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.
It's good to have a consistent set of hash_size/flags for all aspects of
an open file handle. Currently we're using 4 databases:
smbXsrv_open_global.tdb, leases.tdb, locking.tdb and brlock.tdb.
While at it also crank up the hashsize if the smbXsrv_tcon and smbXsrv_session
TDBs. The default TDB hash size is insanely small and disk space is cheap these
days, by going with the much larger hash size we get O(1) lookup instead of O(n)
for moderate to large loads with a few thousand objects.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(master): Mon Dec 19 16:40:15 UTC 2022 on sn-devel-184
This avoids using fcntl() locks for dbwrap_delete()
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15125
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
This will become important in the following commits when the
dbwrap_watch layer will only wake up one watcher at a time
and each woken watcher will wakeup the next one.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15125
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
This makes sure we cleanup the locked record in all cases.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15125
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
The destructor triggered by dbwrap_watched_watch_recv() will
remove the watcher instance via a dedicated dbwrap_do_locked(),
just calling dbwrap_watched_watch_remove_instance() inside.
But the typical caller triggers a dbwrap_do_locked() again after
dbwrap_watched_watch_recv() returned. Which means we call
dbwrap_do_locked() twice.
We now allow dbwrap_watched_watch_recv() to return the existing
instance id (if it still exists) and removes the destructor.
That way the caller can pass the given instance id to
dbwrap_watched_watch_remove_instance() from within its own dbwrap_do_locked(),
when it decides to leave the queue, because it's happy with the new
state of the record. In order to get the best performance
dbwrap_watched_watch_remove_instance() should be called before any
dbwrap_record_storev() or dbwrap_record_delete(),
because that will only trigger a single low level storev/delete.
If the caller found out that the state of the record doesn't meet the
expectations and the callers wants to continue watching the
record (from its current position, most likely the first one),
dbwrap_watched_watch_remove_instance() can be skipped and the
instance id can be passed to dbwrap_watched_watch_send() again,
in order to resume waiting on the existing instance.
Currently the watcher instance were always removed (most likely from
the first position) and re-added (to the last position), which may
cause unfair latencies.
In order to improve the overhead of adding a new watcher instance
the caller can call dbwrap_watched_watch_add_instance() before
any dbwrap_record_storev() or dbwrap_record_delete(), which
will only result in a single low level storev/delete.
The returned instance id is then passed to dbwrap_watched_watch_send(),
within the same dbwrap_do_locked() run.
It also adds a way to avoid alerting any callers during
the current dbwrap_do_locked() run.
Layers above may only want to wake up watchers
during specific situations and while it's useless to wake
others in other situations.
This will soon be used to add more fairness to the g_lock code.
Note that this commit only prepares the api for the above to be useful,
the instance returned by dbwrap_watched_watch_recv() is most likely 0,
which means the watcher entry was already removed, but that will change
in the following commits.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=15125
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
There might be other places than smb2srv_update_crypto_flags(), which
may call smbXsrv_session_update() with a fake session, they should
return in error instead of segfaulting.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
When we make use of this we only in order to provide the correct
error codes anyway.
This actually fixes even more error codes.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
I saw systems with locking.tdb records being part of:
ctdb catdb smbXsrv_tcon_global.tdb
It's yet unknown how that happened, but we should not panic in srvsvc_*
calls because the info0 pointer was NULL.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14752
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Volker Lendecke <vl@samba.org>
Autobuild-User(master): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(master): Tue Jul 6 11:08:43 UTC 2021 on sn-devel-184
The session is valid for the lifetime of the requests anyway
and there's no point in having special handling for compound requests.
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>
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>
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>
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 actually fixes crashes due to stale pointers.
With multi-channel and with 2 (or more) connections,
we'll call smbXsrv_session_disconnect_xconn() when a connection
gets disconnected, but we'll leave smbXsrv_client and all other
connections in place.
However smbXsrv_session_disconnect_xconn_callback() left
a stale session->db_rec pointer in place, which means
a following smbXsrv_session_logoff() will call
dbwrap_record_delete(local_rec) on a stale pointer.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14532
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
It's a bit shocking how many references we have to global
contexts. Make this a bit more obvious.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Pointer values can be reused (yes, I hit that during my testing!).
