IF YOU WOULD LIKE TO GET AN ACCOUNT, please write an
email to Administrator. User accounts are meant only to access repo
and report issues and/or generate pull requests.
This is a purpose-specific Git hosting for
BaseALT
projects. Thank you for your understanding!
Только зарегистрированные пользователи имеют доступ к сервису!
Для получения аккаунта, обратитесь к администратору.
This makes LDAP_DIRSYNC_OBJECT_SECURITY the only behaviour provided by
Samba.
Having a second access control system withing the LDAP stack is unsafe
and this layer is incomplete.
The current system gives all accounts that have been given the
GUID_DRS_GET_CHANGES extended right SYSTEM access. Currently in Samba
this equates to full access to passwords as well as "RODC Filtered
attributes" (often used with confidential attributes).
Rather than attempting to correctly filter for secrets (passwords) and
these filtered attributes, as well as preventing search expressions for
both, we leave this complexity to the acl_read module which has this
facility already well tested.
The implication is that callers will only see and filter by attribute
in DirSync that they could without DirSync.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15424
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
The aim here is to document the expected (even if not implemented)
SEARCH_FLAG_RODC_ATTRIBUTE vs SEARCH_FLAG_CONFIDENTIAL, behaviour, so
that any change once CVE-2023-4154 is fixed can be noted.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15424
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
If the client requested FILE_OVERWRITE[_IF], we're implicitly adding
FILE_WRITE_DATA to the open_access_mask in open_file_ntcreate(), but for the
access check we're using access_mask which doesn't contain the additional
right, which means we can end up truncating a file for which the user has
only read-only access via an SD.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15439
Signed-off-by: Ralph Boehme <slow@samba.org>
We correctly handle this and just return ENOENT (NT_STATUS_OBJECT_NAME_NOT_FOUND).
Remove knowfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15422
Signed-off-by: Jeremy Allison <jra@samba.org>
The raw SMB2-INVALID-PIPENAME test passes against Windows 2022,
as it just returns NT_STATUS_OBJECT_NAME_NOT_FOUND.
Add the knownfail.
BUG:https://bugzilla.samba.org/show_bug.cgi?id=15422
Signed-off-by: Jeremy Allison <jra@samba.org>
In the fd_close() fsp->fsp_flags.fstat_before_close code path.
If this is a stream and delete-on-close was set, the
backing object (an xattr from streams_xattr) might
already be deleted so fstat() fails with
NT_STATUS_NOT_FOUND. So if fsp refers to a stream we
ignore the error and only bail for normal files where
an fstat() should still work. NB. We cannot use
fsp_is_alternate_stream(fsp) for this as the base_fsp
has already been closed at this point and so the value
fsp_is_alternate_stream() checks for is already NULL.
Remove knownfail.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=15487
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Tue Oct 10 09:39:27 UTC 2023 on atb-devel-224
Show that smbd crashes if asked to return full information on close of a
stream handle with delete on close disposition set.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=15487
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Most tests were prepared in advance, but we left these ones to test
the change.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
We have two sets of tests: one that will succeed, and one that is going
to remain a knownfail. The latter involves Resource Attribute ACEs that
have the TX type, meaning "byte string".
In MS-DTYP, a bytestring is defined like "#6869210a", with a hash,
followed by an even number of hex digits. In other places on the web, it
is mentioned that zeroes in the string can be replaced by hashes, like so
"#686921#a". We discover via indirect fuzzing that a TX RA ACE can also
take bare integers, like "6869210a" or "2023". As it would be tricky to
support this, and there is no evidence of this occurring in the wild, we
will probably leave this as a knownfail.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
ACL revision 4 (SECURITY_ACL_REVISION_ADS) is effectively a superset
of revision 2 (SECURITY_ACL_REVISION_NT4), so any revision 2
ACL can be called revision 4 without any problem. But not vice versa:
a revision 4 ACL can contain ACE types that a revision 2 ACL can't. The
extra ACE types relate to objects.
