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.
In order to allow replays of requests on a channel failure, we should
not cancel pending requests, the strategie that seems to make windows
clients happy is to let the requests running and return
NT_STATUS_FILE_NOT_AVAILABLE as long as the original request is still
pending.
Here we introduce xconn->transport.shutdown_wait_queue, this is used
to keep the xconn alive for the lifetime of pending requests.
Now we only cancel pending requests if the disconnected connection
is the last channel for a session.
In that case smbXsrv_session_remove_channel() and
smb2srv_session_shutdown_send() will take care of it.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14449
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Before processing an open we need to reserve the replay cache entry
in order to signal that we're still in progress.
If a reserved record is already present we need to return
FILE_NOT_AVAILABLE in order to let the client retry again.
[MS-SMB2] contains this:
<152> Section 3.2.5.1: For the following error codes, Windows-based clients
will retry the operation up to three times and then retry the operation every 5
seconds until the count of milliseconds specified by Open.ResilientTimeout is
exceeded:
- STATUS_SERVER_UNAVAILABLE
- STATUS_FILE_NOT_AVAILABLE
- STATUS_SHARE_UNAVAILABLE
This works fine for windows clients, but current windows servers seems to
return ACCESS_DENIED instead of FILE_NOT_AVAILABLE.
A Windows server doesn't do any replay detection on pending opens,
which wait for a HANDLE lease to be broken (because of a
SHARING_VIOLATION), at all.
As this is not really documented for the server part of the current [MS-SMB2],
I found the key hint in "SMB 2.2: Bigger. Faster. Scalier - (Parts 1 and 2)"
on page 24. There's a picture showing that a replay gets FILE_NOT_AVAILABLE
as long as the original request is still in progress. See:
https://www.snia.org/educational-library/smb-22-bigger-faster-scalier-parts-1-and-2-2011
A Windows client is unhappy with the current windows server behavior if it
such a situation happens. There's also a very strange interaction with oplock
where the replay gets SHARING_VIOLATION after 35 seconds because it conflicts with
the original open.
I think it's good to follow the intial design from the 2011 presentation and
make the clients happy by using FILE_NOT_AVAILABLE (and differ from Windows).
I'll report that to dochelp@microsoft.com in order to get this hopefully fixed in
their server too).
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14449
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
There's nothing special regarding the last channel,
as the smb2.session.bind2 test demonstrates.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14449
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
This demonstrates that a session and it's open handles is destroyed
when the last explicitly bound channel gets disconnected.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14449
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
These demonstrate that the replay detection for pending opens
either doesn't exist (for the share_access=NONE => SHARING_VIOLATION
case) or return the wrong status code => ACCESS_DENIED instead of
FILE_NOT_AVAILABLE.
Windows clients transparently retry after FILE_NOT_AVAILABLE,
while they pass ACCESS_DENIED directly to the application.
I'll report that to dochelp@microsoft.com in order to
clarify the situation.
In the meantime I added tests with a '-windows' suffix,
which demostrate the current windows server behavior,
while the tests with a '-sane' suffix expect the behavior
that whould make windows clients happy.
For Samba I'll implement the '-sane' behavior that
detects all replays and returns FILE_NOT_AVAILABLE
if the original request is still pending.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14449
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
A DN string with lots of trailing space can cause ldb_dn_explode() to
put a zero byte in the wrong place in the heap.
When a DN string has a value represented with trailing spaces,
like this
"CN=foo ,DC=bar"
the whitespace is supposed to be ignored. We keep track of this in the
`t` pointer, which is NULL when we are not walking through trailing
spaces, and points to the first space when we are. We are walking with
the `p` pointer, writing the value to `d`, and keeping the length in
`l`.
"CN=foo ,DC= " ==> "foo "
^ ^ ^
t p d
--l---
The value is finished when we encounter a comma or the end of the
string. If `t` is not NULL at that point, we assume there are trailing
spaces and wind `d and `l` back by the correct amount. Then we switch
to expecting an attribute name (e.g. "CN"), until we get to an "=",
which puts us back into looking for a value.
Unfortunately, we forget to immediately tell `t` that we'd finished
the last value, we can end up like this:
"CN=foo ,DC= " ==> ""
^ ^ ^
t p d
l=0
where `p` is pointing to a new value that contains only spaces, while
`t` is still referring to the old value. `p` notices the value ends,
and we subtract `p - t` from `d`:
"CN=foo ,DC= " ==> ? ""
^ ^ ^
t p d
l ~= SIZE_MAX - 8
At that point `d` wants to terminate its string with a '\0', but
instead it terminates someone else's byte. This does not crash if the
number of trailing spaces is small, as `d` will point into a previous
value (a copy of "foo" in this example). Corrupting that value will
ultimately not matter, as we will soon try to allocate a buffer `l`
long, which will be greater than the available memory and the whole
operation will fail properly.
However, with more spaces, `d` will point into memory before the
beginning of the allocated buffer, with the exact offset depending on
the length of the earlier attributes and the number of spaces.
