Previously, this would fail the assertion
"connection->initialization_error != NULL" after the label "out".
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=665067
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: David Zeuthen <davidz@redhat.com>
This happens to work at the moment (because GDBusWorker.frozen is a
gboolean and not just a 1-bit bitfield), but isn't right: the gboolean
ends up with values 0 or G_DBUS_CONNECTION_FLAGS_DELAY_MESSAGE_PROCESSING
(which is more than 1).
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=664558
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: David Zeuthen <davidz@redhat.com>
Strictly speaking, neither of the two uses that aren't under the lock
*needs* to be atomic, but it seems better to be obviously correct (and
we save another 4 bytes of struct).
One of these uses is in g_dbus_connection_is_closed(), any use of which
is inherently a race condition anyway.
The other is g_dbus_connection_flush_sync, which as far as I can tell
just needs a best-effort check, to not waste effort on a connection that
has been closed for a while (but I could be wrong).
I removed the check for the closed flag altogether in
g_dbus_connection_send_message_with_reply_unlocked, because it turns out
to be redundant with one in g_dbus_connection_send_message_unlocked,
which is called immediately after.
g_dbus_connection_close_sync held the lock to check the closed flag,
which is no longer needed.
As far as I can tell, the only reason why the lock is still desirable
when setting the closed flag is so that remove_match_rule can't fail
by racing with close notification from the worker thread - but
on_worker_closed needs to hold the lock anyway, to deal with other
data structures, so there's no point in trying to eliminate the
requirement to hold the lock.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=661992
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: David Zeuthen <davidz@redhat.com>
Also, a few that don't need to be.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=661992
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: David Zeuthen <davidz@redhat.com>
This isn't strictly necessary, because in every location where it's
checked, if the reading thread misses an update from another thread,
it's indistinguishable from the reading thread having been scheduled
before the writing thread, which is an unavoidable race condition that
callers need to cope with anyway. On the other hand, merging exit_on_close
into atomic_flags gives the least astonishing semantics to library users
and saves 4 bytes of struct, and if you're accessing exit-on-close often
enough for it to be a performance concern, you're probably doing it wrong.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=661992
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: David Zeuthen <davidz@redhat.com>
The thread shared between all GDBusWorker instances was variously called
the "worker thread" or "message handler thread", which I mostly changed to
"the GDBusWorker thread" to avoid ambiguity.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=661992
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: David Zeuthen <davidz@redhat.com>
The only exceptions are those of the trivial getters/setters that don't
already need the initialization check for its secondary role as a memory
barrier (this is consistent with GSocket, where trivial getters/setters
don't check):
* g_dbus_connection_set_exit_on_close
* g_dbus_connection_get_exit_on_close
* g_dbus_connection_is_closed
g_dbus_connection_set_exit_on_close needs to be safe for
use before initialization anyway, so it can be set at construct-time.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=661689
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: David Zeuthen <davidz@redhat.com>
Also document which fields require such a check in order to have correct
threading semantics.
This usage doesn't matches the GInitable documentation, which suggests
use of a GError - but using an uninitialized GDBusConnection is
programming error, and not usefully recoverable. (The GInitable
documentation may have been a mistake - GNOME#662208.) Also, not all of
the places where we need it can raise a GError.
The check serves a dual purpose: it turns a non-deterministic crash into
a deterministic critical warning, and is also a memory barrier for
thread-safety. All of these functions dereference or return fields that
are meant to be protected by FLAG_INITIALIZED, so they could crash or
return an undefined value to their caller without this, if called from a
thread that isn't the one that called initable_init() (although I can't
think of any way to do that without encountering a memory barrier,
undefined behaviour, or a race condition that leads to undefined
behaviour if the non-initializing thread wins the race).
One exception is that initable_init() itself makes a synchronous call.
We deal with that by passing new internal flags up the call stack, to
reassure g_dbus_connection_send_message_unlocked() that it can go ahead.
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=661689
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=661992
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Reviewed-by: David Zeuthen <davidz@redhat.com>
The comment implied that even failed initialization would set
is_initialized = TRUE, but this wasn't the case - failed initialization
would only set initialization_error, and it was necessary to check both.
It turns out the documented semantics are nicer than the implemented
semantics, since this lets us use atomic operations, which are also
memory barriers, to avoid needing separate memory barriers or locks
for initialization_error (and other members that are read-only after
construction).
I expect to need more than one atomically-accessed flag to fix thread
safety, so instead of a minimal implementation I've turned is_initialized
into a flags word.
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=661689
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=661992
Reviewed-by: David Zeuthen <davidz@redhat.com>
If the connection to the bus is lost while a method call is ongoing,
the method call does not get cancelled. Instead it just sits around
until it times out.
This is visible here on XO laptops when stopping the display manager
during shutdown. imsettings starts sending a sync message to give up
its bus name (via g_bus_unown_name()), then systemd terminates the
session bus at approximately the same time. imsettings then hangs for
about 20 seconds before timing out the message.
http://lists.freedesktop.org/archives/dbus/2011-September/014717.html
imsettings behaviour could be improved as described in that thread,
but I think this is a glib bug. I've also come up with the attached
patch which fixes it.
