This lets us run a subset of the tests that previously relied on extra
test-only code being compiled into libdbus.
Signed-off-by: Simon McVittie <smcv@collabora.com>
We don't actually complete successful authentication, because that
would require us to generate a cookie and compute the correct SHA1,
which is difficult to do in a deterministic authentication script.
However, we do assert that dbus#269 (CVE-2019-12749) has been fixed.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Some NSS modules like those for sssd and LDAP might allocate fds
on a one-per-process basis, for example a socket to talk to sssd.
Make sure those have already been allocated before we enter the code
under test, so that they don't show up as having been "leaked" by the
first module of code under test that happens to do a NSS lookup.
The call to _dbus_test_check_memleaks tears down libdbus' own memory
allocations, but not any hidden state in libc or NSS.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Now that there is no code outside test/ that calls into this, we can
move it into test/, reducing the size of libdbus.
dbus-test-tap.[ch] still need to stay in dbus/ as long as there is
code in dbus/ or bus/ relying on them, and also need to be linked into
libdbus as long as there is other code in libdbus relying on them,
so they stay where they are for now. Similarly, dbus-tests.h lists
the tests that are still embedded in libdbus, and must stay where
it is for the moment.
With this move, various tests now need to be linked to the dbus-testutils
convenience library.
Signed-off-by: Simon McVittie <smcv@collabora.com>
As far as I can tell, the only significant contributions have been
from Red Hat and Collabora, both of which have given permission to
relicense their parts of the dbus codebase to MIT/X11; so let's go
with that.
I'm assuming here that tree-wide changes from
_DBUS_GNUC_WARN_UNUSED_RESULT to _DBUS_WARN_UNUSED_RESULT, and tree-wide
changes to how we include config.h, are not significant for copyright
purposes.
Signed-off-by: Simon McVittie <smcv@collabora.com>
test_object_try_whatever() now has libdbus-like OOM handling,
while test_object_whatever() has GLib-like OOM handling. This is
because an overwhelming majority of the callers of these functions
either didn't check for OOM anyway, or checked for it but then
aborted. In the uncommon case where we do care, we can use the _try_
version.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=100317
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Signed-off-by: Simon McVittie <smcv@collabora.com>
If _dbus_loop_queue_dispatch fails with OOM, we'd try to free cd,
while cd is already owned by the connection's timeout functions.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=100317
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Signed-off-by: Simon McVittie <smcv@collabora.com>
This gets rid of a potential circular dependency, which is annoying
when bootstrapping. It is nice to have the regression tests use
the shared libdbus, but we're about to make it possible to
do that anyway, even though some of them use internal symbols.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=83115
Reviewed-by: Ralf Habacker <ralf.habacker@freenet.de>
In the process, make test_kill_pid() safer: do not try to terminate
more than one pid, or the NULL handle.
Also stop leaking the address_fd in spawn_dbus_daemon, a pre-existing
bug that was spotted by Philip Withnall during review.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=88810
Reviewed-by: Philip Withnall
We only use dbus-glib for its main loop; within dbus, DBusLoop is
available as an alternative, although it isn't thread-safe and
isn't public API.
For tests that otherwise only use libdbus public API, it's desirable to
be able to avoid DBusLoop, so we can run them against an installed
libdbus as an integration test. However, if we don't have dbus-glib,
we're going to have to use an in-tree main loop, which might as well
be DBusLoop.
The major disadvantage of using dbus-glib is that it isn't safe to
link both dbus-1 and dbus-internal at the same time. This is awkward
for a future test case that wants to use _dbus_getsid() in dbus-daemon.c,
but only on Windows (fd.o #54445). If we use the same API wrapper around
both dbus-glib and DBusLoop, we can compile that test against dbus-glib
or against DBusLoop, depending on the platform.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=68852
Reviewed-by: Ralf Habacker <ralf.habacker@freenet.de>
Similar to the previous commit, almost every use of DBusWatch can just
have the main loop call dbus_watch_handle.