What about a longer DN with more attributes? For example,
"CN=foo ,DC= ,DC=example,DC=com" -- since `d` has moved out of
bounds, won't we continue to use it and write more DN values into
mystery memory? Fortunately not, because the aforementioned allocation
of `l` bytes must happen first, and `l` is now huge. The allocation
happens in a talloc_memdup(), which is by default restricted to
allocating 256MB.
So this allows a person who controls a string parsed by ldb_dn_explode
to corrupt heap memory by placing a single zero byte at a chosen
offset before the allocated buffer.
An LDAP bind request can send a string DN as a username. This DN is
necessarily parsed before the password is checked, so an attacker does
not need proper credentials. The attacker can easily cause a denial of
service and we cannot rule out more subtle attacks.
The immediate solution is to reset `t` to NULL when a comma is
encountered, indicating that we are no longer looking at trailing
whitespace.
Found with the help of Honggfuzz.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14595
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
The old behaviour attempted to check for and remove files with duplicate
names, but did not do so due to a bug, and would have left undetermined
which files were given priority when duplicate filenames were present.
Now when hardlinks are present, only one instance of each file is
chosen, with files in the private directory having priority. If one
backup dir is nested inside another, the files contained in the nested
directory are only added once. Additionally, the BIND DNS database is
omitted from the backup.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14027
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz
This test verifies that when performing an offline backup of a domain
where one of the directories to be backed up is nested inside another,
the contained files are only included once in the backup.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14027
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz
This test verifies that when performing an offline backup of a domain
where the directories to be backed up contain hardlinks, only one
instance of each file is backed up, and that files in the private
directory take precedence.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14027
Signed-off-by: Joseph Sutton <josephsutton@catalyst.net.nz>
Reviewed-by: Andrew Bartlett <abartlet@samba.org>
Reviewed-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz
Signed-off-by: David Mulder <dmulder@suse.com>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Thu Mar 18 20:02:50 UTC 2021 on sn-devel-184
We no longer run any *python2* or *python3* specific tests, so
these knownfail lines are just clutter.
Signed-off-by: Douglas Bagnall <douglas.bagnall@catalyst.net.nz>
Reviewed-by: Jeremy Allison <jra@samba.org>
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14512
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Wed Mar 17 01:56:37 UTC 2021 on sn-devel-184
The ACCESS_DENIED errors happened as we didn't expected to signing
algo is attached to the session key. So our client calculated the
wrong signature.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14512
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
When the session is not valid on the current connection it should not be
possible to start a reauth.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14512
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
If someone tries to operate on a session that is not yet valid on the
current connection and the current session setup fails, then we should
not shutdown the session.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14512
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
The key is that we need to have the signing key in order to pass the
signing checks and give the correct session bind error status.
This should fix the MultipleChannel_Negative_SMB2002 testcase
of the Windows Protocol Test Suite (FileServer).
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14512
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reported-by: Jones Syue <jonessyue@qnap.com>
Reviewed-by: Jeremy Allison <jra@samba.org>
This allows us to test session binds with different users.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14512
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
These demonstrate that a failing bind does not destroy
the existing session and binding with a different user results
in ACCESS_DENIED.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14512
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
'smb2.session.bind_negative_smb202' is similar to the MultipleChannel_Negative_SMB2002 test
from the Windows Protocol Test Suite.
It demonstrates that the server needs to do lookup
in the global session table in order to get the signing
and error code of invalid session setups correct.
In order to work out the details I've added more similar tests.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14512
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
The current_user_info is updated at the SMB level, but currently not at the RPC
level in the RPC impersonation function smbd_become_authenticated_pipe_user().
For RPC services running embedded this is not an issue as the SMB level
impersonation has already taken care of current_user_info, but for RPC services
running as external daemons, eg spoolssd, the omission of updating
current_user_info results in variable expansion of eg %U (username) to be
broken.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14668
MR: https://gitlab.com/samba-team/samba/-/merge_requests/1834
RN: %U variable expansion not working in spoolsd
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Björn Baumbach <bb@sernet.de>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Fri Mar 12 00:54:01 UTC 2021 on sn-devel-184
If we don't anticipate a missing principal name,
samba-tool crashes. Also, principal names could
be in dispersed listelements.
Signed-off-by: David Mulder <dmulder@suse.com>
Reviewed-by: Björn Baumbach <bb@sernet.de>
Signed-off-by: David Mulder <dmulder@suse.com>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Mon Mar 8 20:57:50 UTC 2021 on sn-devel-184
If we don't anticipate a missing principal name,
the extension crashes. Also, principal names could
be in dispersed listelements.
Signed-off-by: David Mulder <dmulder@suse.com>
Reviewed-by: Jeremy Allison <jra@samba.org>
Running samba-gpupdate on a client is causing an
error in gp_access_ext, due to it attempting to
access sam.ldb before detecting whether we are on
an ad-dc.
Signed-off-by: David Mulder <dmulder@suse.com>
Reviewed-by: Jeremy Allison <jra@samba.org>