Credits for the bug-fix goes to Daniel Drake <dsd@laptop.org>. The test
case was written by David Zeuthen <zeuthen@gmail.com>.
https://bugzilla.gnome.org/show_bug.cgi?id=660637
Signed-off-by: David Zeuthen <davidz@redhat.com>
Add g_main_context_ref_thread_default(), which always returns a
reffed GMainContext, rather than sometimes returning a (non-reffed)
GMainContext, and sometimes returning NULL. This simplifies the
bookkeeping in any code that needs to keep a reference to the
thread-default context for a while.
https://bugzilla.gnome.org/show_bug.cgi?id=660994
We can't safely close the output part of the I/O stream until any
pending write or flush has been completed. In the worst case, this could
lead to an assertion failure in the worker (when the close wins the
race) or not closing the stream at all (when the write wins the race).
Bug: https://bugzilla.gnome.org/show_bug.cgi?id=651268
Bug-NB: NB#271520
Signed-off-by: Simon McVittie <simon.mcvittie@collabora.co.uk>
Signed-off-by: David Zeuthen <davidz@redhat.com>
Also add convenience _with_unix_fd_list variants to GDBusConnection,
GDBusProxy and GDBusMethodInvocation types to easily support this.
Signed-off-by: David Zeuthen <davidz@redhat.com>
This function implements the following logic:
if (g_variant_is_floating (value))
g_variant_ref_sink (value);
which is used for consuming the return value of callbacks that may or
may not return floating references.
This patch also replaces a few instances of the above code with the new
function (GSettings, GDBus) and lifts a long-standing restriction on the
use of floating values as the return value for signal handlers by
improving g_value_take_variant().
https://bugzilla.gnome.org/show_bug.cgi?id=627974
To help cross compilation, don't use glib-genmarshal in our
build. This is easy now that we have g_cclosure_marshal_generic().
In gobject/, add gmarshal.[ch] to git (making the existing entry
points stubs).
In gio/, simply switch to using g_cclosure_marshal_generic().
https://bugzilla.gnome.org/show_bug.cgi?id=652168
If g_bus_get_sync() fails in authentication (because e.g. the process
uid, doesn't match the expected in EXTERNAL), a secondary call to
g_bus_get_sync() would notice we aren't initialized, and try
to initialize.
The assertion here is just wrong; we now explicitly and clearly handle
both cases where we already have an error, or we already succeeded.
https://bugzilla.gnome.org/show_bug.cgi?id=635694
If specified, the signal subscription is setup client-side but the match
rule is not sent to the server. This allows the caller to manually
register more detailed match rules.
It doesn't really work right now because of a dbus-daemon(1) bug - see
the comment added in the TODO section of gdbusconnection.c. So revert
to old behavior. The downside is a lot of files in /tmp but right now
that's better than not being able to run tests in a loop.
Signed-off-by: David Zeuthen <davidz@redhat.com>
Without this fix, the ./gdbus-connection test case occasionally fails, see
https://bugzilla.gnome.org/show_bug.cgi?id=629945#c5
like this
/gdbus/connection/basic: OK
/gdbus/connection/life-cycle: **
ERROR:gdbus-connection.c:223:test_connection_life_cycle: assertion failed:
(!quit_mainloop_fired)
cleaning up bus with pid 21794
Aborted (core dumped)
because the callback didn't happen on the same thread as where we are
running the loop.
Signed-off-by: David Zeuthen <davidz@redhat.com>
Turns out that GDBusWorker will issue callbacks (in its own thread)
even after g_dbus_worker_stop() has been called. This would rarely
happen (and unreffing a connection is even rarer) so only saw this bug
occasionally when running the gdbus-connection test case in a loop.
Fix up this issue by maintaining a set of GDBusConnection objects that
are currently "alive" and do nothing in the callbacks if the passed
user_data pointer is not in this set.
Also attempted to fix up a race condition with
_g_object_wait_for_single_ref_do() and its use of GObject toggle
references - for now, just resort to busy waiting, thereby
sidestepping the toggle reference mess altogether.
Signed-off-by: David Zeuthen <davidz@redhat.com>
==7269== 144 bytes in 6 blocks are definitely lost in loss record 1,282 of 1,325
==7269== at 0x4005BDC: malloc (vg_replace_malloc.c:195)
==7269== by 0x4056B74: g_malloc (gmem.c:164)
==7269== by 0x406EDB6: g_slice_alloc (gslice.c:842)
==7269== by 0x406EDFB: g_slice_alloc0 (gslice.c:854)
==7269== by 0x413C627: g_type_create_instance (gtype.c:1867)
==7269== by 0x412276A: g_object_constructor (gobject.c:1480)
==7269== by 0x4121E5D: g_object_newv (gobject.c:1264)
==7269== by 0x4121BD3: g_object_new (gobject.c:1176)
==7269== by 0x417CFB9: g_credentials_new (gcredentials.c:156)
==7269== by 0x41D9DBC: g_unix_credentials_message_deserialize (gunixcredentialsmessage.c:149)
==7269== by 0x41C422C: g_socket_control_message_deserialize (gsocketcontrolmessage.c:198)
==7269== by 0x41BFCE3: g_socket_receive_message (gsocket.c:3289)
==7269== by 0x41D99CE: g_unix_connection_receive_credentials (gunixconnection.c:476)
==7269== by 0x41FA829: _g_dbus_auth_run_server (gdbusauth.c:987)
==7269== by 0x4205DDB: initable_init (gdbusconnection.c:2196)
Bug #629689.
Rework filter functions as per
https://bugzilla.gnome.org/show_bug.cgi?id=624546#c8
This commit breaks ABI. However, this ABI break affects only
applications using filter functions. The only known user of is dconf.
Signed-off-by: David Zeuthen <davidz@redhat.com>
