Volker Lendecke [Wed, 21 Nov 2018 13:55:10 +0000 (14:55 +0100)]
krb5_wrap: Fix a typo
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 Nov 28 21:15:31 CET 2018 on sn-devel-144
Volker Lendecke [Wed, 21 Nov 2018 14:28:42 +0000 (15:28 +0100)]
credentials: Fix set_ccache with empty creds cache
This is an extension of bb2f7e3aee7e9b8: Without this fix in the
"empty ccache" case we never set cred->ccache, so the whole call to
cli_credentials_set_ccache became pointless
Signed-off-by: Volker Lendecke <vl@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org>
Ralph Boehme [Tue, 20 Nov 2018 14:50:52 +0000 (15:50 +0100)]
tfork: add a README how to run test torture test under valgrind
Signed-off-by: Ralph Boehme <slow@samba.org> Reviewed-by: Stefan Metzmacher <metze@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Wed Nov 28 15:57:43 CET 2018 on sn-devel-144
Ralph Boehme [Tue, 20 Nov 2018 15:03:03 +0000 (16:03 +0100)]
tfork: add a suppresssions file for drd
drd reports:
initialized twice: cond 0x514f188
at 0x4C3A399: pthread_cond_init_intercept (drd_pthread_intercepts.c:1022)
by 0x4C3A399: pthread_cond_init@* (drd_pthread_intercepts.c:1030)
by 0x50F3FF3: tfork_atfork_child (tfork.c:250)
by 0x9A4B95D: fork (fork.c:204)
by 0x50F4834: tfork_start_waiter_and_worker (tfork.c:581)
by 0x50F4CDB: tfork_create (tfork.c:780)
by 0x2F7469: tfork_thread (tfork.c:431)
by 0x4C358F8: vgDrd_thread_wrapper (drd_pthread_intercepts.c:444)
by 0x8D46593: start_thread (pthread_create.c:463)
by 0x9A7EE6E: clone (clone.S:95)
cond 0x514f188 was first observed at:
at 0x4C3A399: pthread_cond_init_intercept (drd_pthread_intercepts.c:1022)
by 0x4C3A399: pthread_cond_init@* (drd_pthread_intercepts.c:1030)
by 0x50F413A: tfork_global_initialize (tfork.c:287)
by 0x8D4DEA6: __pthread_once_slow (pthread_once.c:116)
by 0x4C377FD: pthread_once_intercept (drd_pthread_intercepts.c:800)
by 0x4C377FD: pthread_once (drd_pthread_intercepts.c:806)
by 0x50F4C0E: tfork_create (tfork.c:743)
by 0x2F7469: tfork_thread (tfork.c:431)
by 0x4C358F8: vgDrd_thread_wrapper (drd_pthread_intercepts.c:444)
by 0x8D46593: start_thread (pthread_create.c:463)
by 0x9A7EE6E: clone (clone.S:95)
This is intentional, the reinit is in a child process. Cf the comment in
tfork.c.
Signed-off-by: Ralph Boehme <slow@samba.org> Reviewed-by: Stefan Metzmacher <metze@samba.org>
Ralph Boehme [Mon, 19 Nov 2018 14:18:34 +0000 (15:18 +0100)]
tfork: add a suppresssions file for helgrind
tfork_atexit_unknown[1|2]:
No idea what triggers this, definitely not tfork itself.