Introduce a channel_id to identify connections and also add
some timestamps to make debugging easier.
This makes smbXsrv_session_find_auth() much more robust.
This is a similar change as 0cec96526b:
"smb2_server: make sure we detect stale smbXsrv_connection pointers in smbXsrv_channel_global"
BUG: https://bugzilla.samba.org/show_bug.cgi?id=11898
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Günther Deschner <gd@samba.org>
This removes the connection references from the session channel
array for each session that's used on the connection.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Guenther Deschner <gd@samba.org>
Pointer values can be reused (yes, I hit that during my testing!).
Introduce a channel_id to identify connections and also add
some timestamps to make debugging easier.
This makes smbXsrv_session_find_channel() much more robust.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Guenther Deschner <gd@samba.org>
At last, the nail in the coffin. :)
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 Jan 13 21:09:01 UTC 2020 on sn-devel-184
In memory of get_valid_user_struct() and functionally equivalent it only returns
the session if session setup was successfully completed and
session->global->auth_session_info is valid.
This function is similar to smbXsrv_session_local_lookup() and it's wrappers,
but it doesn't implement the state checks of those. get_valid_smbXsrv_session()
is NOT meant to be called to validate the session wire-id of incoming SMB
requests, it MUST only be used in later internal processing where the session
wire-id has already been validated.
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
This is not needed anymore because a previous commit changes this to use
session->global->session_wire_id insteaf of session->compat->vuid, so we're not
depending on session->compat anymore.
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
session->compat->vuid is set to session->global->session_wire_id after a
successful session setup, so both variables will always carry the same value. Cf
the next commit which removes vuid from user_struct.
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
This reflects that the messaging context is also used outside of the
server processes.
The command used for the rename:
find . -name '*.[hc]' -print0 | xargs -0 sed -i 's/server_messaging_context/global_messaging_context/'
Signed-off-by: Christof Schmitt <cs@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Bug: https://bugzilla.samba.org/show_bug.cgi?id=13549
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): Sat Sep 1 01:26:35 CEST 2018 on sn-devel-144
This avoids a race in durable handle reconnects if the reconnect comes
in while the old session is still in the tear-down phase.
The new session is supposed to rendezvous with and wait for destruction
of the old session, which is internally implemented with
dbwrap_watch_send() on the old session record.
If the old session deletes the session record before calling
file_close_user() which marks all file handles as disconnected, the
durable handle reconnect in the new session will fail as the records are
not yet marked as disconnected which is a prerequisite.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=13549
Signed-off-by: Ralph Boehme <slow@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Point out in the API that "backend" talloc_moves into the watched
database.
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 Aug 17 21:29:15 CEST 2018 on sn-devel-144
Fix a confusing API: Many places TALLOC_FREE the path where it's not
clear you have to do it.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
This is the same pointer and we don't have a lot of callers,
so we can just use one pointer.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
That makes it clearer that no tevent_context wrapper is used here
and the related code should really run without any (active) impersonation
as before.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
The initial idea was to have some "atomicity" in this API. Every
caller interested in a record would have to do something with
it once it changes. However, only one caller really used this
feature, and that is easily changed to not use it. So
remove the complexity.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
This is the only user of the "prec" argument of
dbwrap_watched_watch_recv. The next patch will remove this
functionality, as it's easily replaced here.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Typicall, when we watch a record, we wait for a process to give up some
resource. Be it an oplock, a share mode or the g_lock. If everything goes well,
the blocker sends us a message. If the blocker dies hard, we want to also be
informed immediately.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
smb2srv_session_close_previous_check crashes if
ndr_pull_smbXsrv_session_globalB fails for some reason. It depends on "is_free"
to be correctly set. All we can do for an invalid database is to discard the
record and set it free.
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 11 00:12:18 CET 2016 on sn-devel-144
This allows to create more than one channel.
Pair-Programmed-With: Michael Adam <obnox@samba.org>
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Michael Adam <obnox@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Thu Sep 10 21:33:47 CEST 2015 on sn-devel-104