Samba currently simplifies things by calling all its ACLs revision 4,
even if (as is commonly the case) the ACLs contain only revision 2 ACEs.
On the other hand, Windows will use revision 2 whenever it can. In other
tests we skip past this by forcing Windows ACLs to v4 before comparison.
This test is to remind us of the incompatibility.
It would not be hard to fix.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
If there are multiple identical ACEs in an SDDL ACL, Windows will decode
them all and put extra trailing zeroes at the end of the ACL.
In contrast, Samba will decode the ACEs and not put extra zeroes at the
end.
The problem comes when Samba tries to read a binary ACL from Windows that
has the extra zeroes, because Samba's ACL size calculation is based on
the size of its constituent ACEs, not the ACL size field.
There is no good reason for an ACL to have repeated ACEs, but they could
be added accidentally.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Matches file and directory closes.
Remove knownfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15423
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Wed Sep 20 02:43:18 UTC 2023 on atb-devel-224
Shows the server crashes if we open a named pipe, do an async read
and then disconnect.
Adds knownfail:
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15423
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
If nss_winbind is loaded into a process that uses fork multiple times
without any further calls into nss_winbind, wb_atfork_child handler
was using a wb_global_ctx.key that was no longer registered in the
pthread library, so we operated on a slot that was potentially
reused by other libraries or the main application. Which is likely
to cause memory corruption.
So we better don't call pthread_key_delete() in wb_atfork_child().
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15464
Reported-by: Krzysztof Piotr Oledzki <ole@ans.pl>
Tested-by: Krzysztof Piotr Oledzki <ole@ans.pl>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Untested code is broken code. For symlinks we need to hand over the
full reparse buffer into symlink_reparse_buffer_parse(), as this is
also used for the smb2 error response handling. For that, the
"reserved" field in [MS-FSCC] 2.1.2.4 Symbolic Link Reparse Data
Buffer is used for the "unparsed" field.
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 Sep 8 17:24:19 UTC 2023 on atb-devel-224
The blob was taken from a smbclient allinfo command for a Windows
symlink. Show that reparse_data_buffer_parse() is broken.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
This way we can run the tests and more easily put them into knownfail
individually. Before this, everything went into the error category,
which was not so easy to catch in something like knownfail.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
`cache_get_all_attribute_values` returns a dict whereas we need to pass
a list of keys to `remove`. These will be interpolated in the gpdb search.
Signed-off-by: Gabriel Nagy <gabriel.nagy@canonical.com>
Reviewed-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: David Mulder <dmulder@samba.org>
Autobuild-User(master): Andrew Bartlett <abartlet@samba.org>
Autobuild-Date(master): Mon Aug 28 03:01:22 UTC 2023 on atb-devel-224
For this we need to stage a Registry.pol file with certificate
autoenrollment enabled, but with checkboxes unticked.
Signed-off-by: Gabriel Nagy <gabriel.nagy@canonical.com>
Reviewed-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: David Mulder <dmulder@samba.org>
If certificate templates are added or removed, the autoenroll extension
should react to this and reapply the policy. Previously this wasn't
taken into account.
Signed-off-by: Gabriel Nagy <gabriel.nagy@canonical.com>
Reviewed-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: David Mulder <dmulder@samba.org>
Ensure that cepces-submit reporting additional templates and re-applying
will enforce the updated policy.
Signed-off-by: Gabriel Nagy <gabriel.nagy@canonical.com>
Reviewed-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: David Mulder <dmulder@samba.org>
I don't know whether this applies universally, but in our case the
contents of `es['cACertificate'][0]` are binary, so cleanly converting
to a string fails with the following:
'utf-8' codec can't decode byte 0x82 in position 1: invalid start byte
We found a fix to be encoding the certificate to base64 when
constructing the CA list.
Section 4.4.5.2 of MS-CAESO also suggests that the content of
`cACertificate` is binary (OCTET string).