tfork_pthread_get_specific:
Helgrind reports:
Possible data race during read of size 4 at 0x5141304 by thread #3
Locks held: none
at 0x50E602E: tfork_global_get (tfork.c:301)
by 0x50E69B1: tfork_create (tfork.c:737)
by 0x2F7419: tfork_thread (tfork.c:431)
by 0x4C35AC5: mythread_wrapper (hg_intercepts.c:389)
by 0x8D38593: start_thread (pthread_create.c:463)
by 0x9A70E6E: clone (clone.S:95)
This conflicts with a previous write of size 4 by thread #2
Locks held: none
at 0x8D3F7B7: pthread_key_create (pthread_key_create.c:41)
by 0x50E5F79: tfork_global_initialize (tfork.c:280)
by 0x8D3FEA6: __pthread_once_slow (pthread_once.c:116)
by 0x50E6999: tfork_create (tfork.c:728)
by 0x2F7419: tfork_thread (tfork.c:431)
by 0x4C35AC5: mythread_wrapper (hg_intercepts.c:389)
by 0x8D38593: start_thread (pthread_create.c:463)
by 0x9A70E6E: clone (clone.S:95)
Location 0x5141304 is 0 bytes inside global var "tfork_global_key"
declared at tfork.c:122
This is nonsense, tfork_global_get() calls pthread_getspecific, so
we're looking at the pthread_key_create()/pthread_[g|s]etspecific()
API here which works with threads by design.
Signed-off-by: Ralph Boehme <slow@samba.org> Reviewed-by: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Tim Beale <timbeale@catalyst.net.nz> Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Karolin Seeger <kseeger@samba.org>
Autobuild-Date(master): Wed Nov 28 11:31:14 CET 2018 on sn-devel-144
Tim Beale [Mon, 12 Nov 2018 23:24:16 +0000 (12:24 +1300)]
CVE-2018-16857 dsdb/util: Correctly treat lockOutObservationWindow as 64-bit int
Commit 442a38c918ae1666b35 refactored some code into a new
get_lockout_observation_window() function. However, in moving the code,
an ldb_msg_find_attr_as_int64() inadvertently got converted to a
ldb_msg_find_attr_as_int().
ldb_msg_find_attr_as_int() will only work for values up to -2147483648
(about 3.5 minutes in MS timestamp form). Unfortunately, the automated
tests used a low enough timeout that they still worked, however,
password lockout would not work with the Samba default settings.
Tim Beale [Mon, 12 Nov 2018 22:49:56 +0000 (11:49 +1300)]
CVE-2018-16857 tests: Sanity-check password lockout works with default values
Sanity-check that when we use the default lockOutObservationWindow that
user lockout actually works.
The easiest way to do this is to reuse the _test_login_lockout()
test-case, but stop at the point where we wait for the lockout duration
to expire (because we don't want the test to wait 30 mins).
This highlights a problem currently where the default values don't work.
Gary Lockyer [Wed, 7 Nov 2018 02:08:04 +0000 (15:08 +1300)]
CVE-2018-16852 dcerpc dnsserver: refactor common properties handling
dnsserver_common.c and dnsutils.c both share similar code to process
zone properties. This patch extracts the common code and moves it to
dnsserver_common.c.
Gary Lockyer [Mon, 5 Nov 2018 23:16:30 +0000 (12:16 +1300)]
CVE-2018-16852 dcerpc dnsserver: Ensure properties are handled correctly
Fixes for
Bug 13669 - (CVE-2018-16852) NULL
pointer de-reference in Samba AD DC DNS management
The presence of the ZONE_MASTER_SERVERS property or the
ZONE_SCAVENGING_SERVERS property in a zone record causes the server to
follow a null pointer and terminate.
Tests to verify
Bug 13669 - (CVE-2018-16852) NULL
pointer de-reference in Samba AD DC DNS management
The presence of the ZONE_MASTER_SERVERS property or the
ZONE_SCAVENGING_SERVERS property in a zone record causes the server to
follow a null pointer and terminate.
Garming Sam [Mon, 5 Nov 2018 03:18:18 +0000 (16:18 +1300)]
CVE-2018-16851 ldap_server: Check ret before manipulating blob
In the case of hitting the talloc ~256MB limit, this causes a crash in
the server.
Note that you would actually need to load >256MB of data into the LDAP.
Although there is some generated/hidden data which would help you reach that
limit (descriptors and RMD blobs).