The one exception is the bus activation code; it's had a comment
explaining why it's wrong since 2003. We should fix that one day, but for
now, just migrate it to a new _dbus_loop_add_watch_full which preserves
the second-layer callback.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=33342
Reviewed-by: Thiago Macieira <thiago@kde.org>
Instead of supplying 8 tiny wrapper functions around dbus_timeout_handle,
each with a user_data parameter that's a potentially unsafe borrowed
pointer but isn't actually used, we can call dbus_timeout_handle directly
and save a lot of trouble.
One of the wrappers previously called dbus_timeout_handle repeatedly
if it returned FALSE to indicate OOM, but that timeout's handler never
actually returned FALSE, so there was no practical effect. The rest just
ignore the return, which is documented as OK to do.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=33342
Reviewed-by: Thiago Macieira <thiago@kde.org>
This is similar to how ConnectionData works. Without this change, we
deserve to segfault: when the first set of callbacks (either watches or
timeouts) is cleaned up, we unref the server and loop, and free sd;
when the second set of callbacks is cleaned up, we use-after-free sd,
the server and the loop, then double-free sd.
However, due to fd.o #33277 we don't even get that far, because we've
already died with an assertion failure.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=33277
* dbus/dbus-connection.c (_dbus_connection_close_if_only_one_ref):
Add a hack to make DBusNewConnectionFunction work right.
* dbus/dbus-server-socket.c (handle_new_client_fd_and_unlock): use
the hack here. Also, fix the todo about refcount leak.
* dbus/dbus-server-debug-pipe.c (_dbus_transport_debug_pipe_new):
and use the hack here
* dbus/dbus-connection.c: Kill the "shared" flag vs. the
"shareable" flag; this was completely broken, since it meant
dbus_connection_open() returned a connection of unknown
shared-ness. Now, we always hold a ref on anything opened
as shareable.
Move the call to notify dbus-bus.c into
connection_forget_shared_unlocked, so libdbus consistently forgets
all its knowledge of a connection at once. This exposed numerous
places where things were totally broken if we dropped a ref inside
get_dispatch_status_unlocked where
connection_forget_shared_unlocked was previously, so move
connection_forget_shared_unlocked into
_dbus_connection_update_dispatch_status_and_unlock. Also move the
exit_on_disconnect here.
(shared_connections_shutdown): this assumed weak refs to the
shared connections; since we have strong refs now, the assertion
was failing and stuff was left in the hash. Fix it to close
still-open shared connections.
* bus/dispatch.c: fixup to use dbus_connection_open_private on the
debug pipe connections
* dbus/dbus-connection.c (dbus_connection_dispatch): only notify
dbus-bus.c if the closed connection is in fact shared
(_dbus_connection_close_possibly_shared): rename from
_dbus_connection_close_internal
(dbus_connection_close, dbus_connection_open,
dbus_connection_open_private): Improve docs to explain the deal
with when you should close or unref or both
* dbus/dbus-bus.c
(_dbus_bus_notify_shared_connection_disconnected_unlocked): rename
from _dbus_bus_check_connection_and_unref_unlocked and modify to
loop over all connections
* test/test-utils.c (test_connection_shutdown): don't try to close
shared connections.
* test/name-test/test-threads-init.c (main): fix warnings in here
* dbus/dbus-sysdeps.c (_dbus_abort): support DBUS_BLOCK_ON_ABORT
env variable to cause blocking waiting for gdb; drop
DBUS_PRINT_BACKTRACE and just call _dbus_print_backtrace()
unconditionally.
* configure.in: add -export-dynamic to libtool flags if assertions enabled
so _dbus_print_backtrace works.
