This snippet was already attempting to create /var/lib/dbus/machine-id,
but would fail on volatile or stateless systems where /var/lib/dbus/
did not already exist. systemd-tmpfiles automatically creates parent
directories for tmpfiles of type 'd', 'D', etc., but not for files
or symlinks (https://github.com/systemd/systemd/issues/7853).
Signed-off-by: Chris Lesiak <chris.lesiak@licor.com>
[smcv: Extended commit message to clarify why we need this]
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=104577
Reviewed-by: Simon McVittie <smcv@collabora.com>
This lets ordinary users create a limited number of app-containers
on the system bus.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101354
These will be enforced in subsequent commits.
Reviewed-by: Philip Withnall <withnall@endlessm.com>
[smcv: Fix whitespace]
Signed-off-by: Simon McVittie <smcv@collabora.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101354
These are debugging interfaces, which are essentially read-only.
By default, Verbose is not available on the system bus at all and
Stats is only available to uid 0, but both are available on the
session bus, and they can be allowed for other uids by configuring
the system bus.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101354
We can relax AddServer() from PRIVILEGED to NOT_CONTAINERS when we've
put resource limits in place, although for now it must remain
PRIVILEGED because it uses up resources.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101354
Even if the uid matches, a contained app shouldn't count as the owner
of the bus.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101354
On the system bus, that would be a denial of service, assuming we
relax the access-control from METHOD_FLAG_PRIVILEGED to a new
METHOD_FLAG_NOT_CONTAINERS later.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101354
We will eventually want to have other ways to signal that a
container server should stop listening, so that the container manager
doesn't have to stay on D-Bus (fd-passing the read end of a pipe
whose write end will be closed by the container manager has been
suggested as easier to deal with for Flatpak/Bubblewrap), but for
now we're doing the simplest possible thing.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101354
We will need this to be able to shut down the container when its
creator vanishes.
Reviewed-by: Philip Withnall <withnall@endlessm.com>
[smcv: Fix minor conflict]
Signed-off-by: Simon McVittie <smcv@collabora.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101354
If we're strict now, we can relax this later (either with a named
parameter or always); but if we're lenient now, we'll be stuck with it
forever, so be strict.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101354
This means we can accept connections on the new socket. For now, we
don't process them and they get closed.
For the system bus (or root's session bus, where the difference is
harmless but makes automated testing easier), rely on system-wide
infrastructure to create /run/dbus/containers. The upstream dbus
distribution no longer contains integration glue for non-systemd boot
systems, but downstreams that maintain a non-systemd boot system and are
interested in the Containers interface should create /run/dbus/containers
during boot.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101354
We still don't actually create a DBusServer for incoming connections
at this point, much less accept incoming connections.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101354
In particular, we now fail early if we can't extract the file
descriptor, or if there are named parameters (none are supported yet).
Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101354
For now, this is considered to be a privileged operation, because the
resource-limiting isn't wired up yet. It only contains the bare minimum
of API.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101354
This lets us see which bits are painfully slow. (Spoilers:
check_existent_service_no_auto_start.)
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Signed-off-by: Simon McVittie <smcv@collabora.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=103601
The echo service frequently fails to connect to the bus when we are
testing OOM code paths, again causing a lot of noise in the log.
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Signed-off-by: Simon McVittie <smcv@collabora.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=103601
In parts of the OOM testing, our logging produces multiple megabytes
of output. Let's not do that.
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Signed-off-by: Simon McVittie <smcv@collabora.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=103601
This is a little more self-documenting - it justifies why it's
acceptable to fail hard on out-of-memory conditions. _dbus_test_fatal()
isn't compiled unless we are compiling embedded tests, so compiling
with embedded tests disabled provides reasonable confidence that we
aren't using _dbus_test_fatal() inappropriately.
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Signed-off-by: Simon McVittie <smcv@collabora.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=103601
Unlike _dbus_assert_not_reached(), this new function takes a printf-style
format string, so we don't need to use a _dbus_warn() to explain why
the failure occurred (unless the failure message is multi-line).
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Signed-off-by: Simon McVittie <smcv@collabora.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=103601
Printing to stdout would interfere with generating TAP syntax.
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Signed-off-by: Simon McVittie <smcv@collabora.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=103601
While reviewing fd.o#101354, Philip Withnall pointed out that if we
rejected a connection in the new code there, we didn't log why. It
turns out we didn't log that in the more normal code path either.
Redo the error handling so that failure to set up a connection
is logged.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=103592
If an async signal interrupts some function, we can have this
anti-pattern:
/* in normal code */
result = some_syscall (); /* fails, e.g. errno = EINVAL */
/* interrupted by async signal handler */
write (...); /* fails, e.g. errno = ENOBUFS */
/* back to normal code */
if (errno == EINVAL) /* problem! it should be but it isn't */
The solution is for signal handlers to save and restore errno.