Andrew Bartlett [Wed, 24 Oct 2018 02:41:28 +0000 (15:41 +1300)]
CVE-2018-16841 selftest: Check for mismatching principal in certficate compared with principal in AS-REQ
BUG: https://bugzilla.samba.org/show_bug.cgi?id=13628 Signed-off-by: Andrew Bartlett <abartlet@samba.org> Reviewed-by: Gary Lockyer <gary@catalyst.net.nz>
Volker Lendecke [Mon, 26 Nov 2018 15:21:16 +0000 (16:21 +0100)]
notifyd: Improve a debug message
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): Tue Nov 27 21:42:24 CET 2018 on sn-devel-144
Volker Lendecke [Sat, 24 Nov 2018 12:25:25 +0000 (13:25 +0100)]
winbind: Use dom_sid_str_buf
Signed-off-by: Volker Lendecke <vl@samba.org> Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Tue Nov 27 10:38:11 CET 2018 on sn-devel-144
Ralph Boehme [Wed, 14 Nov 2018 12:45:11 +0000 (13:45 +0100)]
s4:torture: add a test-suite for VSS
This test will not be run from the main torture test runner in selftest,
as there we don't pass the required arguments 'twrp_file' and
'twrp_snapshot'.
The test needs a carefully prepared environment with provisioned
snapshot data, so the test will be started from a blackbox test
script. That comes next.
Tim Beale [Thu, 22 Nov 2018 03:56:22 +0000 (16:56 +1300)]
tests: Rework backup test inheritance to make LP constraints clearer
The backup tests have a special constraint where we always want to use
check_output() over runcmd(). The reason is we need the samba-tool
backup/restore commands executed in a separate process. Otherwise the
global underlying LoadParm can accumulate settings from earlier test
case runs.
We can avoid someone in future inadvertently running runcmd() by
mistake, by simply changing the inheritance so we no longer inherit from
SambaToolCmdTest (so the runcmd functions are no longer present).
The comment explaining this has been moved to the top of the file.
Note that the TestCaseInTempDir inheritance was redundant.
BlackboxTestCase inherits from TestCaseInTempDir (and SambaToolCmdTest
was inheriting from BlackboxTestCase).
Signed-off-by: Tim Beale <timbeale@catalyst.net.nz> Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Autobuild-User(master): Tim Beale <timbeale@samba.org>
Autobuild-Date(master): Tue Nov 27 06:57:03 CET 2018 on sn-devel-144
Tim Beale [Thu, 22 Nov 2018 20:46:38 +0000 (09:46 +1300)]
tests: Work out DOMSID via samdb rather than environs
Not all testenvs have the DOMSID set as an environment variable.
However, it's easy enough to work out from querying the samdb.
This is a slight change in that we use a source4-generated loadparm
to connect to the DB (self.lp is source3-generated, presumably for
some SMB connection dependency).
This change is so we can run the ntacls_backup tests against a DC with
SMBv1 disabled (the restoredc). Note that currently the tests fail in
the smb.SMB() connection in the setUp(), so we can't run them as part
of autobuild just yet (because we can't known-fail test errors).
Tim Beale [Thu, 22 Nov 2018 01:05:01 +0000 (14:05 +1300)]
tests: Run backup tests against restoredc (SMBv1 disabled)
Running the backup tests against the restoredc highlights that the
backup online/rename commands don't work if SMBv1 is disabled. Note that
the offline commands still work because they don't rely on an SMB
connection to the server.
(Note that running the backup tests against the restoredc is probably a
good idea anyway, to prove that there's no limit to the number of times
you can restore a domain from backup, i.e. we support more than just a
one-off restore).
Noel Power [Mon, 5 Nov 2018 19:00:20 +0000 (19:00 +0000)]
python/samba/test: PY3 port samba.tests.domain_backup
The restoredc already runs under python3, so before we can run the
domain_backup tests against the restoredc, we need to make sure they
work under python3.