Signed-off-by: Gabriel Nagy <gabriel.nagy@canonical.com>
Reviewed-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: David Mulder <dmulder@samba.org>
This fails all GPO-related tests that call `gpupdate --rsop`.
Signed-off-by: Gabriel Nagy <gabriel.nagy@canonical.com>
Reviewed-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: David Mulder <dmulder@samba.org>
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): Thu Aug 24 03:47:08 UTC 2023 on atb-devel-224
Can Samba understand Windows security descriptors? Does it parse SDDL
the same way?
Here we test on over 7000 SDDL/descriptor pairs and find the answer
is pleasing. In later commits we will add more tests using different
classes of ACE.
The test cases are derived from fuzz seeds, exported to Windows via
the script in the last commit, with the Windows descriptor bytes found
using libcli/security/tests/windows/windows-sddl-test.py.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Soon we will get Conditional ACEs and Resource Attribute ACES, each of
which have trailing bytes at the end of the ACE. Here's a diagram:
____ The ACE size field may indicate a size bigger
.type / | than the known parts, even when you take
.flags / | rounding to a multiple of four into account.
.size --' | This extra data is meaningful in some ACEs.
.access_mask |
.trustee (sid) _| <- known data ends here.
:
"coda" ___: <- the trailing part, Zero size unless the size
field points beyond the end of the known data.
Probably empty for ordinary ACE types.
Until now we have thrown away these extra bytes, because they have no
meaning in the ACE types we recognise. But with conditional and
resource attribute ACEs we need to catch and process these bytes, so
we add an extra field for that.
Thus we can drop the manually written ndr_pull_security_ace() that
discarded the trailing bytes, because we just allow it to be pulled
into an unused blob. In the very common case, the blob will be empty.
Microsoft does not use a common name across different ACE types to
describe this end-data -- "coda" is a Samba term.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
This field is supposed to be aligned to eight bytes, but the ‘dlong’
type is aligned to only four bytes. This discrepancy resulted in claims
being encoded and decoded incorrectly.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15452
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Remove the now unneeded req->xxx = NULL assignments (and the
deliberately bogus req->session = (void *)0xDEADBEEF one
used to demonstrate the bug).
Remove knownfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15432
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Noel Power <npower@samba.org>
Autobuild-User(master): Noel Power <npower@samba.org>
Autobuild-Date(master): Tue Aug 15 12:06:36 UTC 2023 on atb-devel-224
Found by Robert Morris <rtm@lcs.mit.edu>.
Adds knownfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15432
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Noel Power <nopower@samba.org>
Remove knownfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15430
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Noel Power <npower@samba.org>
Autobuild-User(master): Noel Power <npower@samba.org>
Autobuild-Date(master): Mon Aug 14 19:52:49 UTC 2023 on atb-devel-224
Robert Morris <rtm@lcs.mit.edu> noticed a missing
return in reply_exit_done().
Adds knownfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15430
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Noel Power <npower@samba.org>
Robert Morris <rtm@lcs.mit.edu> noticed that in the case
where srvstr_pull_req_talloc() is being called with
buffer remaining == 0, we don't NULL out the destination
pointed which is *always* done in the codepaths inside
pull_string_talloc(). This prevents a crash in the caller.
Remove knownfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15420
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Shows that we indirect through an uninitialized pointer and the client crashes
it's own smbd.
Add knownfail.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15420
Signed-off-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
We send the NC root first, as a special case for every chunk
that we send until the natural point where it belongs.
We do not bump the tmp_highest_usn in the highwatermark that
the client and server use (it is meant to be an opauqe cookie)
until the 'natural' point where the object appears, similar
to the cache for GET_ANC.
The issue is that without this, because the NC root was sorted
first in whatever chunk it appeared in but could have a 'high'
highwatermark, Azure AD Connect will send back the same
new_highwatermark->tmp_highest_usn, and due to a bug,
a zero reserved_usn, which makes Samba discard it.