This is unnecessary for signal handlers that can't touch errno (like
the one in dbus-launch that just sets a flag), and for signal handlers
that never return (like the one in test-utils-glib for timeouts).
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=103010
Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
LSB-style (SysV-style) init scripts have not historically been
portable between distributions, as evidenced by the presence of both
"Red Hat" and "Slackware" init scripts in dbus. Many distributors
prefer to maintain them downstream, as is done in Debian (and its
derivatives) and in Slackware, so that the init script can follow
OS conventions (for example regarding boot messages) and make use
of OS-provided facilities (for example, the Debian init script uses
dpkg's start-stop-daemon utility).
The Slackware and Red Hat init scripts removed by this commit are not
tested or maintained in practice, and so are likely to have bugs. The
Slackware init-script provided here is not used on actual Slackware
systems, which provide a different implementation of rc.messagebus in
their packaging, while the Red Hat init script has been superseded by
the systemd unit in current Fedora, CentOS and RHEL versions.
The Cgywin messagebus-config provided here does appear to be used in
production in cygwin-ports, but it's full of Cygwin-specifics with which
the dbus maintainers are not familiar, so it is probably more appropriate
for it to be tracked downstream as part of the Cygwin packaging.
The systemd unit is not removed, since it is used on multiple Linux
distributions with little or no modification, and receives regular
testing and maintenance; this makes it appropriate to maintain upstream.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Bug: https://bugs.freedesktop.org/101706
Reviewed-by: Philip Withnall <withnall@endlessm.com>
If we add a rule like
<allow send_destination="com.example" send_broadcast="true"/>
then it cannot possibly match anything, because to be a broadcast, the
message would have to have no destination. The only value of
send_destination that can be combined with send_broadcast="true" is
the wildcard "*", but by this point in the function we already
replaced "*" with NULL.
Adapted from an earlier implementation of send_broadcast by
Alban Crequy.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Bug: https://bugs.freedesktop.org/92853
Reviewed-by: Philip Withnall <withnall@endlessm.com>
<allow send_broadcast="true" ...> only matches broadcasts,
which are signals with a NULL destination. There was previously
no way for the policy language to express "NULL destination",
only "any destination".
<allow send_broadcast="false" ...> only matches non-broadcasts,
which are non-signals or signals with a non-NULL destination.
There was previously no way for the policy language to express
"any non-NULL destination", only "any destination".
Reviewed-by: Philip Withnall <withnall@endlessm.com>
[smcv: improved documentation as per Philip's review]
Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Thiago Macieira <thiago@kde.org>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=92853
The giant conditionals used to check policy attributes are increasingly
unwieldy, so let's try something else. Bundle together the send_
attributes, the receive_ attributes, the eavesdrop attribute
(which can go on either send or receive rules) and the other attributes
into equivalence classes, and write the conditionals in terms of those
equivalence classes.
In particular, this correctly forbids
<allow receive_type="..." send_destination="..."/>
which was previously allowed but nonsensical (the send part took
precedence and the receive part was ignored).
Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Reviewed-by: Thiago Macieira <thiago@kde.org>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=92853
By default, Expat uses cryptographic-quality random numbers as a salt for
its hash algorithm, and since 2.2.1 it gets them from the getrandom
syscall on Linux. That syscall refuses to return any entropy until the
kernel's CSPRNG (random pool) has been initialized. Unfortunately, this
can take as long as 40 seconds on embedded devices with few entropy
sources, which is too long: if the system dbus-daemon blocks for that
length of time, important D-Bus clients like systemd and systemd-logind
time out and fail to connect to it.
We're parsing small configuration files here, and we trust them
completely, so we don't need to defend against hash collisions: nobody
is going to be crafting them to cause pathological performance.
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101858
Tested-by: Christopher Hewitt <hewitt@ieee.org>
[smcv: Adjust build-system changes for 1.11.x]
Signed-off-by: Simon McVittie <smcv@debian.org>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
This lets _dbus_warn() and _dbus_warn_check_failed() fall through
to flushing stderr and calling _dbus_abort(), meaning that failed
checks and warnings can result in a core dump as intended.
By renaming the FATAL severity to ERROR, we ensure that any code
contributions that assumed the old semantics will fail to compile.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101568
Now that we're starting to implement methods in more places, it makes
sense to share this code. The Stats interface can already benefit.
Signed-off-by: Simon McVittie <smcv@collabora.com>
Reviewed-by: Philip Withnall <withnall@endlessm.com>
Bug: https://bugs.freedesktop.org/show_bug.cgi?id=101567