Tim Beale [Thu, 22 Nov 2018 01:35:58 +0000 (14:35 +1300)]
tests: Handle backup command exceptions as test failures, not errors
If the backup command fails (i.e. throws an exception), we want the test
to fail. This makes it easier to mark tests as 'knownfail' (because we
can't knownfail test errors).
In theory, this should just involve updating run_cmd() to catch any
exceptions from the command and then call self.fail().
However, if the backup command fails, it can leave behind files in the
targetdir. Partly this is intentional, as these files may provide clues
to users as to why the command failed. However, in selftest, it causes
the TestCaseInTempDir._remove_tempdir() assertion to fire. Because this
assert actually gets run as part of the teardown, the assertion gets
treated as an error rather than a failure (and so we can't knownfail the
backup tests). To get around this, we remove any files in the tempdir
prior to calling self.fail().
self.create_backup() uses self.run_cmd(), which is a wrapper around
self.check_output(). Rework the code to call the underlying
check_output() function directly instead.
The reason we're doing this is we want run_cmd() to catch exceptions and
fail the test (i.e. in the next patch). However, we can't do that because
this test case relies on receiving the exceptions.
Tim Beale [Thu, 22 Nov 2018 00:22:19 +0000 (13:22 +1300)]
selftest: Designate one testenv as having SMBv1 disabled
We recommend users disable SMBv1 to avoid potential security holes.
However, none of the AD DC testenvs have SMBv1 disabled.
This patch disables SMBv1 on an arbitrarily-chosen testenv (restoredc).
I chose restoredc as we'll want to run the backup tool tests against
this target, and it might be useful to check we can backup a DC if it's
already been restored once.
Note that SMBv2 doesn't support POSIX extensions (only SMBv1 does),
which is why we haven't just disabled SMBv1 on *all* testenvs.
Tim Beale [Sun, 25 Nov 2018 22:58:31 +0000 (11:58 +1300)]
tests: Use s3 smbclient for testenv smoketests
smbclient4 doesn't support SMBv2 connections, i.e. it won't work against
a DC with SMBv1 disabled. The smoke-test here is that the DC accepts
the connection, so we don't really care what SMB tool we use to connect.
Tim Beale [Mon, 26 Nov 2018 01:28:59 +0000 (14:28 +1300)]
selftest: Make testenv NTVFS usage match --use-ntvfs
Regardless of whether the testenv uses --use-ntvfs as part of its
provision options, the s4 testenvs all default to using the NTVFS file
server.
It's not particularly obvious that this is happening. The new restore
DCS (restoredc, renamedc, labdc) were all using NTVFS unintentionally.
The problem is the s4 testenvs default to using services '-s3fs +smb".
provision_ad_dc() explicitly overrides this to use s3fs again
(technically it ends up with both in its smb.conf and just uses whatever
comes last).
This patch changes the testenv setup to check for the presence of the
'--use-ntvfs' option and to set the 'server services' config option
appropriately. This way, the provision command and the smb.conf options
should always line up, with respect to NTVFS.
Tim Beale [Mon, 26 Nov 2018 01:59:06 +0000 (14:59 +1300)]
selftest: Make chgdcpass's NTVFS usage more obvious
The chgdcpass testenv was not passing --use-ntvfs to the provision
command, but it was still actually using NTVFS.
The reason is the smb.conf generated by provision_raw_step1() would
always try to use the s4/NTVFS file server. Because the smb.conf already
existed, this trumped what was passed to the provision command.
This patch doesn't change the chgdcpass file server. It just makes it
more obvious that chgdcpass is using NTVFS.
Volker Lendecke [Fri, 23 Nov 2018 07:40:57 +0000 (08:40 +0100)]
libads: Align integer types
Signed-off-by: Volker Lendecke <vl@samba.org> Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Fri Nov 23 20:23:57 CET 2018 on sn-devel-144
Tim Beale [Wed, 21 Nov 2018 21:35:01 +0000 (10:35 +1300)]
source4 smbd prefork: Add code comments
Add some comments to the prefork code explaining what's going on.