The reserved_usn is now much less likely to ever be set because
the tmp_higest_usn is now always advancing.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15401
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
This changes the GetNCChanges server to use a per-call state for
extended operations like RID_ALLOC or REPL_OBJ and only maintain
and (more importantly) invalidate the state during normal replication.
This allows REPL_OBJ to be called during a normal replication cycle
that continues using after that call, continuing with the same
highwatermark cookie.
Azure AD will do a sequence of (roughly)
* Normal replication (objects 1..100)
* REPL_OBJ (of 1 object)
* Normal replication (objects 101..200)
However, if there are more than 100 (in this example) objects in the
domain, and the second replication is required, the objects 1..100
are sent, as the replication state was invalidated by the REPL_OBJ call.
RN: Improve GetNChanges to address some (but not all "Azure AD Connect")
syncronisation tool looping during the initial user sync phase.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15401
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
The NC root, on any replication when it appears, is the first object to be
replicated, including for all subsequent chunks in the replication.
However the tmp_highest_usn is not updated by that USN, it must
only be updated for the non-NC changes (to match Windows exactly),
or at least only updated with the non-NC changes until it would
naturally appear.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15401
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
This demonstrates the behaviour used by the "Azure AD Connect" cloud sync tool.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15401
Signed-off-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
If a client opens multiple connection with the same
client guid in parallel, our connection passing is likely
to hit a race.
Assume we have 3 processes:
smbdA: This process already handles all connections for
a given client guid
smbdB: This just received a new connection with an
SMB2 neprot for the same client guid
smbdC: This also received a new connection with an
SMB2 neprot for the same client guid
Now both smbdB and smbdC send a MSG_SMBXSRV_CONNECTION_PASS
message to smbdA. These messages contain the socket fd
for each connection.
While waiting for a MSG_SMBXSRV_CONNECTION_PASSED message
from smbdA, both smbdB and smbdC watch the smbXcli_client.tdb
record for changes (that also verifies smbdA stays alive).
Once one of them say smbdB received the MSG_SMBXSRV_CONNECTION_PASSED
message, the dbwrap_watch logic will wakeup smbdC in order to
let it recheck the smbXcli_client.tdb record in order to
handle the case where smbdA died or deleted its record.
Now smbdC rechecks the smbXcli_client.tdb record, but it
was not woken because of a problem with smbdA. It meant
that smbdC sends a MSG_SMBXSRV_CONNECTION_PASS message
including the socket fd again.
As a result smbdA got the socket fd from smbdC twice (or even more),
and creates two (or more) smbXsrv_connection structures for the
same low level tcp connection. And it also sends more than one
SMB2 negprot response. Depending on the tevent logic, it will
use different smbXsrv_connection structures to process incoming
requests. And this will almost immediately result in errors.
The typicall error is:
smb2_validate_sequence_number: smb2_validate_sequence_number: bad message_id 2 (sequence id 2) (granted = 1, low = 1, range = 1)
But other errors would also be possible.
The detail that leads to the long delays on the client side is
that our smbd_server_connection_terminate_ex() code will close
only the fd of a single smbXsrv_connection, but the refcount
on the socket fd in the kernel is still not 0, so the tcp
connection is still alive...
Now we remember the server_id of the process that we send
the MSG_SMBXSRV_CONNECTION_PASS message to. And just keep
watching the smbXcli_client.tdb record if the server_id
don't change. As we just need more patience to wait for
the MSG_SMBXSRV_CONNECTION_PASSED message.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15346
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Andreas Schneider <asn@cryptomilk.org>
Autobuild-Date(master): Tue Aug 8 13:59:58 UTC 2023 on atb-devel-224
This demonstrates the race quite easily against
Samba and works fine against Windows Server 2022.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15346
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
It seems commit 259129e8f4bc8cacd1850eba3f6551134835d079 was partly just
fantasy...
Windows clients just use 16 bytes for DCERPC_PKT_CO_CANCEL and
DCERPC_PKT_ORPHANED pdus.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15446
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Tue Aug 8 08:57:46 UTC 2023 on atb-devel-224
