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.
Don't close unconsumed file descriptors in messaging_recv_cb(). Via
multiple registrations on different tevent contexts we might call
messaging_recv_cb() multiple times: All but the first tevent context
handled in the loop in msg_dgm_ref_recv() will not see file
descriptors anymore, it will just get a -1, even if the first
reference had no receiver interested in the fds.
Change the API such that consumers can set the file descriptor to -1
if it's consumed. If nobody wanted them, do the close where they were
created via recvmsg, in messages_dgm.c.
If you want multiple handlers to consume the file descriptors, you
should dup() them in the filter function handed to
messaging_filtered_read_send and save the duplicate in your private
data for later consumption.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Autobuild-User(master): Volker Lendecke <vl@samba.org>
Autobuild-Date(master): Fri Mar 19 08:18:26 UTC 2021 on sn-devel-184
In messaging_rec_create(), messaging_recv_cb() and
messaging_dispatch_rec(), variable length arrays of file descriptors are
declared using an incoming num_fds parameter.
However, there are several scenarios where num_fds can be zero, and
declaring a zero-length VLA is undefined behavior. This can lead to
segmentation faults and/or other crashes when compiling with recent
versions of clang at high optimization levels.
To avoid ever using zero as the length for these declarations, use
MAX(1, length) instead.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=14605
Signed-off-by: Dimitry Andric <dimitry@andric.com>
Reviewed-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
Autobuild-User(master): Ralph Böhme <slow@samba.org>
Autobuild-Date(master): Mon Jan 4 10:50:07 UTC 2021 on sn-devel-184
Early talloc_free() for the msg_rec if it's not picked up
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 Oct 21 20:27:57 UTC 2020 on sn-devel-184
This is a bit confusing now, sorry for that:
register_msg_pool_usage() in the ctdb case uses
messaging_ctdb_register_tevent_context(), which talloc_reference()s
the central struct messaging_ctdb_fde_ev of the
messaging_ctdb_context. In messaging_reinit(), we talloc_free only one
of those references and allocate a new messaging_ctdb_fde_ev. The
remaining messaging_ctdb_fde_ev should have been deleted as well, but
due to the second reference this does not happen. When doing the
shutdown messaging_ctdb_fde_ev_destructor() is called twice, once on
the properly reinitialized fde_ev, and once much later on the leftover
one which references invalid data structures.
By the way, this is not a problem with talloc_reference(), this would
have happened with explicit refcounting too.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=14281
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Autobuild-User(master): Björn Baumbach <bb@sernet.de>
Autobuild-Date(master): Tue Feb 18 13:05:53 UTC 2020 on sn-devel-184
Consolidate "msg_dgm_ref" and "msg_ctdb_ref": The only purpose of
those pointers was to TALLOC_FREE() them in messaging_reinit(). We'll
have a third entity to talloc_free() in the next commit, make that
simpler.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=14281
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Add a simple way to deactivate the registration
Bug: https://bugzilla.samba.org/show_bug.cgi?id=14281
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Martin Schwenke <martin@meltin.net>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Fixes:
source3/lib/messages.c:483:11: warning: Value stored to 'status' during its initialization is never read <--[clang]
NTSTATUS status = NT_STATUS_UNSUCCESSFUL;
^~~~~~ ~~~~~~~~~~~~~~~~~~~~~~
1 warning generated.
Signed-off-by: Noel Power <noel.power@suse.com>
Reviewed-by: Gary Lockyer <gary@catalyst.net.nz>
This is spamming the console when running 'make test' and we set the log
level to 1 by default in selftest.
Signed-off-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
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>
As it gets 'messaging_context' as argument, we're sure a messaging context
with a raw tevent context already exist.
It means we can allow a wrapper tevent context that wrapps the main tevent
context of the messaging context.
The use of tevent_req_defer_callback() makes sure that the callers
callback function calls messaging_filtered_read_recv() from the
correct "wrapped" environment.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
This makes a lot of assumtion easier to understand and the introduction
of wrapper tevent contexts will not change the existing behaviour.
We'll relax this a bit in the next commits.
Signed-off-by: Stefan Metzmacher <metze@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
We should create the immediate event at the beginning
were we have a chance to return an error, rather than
ignoring a failure later.
As a side effect this also reuses the immediate event
after the refcount went to 0 and up again.
Pair-Programmed-With: Stefan Metzmacher <metze@samba.org>
Signed-off-by: Ralph Boehme <slow@samba.org>
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 Apr 24 14:30:20 CEST 2018 on sn-devel-144
Commit ad8c7171ba accidently
moved sec_init() to the point after sec_initial_uid() is
called in the call to directory_create_or_exist_strict().
I missed this in the review (sorry). This works as root
as initial_uid/initial_gid are static (and so initialized
as zero) but doesn't work on ChromeOS as this code isn't
running as root.