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): Fri Nov 23 11:40:49 CET 2018 on sn-devel-144
Gary Lockyer [Thu, 13 Sep 2018 21:43:59 +0000 (09:43 +1200)]
source4 messaging: clean up terminated processes
Now that the smbd pre-fork process model restarts failed processes rather than
terminating, we end up with names registered to defunct processes.
This patch adds a function to clean up all the names registered to a process.
Signed-off-by: Gary Lockyer <gary@catalyst.net.nz> Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Gary Lockyer [Tue, 4 Sep 2018 19:31:22 +0000 (07:31 +1200)]
source4 smbd prefork: Add backoff to process restart
Add new smbd.conf variables 'prefork backoff increment' and
'prefork maximum backoff' to control the rate at which failed pre-forked
processes are restarted.
Signed-off-by: Gary Lockyer <gary@catalyst.net.nz> Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Gary Lockyer [Mon, 3 Sep 2018 22:09:38 +0000 (10:09 +1200)]
source4 smbd process: pass the fatal flag to terminate
Pass the fatal flag supplied to task_server_terminate to the process
task_terminate method. It will be used by the task_terminate methods to
set an appropriate exit code. The process_prefork model will use a non
zero exit code to indicate that the process should be restarted.
Signed-off-by: Gary Lockyer <gary@catalyst.net.nz> Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Tim Beale [Tue, 20 Nov 2018 03:02:05 +0000 (16:02 +1300)]
replmd: remove unnecessary indent
The previous refactor now means we return early if we don't need to
re-apply isDeleted to the object. The 'else' is redundant and we can
remove it to avoid unnecessary indent.
This patch is basically just a whitespace change. It should not alter
functionality.
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): Fri Nov 23 08:10:41 CET 2018 on sn-devel-144
Tim Beale [Mon, 19 Nov 2018 22:45:07 +0000 (11:45 +1300)]
replmd: Avoid passing msg to replmd_process_linked_attribute()
We can prevent anyone from inadvertently adding/removing msg->elements[]
in replmd_process_linked_attribute() by just not passing msg into the
function. Currently we only actually need the source DN and a memory
context for reallocating old_el->values.
The warning comment has been moved to a more appropriate place.
Signed-off-by: Tim Beale <timbeale@catalyst.net.nz> Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Tim Beale [Tue, 20 Nov 2018 04:30:37 +0000 (17:30 +1300)]
tests: Fix flappiness in DRS tests due to RID Set changing
The test_link_utdv_hwm test case in getnc_exop has started getting
slightly flappy (8 failures in the last 2 weeks). The problem is the
test case creates a new computer, which can occasionally result in a new
RID pool being allocated.
The problem can be reproduced by running the test case repeatedly (it
usually fails after ~250 times).
This patch updates the _check_ctr6() assertion to filter out the 'CN=RID
Set' object, if it happens to be present.
Signed-off-by: Tim Beale <timbeale@catalyst.net.nz> Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Tim Beale [Tue, 20 Nov 2018 04:15:41 +0000 (17:15 +1300)]
tests: Add better error for DRS tests
We've got a flappy test hitting this assertion failure, but we can't
tell why it's failing intermittently (probably because we're bumping the
RID-Set, but there's no way to confirm this).
Add some extra debug info if the test assertion fails.
Signed-off-by: Tim Beale <timbeale@catalyst.net.nz> Reviewed-by: Andrew Bartlett <abartlet@samba.org>
librpc/ndr/ndr.c:1430 runtime error: left shift of 1 by 31 places cannot
be represented in type 'int'
Signed-off-by: Andreas Schneider <asn@samba.org> Reviewed-by: Gary Lockyer <gary@catalyst.net.nz>
Autobuild-User(master): Gary Lockyer <gary@samba.org>
Autobuild-Date(master): Fri Nov 23 01:23:09 CET 2018 on sn-devel-144
projects / thirdparty / samba.git / log
