse function to gp_ext,
and requires subclasses to implement the read()
function to interpret the data.

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

gpo: Move gp_sec_ext __init__ to base class

For this class to be extensible, the constructor
should be available to subclasses.

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

gpo: Rename the inf_to class to gp_ext_setter

This class will be subclassed and used for more
than just inf settings application.

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

Check "auth event notification" param in log_json

When Samba is compiled with JSON support it logs an error message
at log level 3 by default for every authentication event:

get_auth_event_server: Failed to find 'auth_event' registered on the
message bus to send JSON authentication events to:
NT_STATUS_OBJECT_NAME_NOT_FOUND

This patch adjusts the log_json function to consider the smb.conf
parameter "auth event notification" before calling auth_message_send.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13381

Signed-off-by: Arvid Requate <requate@univention.de>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Gary Lockyer <gary@catalyst.net.nz>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Thu Apr 12 08:12:39 CEST 2018 on sn-devel-144

ldb-samba: require pid match for cached ldb

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

ldb_wrap: Remove the magic cache of database handles except for sam.ldb

sam.ldb is handled in samdb_connect_url(), not this function.

This cache caused issues when "private dir" was changed in a testing script, but also
just generates many-owner shared mutable state that is frowned upon these days.

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

dsdb: Use talloc_get_type_abort() in schema_load_{start,end}_transaction

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13379

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

dsdb: ensure we take out a read lock during the dsdb_init

We have to also take it out in the partitions code when we load the
partition backends.

This ensures that the init handlers hold a whole-db lock just as the
search code does.

To ensure the locking count in schema_load is balanced, the
private data is now created in the first lock_read() call.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13379

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

dsdb: Rework schema reload during the read lock

Rather than refusing the reload based on making cached sequence numbers match
just load it once at the time the DB is globally locked, if required.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13379

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

dsdb: Load schema during the read_lock() hook, not the search

This should trigger slightly less often and is the more correct place, as
we only load it during the first lock when not in a transaction.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13379

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

dsdb: Allow search before init() is called in extended_dn_out

This matches the earlier check of p && p->normalise.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13379

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

dsdb: Move ldb_set_default_dns() into rootdse_get_private_data()

This call needs to be done at the very first chance, in this case
during the first call to the lock_read() hook, otherwise the
schema_data module can't find the schema.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13379

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

dsdb: Create rootdse_get_private_data()

This will get the private data on the first call, allowing that not to be
the init() hook.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13379

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

selftest: Make a transaction before @INDEXLIST etc is checked in dsdb_schema_attributes.py

This helps us remove the write to the database from the (soon to be
read locked) init code.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13379

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

dsdb: Wait until a transaction starts to call dsdb_schema_set_indices_and_attributes()

This avoids starting a transaction in schema_load_init() and allows it
to operate with a read lock held, which will avoid locking issues
(deadlock detected due to lock odering if we do not have a global
read lock).

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13379

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

dsdb: Allow search before init() call in encrypted_secrets

Simply do not decrypt anything until the init call is run.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13379

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

dsdb: Do not create a transaction in partition_init()

This will allow us to lock the databases for read during all of the Samba init
hooks.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13379

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

dsdb: Ensure to cancel the transaction if we fail to save the prefixMap

This rare error case forgot to call ldb_transaction_cancel()

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

ldb_wrap: Remove ldb_transaction_cancel_noerr from ldb_wrap_fork_hook()

Writing to a TDB, without locks (these are per-process) in a forked child is never going to
end well, if a transaction is open at this point we have bigger problems.

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

dsdb: check for dSHeuristics more carefully

This check would pass if the dSHeuristics was treated as always being
000000000 for searches which is not enough, we must check for a value
of 000000001 (userPassword enabled).

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13378

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

dsdb: Check for userPassword support after loading the databases

The net result of this is only that userPassword values (which were
world readable when set) would still be visible after userPassword
started setting the main DB password.

In AD, those values become hidden once the dSHeuristics bit is set,
but Samba lost that when fixing a performance issue with
f26a2845bd42e580ddeaf0eecc9b46b823a0c6bc

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13378

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

s3: tests: Regression test to ensure we can never return a DIRECTORY attribute on a stream.

Tests streams_xattr and also streams_depot.

Inspired from a real-world test case by Andrew Walker <awalker@ixsystems.com>.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13380

Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Böhme <slow@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Thu Apr 12 02:04:28 CEST 2018 on sn-devel-144

s3: smbd. Generic fix for incorrect reporting of stream dos attributes on a directory

According to MS-FSA a stream name does not have
separate DOS attribute metadata, so we must return
the DOS attribute from the base filename. With one caveat,
a non-default stream name can never be a directory.

As this is common to all streams data stores, we handle
it here instead of inside all stream VFS modules.

Otherwise identical logic would have to be added to
all streams modules in their [f]get_dos_attribute_fn()
VFS calls.

Found in real-world use case by Andrew Walker <awalker@ixsystems.com>.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13380

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

s3: vfs: vfs_streams_xattr: Don't blindly re-use the base file mode bits.