BUG: https://bugzilla.samba.org/show_bug.cgi?id=13368
Signed-off-by: Lutz Justen <ljusten@google.com>
Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-by: Andreas Schneider <asn@samba.org>
Autobuild-User(master): Jeremy Allison <jra@samba.org>
Autobuild-Date(master): Wed Apr 4 23:52:02 CEST 2018 on sn-devel-144
For non-clustered messaging this should have never gone through the socket, we
should have caught it before in messaging_send_iov_from.
It can come in on a socket from ctdb when broadcasting in clustered mode. There
ctdb does the broadcasting.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
This leads to cleanupd doing endless MSG_SMB_UNLOCK calls, as it triggers
itself in the send_all. This worked correctly before the serverid.tdb removal
because cleanupd did not register in serverid.tdb (which was a bug, but it
helped us there).
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
This will replace message_send_all. With messaging_dgm_forall we have
a local broadcast mechanism, and ctdb can also broadcast
everywhere. So there's no need for a separate traverse/send mechanism.
There's no good error reporting mechanism for broadcasting, so make
this function void.
This drops the message_type filtering. I believe that this does not matter in
practice, since messaging is a lot cheaper with dgm instead of the old tdb
based messaging. If someone presents a use case where this matters, nowadays
I'd much rather extend the messaging_dgm lock file format (where the unique id
lives right now) with the filter bits.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Signed-off-by: Anoop C S <anoopcs@redhat.com>
Reviewed-by: Andreas Schneider <asn@samba.org>
Reviewed-by: Martin Schwenke <martin@meltin.net>
Autobuild-User(master): Martin Schwenke <martins@samba.org>
Autobuild-Date(master): Wed Aug 16 16:26:12 CEST 2017 on sn-devel-144
We need to listen for the ctdb socket in nested event contexts
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
We must give all event contexts that might be interested the chance to pick up
the message. If we send a message to ourselves via messaging_send_iov_from,
nested event contexts need to get a chance to see the message. Before this
patch only the main event context in msg_ctx got it.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
This is done with an explicit reference-counting and without talloc
destructors. The code is isolated enough for now that explicit refcount
management seems simpler. This might change in the future, but for
now keep it simple.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
No real code change: This makes dispatching to non-classic receives available
for other callers.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
This did not really take off, notifyd was the only user
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 Jul 7 05:11:48 CEST 2017 on sn-devel-144
This survived an autobuild, so no subsystem strictly needs this anymore. In
particular the notify subsystem has been rewritten.
Why this patch? It removes some complexity from core code, and it reduces the
potential memory overconsumption: Right now I'm working on a g_lock_ping_pong
test. This test does a lot of messaging_filtered_read_send calls in a tight
loop on a nested event context. With the current code we let the
messaging_filtered_read code consume the message that arrives, but it also
posts it for consumption by the main event context attached to the messaging
context with its "classic" callback. This test never comes back to the main
event context, so it accumulates more and more self-posted messages. That's
just unnecessary, given that due to the successful autobuild nothing but the
read1 test makes use of the "multicasting" of messages.
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 Jun 21 07:30:39 CEST 2017 on sn-devel-144
messaging_init_internal() blanket returned NT_STATUS_INTERNAL_ERROR
instead of correctly changing the return code to an NTSTATUS code. Also
return more appropriate mem error.
Bug: https://bugzilla.samba.org/show_bug.cgi?id=12828
Signed-off-by: David Mulder <dmulder@suse.com>
Reviewed-by: Jeremy Allison <jra@samba.org>
Reviewed-by: David Disseldorp <ddiss@samba.org>
Autobuild-User(master): David Disseldorp <ddiss@samba.org>
Autobuild-Date(master): Thu Jun 8 08:04:05 CEST 2017 on sn-devel-144
Bug: https://bugzilla.samba.org/show_bug.cgi?id=12509
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Autobuild-User(master): Stefan Metzmacher <metze@samba.org>
Autobuild-Date(master): Tue Jan 10 17:40:58 CET 2017 on sn-devel-144
Only one tevent_fd is possible for every file descriptor.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Stefan Metzmacher <metze@samba.org>
Only look at "classic" messaging_register handlers in the main event
context loop
If we're sitting in a nested event context, only act upon the
messaging_filtered_read ones that are registered in the nested context.
Postpone everything else via an immediate to the main tevent context
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
We need to go through the event loop, which messaging_dgm_send does. We can
also use a tevent_immediate for the same purpose. Right now the main user is
messaging_ctdb: Here strace looks a bit weird when we receive a message.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Jeremy Allison <jra@samba.org>
Just paranoia
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 Sep 28 03:58:22 CEST 2016 on sn-devel-144
Trying to trim down messages.c a bit: Sending to all processes that are
registered in serverid.tdb and filtering to me is not really logic of general
messaging but more of the serverid code.
Signed-off-by: Volker Lendecke <vl@samba.org>
Reviewed-by: Ralph Boehme <slow@samba.org>