* dbus/dbus-sysdeps-unix.c (_dbus_print_backtrace): use fprintf
instead of _dbus_verbose to print the backtrace, and diagnose lack
of -rdynamic/-export-dynamic
- Remove pending call locking todo item
- dbus_connection_open now holds hard ref. Remove todo item
- do proper locking on _dbus_bus_check_connection_and_unref
and handle DBUS_BUS_STARTER. Remove todo item
- Warn on closing of a shared connection. Remove todo item
* bus/bus.c, bus/connection.c, bus/dispatch.c, dbus/dbus-bus.c,
dbus/dbus-connection.c: Use the dbus_connection_close_internal
so we don't get the warning when closing shared connections
* test/test-service.c, test/test-shell-service.c: Applications
don't close shared connections themselves so we unref instead of
close
* test/test-utils.c (test_connection_shutdown): Close the connection
* dbus/dbus-bus.c (_dbus_bus_check_connection_and_unref): Changed to
_dbus_bus_check_connection_and_unref_unlocked since we only call this
method on a locked connection.
Make sure we call _dbus_connection_unref_unlocked instead of
dbus_connection_unref also.
Handle DBUS_BUS_STARTER correctly
* dbus/dbus-connection.c (connection_record_shared_unlocked):
Mark as shared and hard ref the connection
(connection_forget_shared_unlocked): Remove the hard ref from the
connection
(_dbus_connection_close_internal_and_unlock): New internal function
which takes a locked connection and unlocks it after closing it
(_dbus_connection_close_internal): New internal function which acts
like the origonal dbus_connection_close method by grabbing a connection
lock and calling _dbus_connection_close_internal_and_unlock
(dbus_connection_close): Public close method, warns when the app
trys to close a shared connection
dbus/dbus-server-debug-pipe.c, glib/dbus-binding-tool-glib.c
glib/dbus-glib-tool.c, glib/dbus-gparser.c, glib/dbus-gproxy.c
test/test-segfault.c, test/test-utils.c,
test/glib/test-dbus-glib.c, tools/dbus-cleanup-sockets.c
tools/dbus-launch.c, tools/dbus-tree-view.c, tools/dbus-viewer.c:
Various cleanup of dead code and compiler warnings (patch from
Kjartan Maraas <kmaraas at gnome.org>)
* glib/dbus-gmain.c: adapt to watch changes
* bus/bus.c, bus/activation.c, etc.: adjust to watch changes
* dbus/dbus-server.h: remove dbus_server_handle_watch
* dbus/dbus-connection.h: remove dbus_connection_handle_watch
* dbus/dbus-watch.c (dbus_watch_handle): change DBusWatch to work
like DBusTimeout, so we don't need dbus_connection_handle_watch
etc.
* test/test-utils.c: use dispatch status function to fix this up
* bus/connection.c (connection_watch_callback): don't dispatch
from here
(connection_timeout_callback): don't dispatch from here
(bus_connections_setup_connection): set the dispatch status function
(bus_connection_disconnected): unset it
* dbus/dbus-mainloop.c (_dbus_loop_queue_dispatch): new function
used to add a connection to be dispatched
(_dbus_loop_iterate): do the dispatching at the end of each
iteration
* dbus/dbus-connection.c
(dbus_connection_set_dispatch_status_function): new function
allowing us to fix up main loop usage
(_dbus_connection_last_unref): free all the various function
user data
(dbus_connection_dispatch): call the DispatchStatusFunction
whenever this function returns
(dbus_connection_handle_watch): call DispatchStatusFunction
(dbus_connection_send_with_reply_and_block): call DispatchStatusFunction
(reply_handler_timeout): call DispatchStatusFunction
(dbus_connection_flush): call DispatchStatusFunction
* dbus/dbus-bus.c (dbus_bus_register): fix up error handling and
a memory leak
* bus/dispatch.c (check_service_activated): fix bug in test
* dbus/dbus-mainloop.c (check_timeout): fix this up
* dbus/dbus-internals.c (_dbus_verbose_real): include PID in
verbose output so we can sort out output from different processes,
e.g. in the activation case.
* doc/dbus-specification.sgml: require that base service names
start with ':' and that the base service is created/deleted
as first and last things a connection does on the bus
* bus/dispatch.c (check_existent_service_activation): lots more
work on the activation test; it doesn't fully pass yet...
* test/test-service.c (main): fix so we don't memleak the
connection to the message bus
(filter_func): accept a message asking us to exit