When returning the stat struct for an xattr stream,
we originally base the st_ex_mode field on the value
from the base file containing the xattr. If the base
file is a directory, it will have S_IFDIR set in st_ex_mode,
but streams can never be directories, they must be reported
as regular files.

The original code OR'ed in S_IFREG, but neglected to
AND out S_IFDIR.

Note this is not a complete to fix bug 13380 as
it doesn't fix the generic case with all streams
modules. See later fix and regression test.

Found in real-world use case by Andrew Walker <awalker@ixsystems.com>.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13380

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

credentials: Fix CID 1414796 Explicit null dereferenced

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): Wed Apr 11 21:58:00 CEST 2018 on sn-devel-144

credentials: Fix line length

... just because I'll modify that line in the next commit

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

credentials: Revert "credentials: Fix CID 1414796 Explicit null dereferenced"

This reverts commit 90c02ec64d0e3c860f8d6906cf849bdd2c7bcc54.

We have code to take care of password==NULL, this CID must be fixed in a
different way

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

smbd: Fix CID 1414783 Double unlock

The loop is unnecessary, both susv4 as well as the Linux manpage
explicitly say:

> These functions shall not return an error code of [EINTR].

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

dnsrpc: Use TALLOC_FREE instead of an explicit if-statement

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

winbind: Fix CID 1427625 Calling risky function

Probably not really a problem, but we have generate_random(), so why not
use it?

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

vfs_fruit: Fix CID 1416474 Dereference null return value

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

tevent: Fix CID 1414792 Unchecked return value

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

rpc_server: Init local_server_* in make_internal_rpc_pipe_socketpair

Bug: https://bugzilla.samba.org/show_bug.cgi?id=13370
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Wed Apr 11 15:19:19 CEST 2018 on sn-devel-144

lib/util: Call log_stack_trace() in smb_panic_default()

This matches the AD DC with the behaviour in smbd.

Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Wed Apr 11 04:03:07 CEST 2018 on sn-devel-144

lib/util: Move log_stack_trace() to common code

Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

lib/util: Log PANIC before calling pacic action just like s3

This is like the changes made in s3 by
4fa555980070d78b39711ef21d77628d26055bc2

Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

s3-lib: Remove support for libexc for IRIX backtraces

IRIX is long dead, and this code needs become_root() which is not in
the top level code.

Additionally, the check for libexc never made it into waf, so this
has been dead code since Samba 4.1.

Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

libsmb: Pass "account_name/flags" through nb_getdc

Don't hardcode values that we might want to change later

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

libsmb: Pass "account_name/flags" through prep_getdc_request

Don't hardcode values that we might want to change later

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

libsmb: Introduce a helper variable

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

libsmb: Give dsgetdcname.c its own header

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

libsmb: Give namequery.c its own header

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

libsmb: Remove unused trustdom_cache.c

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

nmbd: Fix a typo

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

nmbd: Fix a typo

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

tdbtool: Use tdb_wipe_all in "erase" command

This is a lot quicker on large, fragmented databases. tdb_delete can
leave the freelist in a fragmented mess.

Also, it's a lot more robust: I've got a 4GB tdb file that was affected
by the problem fixed with c7211882a79. These databases have large space
at the end that is not part of any record or freelist
entry. tdb_wipe_all converts this space into a freelist entry. One
downside is that with those broken databases (which should not happen
after c7211882a79) have unallocated blocks in their file range after
this operation.

I think the speed advantage outweighs this disadvantage.

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

s4:ldb/password_hash.c: improve krb5 context error message

When heimdal encounters a MIT krb5.conf that it does not understand,
it would emit an "ldb operations error". Sadly this does not help
or communicate to the administrator the root cause of the issue.

Improve the error message for when krb init fails during password_hash.c

Signed-off-by: William Brown <william@blackhats.net.au>
Reviewed-by: Alexander Bokovoy <ab@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>

s3: smbd: Unix extensions attempts to change wrong field in fchown call.

Cut and paste error.

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13375

Reported-by: Rungta, Vandana <vrungta@amazon.com>
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: David Disseldorp <ddiss@samba.org>
Autobuild-User(master): David Disseldorp <ddiss@samba.org>
Autobuild-Date(master): Tue Apr 10 00:45:56 CEST 2018 on sn-devel-144

s3: smbd: Fix memory leak in vfswrap_getwd()

BUG: https://bugzilla.samba.org/show_bug.cgi?id=13372

Signed-off-by: Andrew Walker <awalker@ixsystems.com>.
Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Mon Apr  9 21:48:12 CEST 2018 on sn-devel-144

s3:smb2_tcon: Add check to prevent non-DFS clients from connecting to an msdfs proxy.

Signed-off-by: Ralph Wuerthner <ralph.wuerthner@de.ibm.com>
Reviewed-by: David Disseldorp <ddiss@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Sat Apr  7 05:05:22 CEST 2018 on sn-devel-144

pdb_nds: Fix CID 1273401 Unused value

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

idmap: Fix CID 1363261 Resource leak

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

idmap: Use TALLOC_FREE instead of explicit code

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

s4/torture: spelling fix in vfs_fruit test

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