Old libdbus has a feature that the process is terminated whenever the the bus
connection receives a disconnect. This is pretty useful on desktop apps (where
a disconnect indicates session termination), as well as on command line apps
(where we really shouldn't stay hanging in most cases if dbus daemon goes
down).
Add a similar feature to sd-bus, but make it opt-in rather than opt-out, like
it is on libdbus. Also, if the bus is attached to an event loop just exit the
event loop rather than the the whole process.
If the server side kicks us from the bus, from our view no names are on the bus
anymore, hence let's make sure to dispatch all tracking objects immediately.
The macro determines the right length of a AF_UNIX "struct sockaddr_un" to pass to
connect() or bind(). It automatically figures out if the socket refers to an
abstract namespace socket, or a socket in the file system, and properly handles
the full length of the path field.
This macro is not only safer, but also simpler to use, than the usual
offsetof() + strlen() logic.
Throughout the tree there's spurious use of spaces separating ++ and --
operators from their respective operands. Make ++ and -- operator
consistent with the majority of existing uses; discard the spaces.
GLIB has recently started to officially support the gcc cleanup
attribute in its public API, hence let's do the same for our APIs.
With this patch we'll define an xyz_unrefp() call for each public
xyz_unref() call, to make it easy to use inside a
__attribute__((cleanup())) expression. Then, all code is ported over to
make use of this.
The new calls are also documented in the man pages, with examples how to
use them (well, I only added docs where the _unref() call itself already
had docs, and the examples, only cover sd_bus_unrefp() and
sd_event_unrefp()).
This also renames sd_lldp_free() to sd_lldp_unref(), since that's how we
tend to call our destructors these days.
Note that this defines no public macro that wraps gcc's attribute and
makes it easier to use. While I think it's our duty in the library to
make our stuff easy to use, I figure it's not our duty to make gcc's own
features easy to use on its own. Most likely, client code which wants to
make use of this should define its own:
#define _cleanup_(function) __attribute__((cleanup(function)))
Or similar, to make the gcc feature easier to use.
Making this logic public has the benefit that we can remove three header
files whose only purpose was to define these functions internally.
See #2008.
There are more than enough calls doing string manipulations to deserve
its own files, hence do something about it.
This patch also sorts the #include blocks of all files that needed to be
updated, according to the sorting suggestions from CODING_STYLE. Since
pretty much every file needs our string manipulation functions this
effectively means that most files have sorted #include blocks now.
Also touches a few unrelated include files.
This introduces two new helpers alongside sd_bus_path_{encode,decode}(),
which work similarly to their counterparts, but accept a format-string as
input. This allows encoding and decoding multiple labels of a format
string at the same time.
If code enqueues a message on one of the default busses, but doesn't
sync on it, and immediately drops the reference to the bus again, it
will stay queued and consume memory. Intrdouce a new call
sd_bus_default_flush_close() that can be invoked at the end of programs
(or threads) and flushes out all unsent messages on any of the default
busses.
Whenever we run in a user context, sd_bus_{default_user,open_user}() and
friends should always connect to the user-bus of the current context,
instead of deriving the uid from getuid(). This allows us running
programs via sudo/su, without the nasty side-effect of accidentally
connecting to the root user-bus.
This patch enforces the idea of making su/sudo *not* opening sessions by
default. That is, all they do is raising privileges, but keeping
everything set as before. You can still use su/sudo to open real sessions
by requesting a login-session (or loading pam_systemd otherwise).
However, in this case XDG_RUNTIME_DIR= will not be set (as usual in these
cases), hence, you will not be able to connect to *any* user-bus.
Long story short: With this patch applied, both:
- ./busctl --user
- sudo ./busctl --user
..will successfully connect to the user-bus of the local user.
Fixes#390.
We should never connect to the host bus as fallback if connecting to a
container failed via one method. Otherwise connecting to a dbus1
container will always result in a connection to the host.
We should not fall back to dbus-1 and connect to the proxy when kdbus
returns an error that indicates that kdbus is running but just does not
accept new connections because of quota limits or something similar.
Based on a patch by Kay.
This reverts commit d4d00020d6. The idea of
the commit is broken and needs to be reworked. We really cannot reduce
the bus-addresses to a single address. We always will have systemd with
native clients and legacy clients at the same time, so we also need both
addresses at the same time.
As it turns out machine_name_is_valid() does the exact same thing as
hostname_is_valid() these days, as it just invoked that and checked the
name length was < 64. However, hostname_is_valid() checks the length
against HOST_NAME_MAX anyway (which is 64 on Linux), hence any
additional check is redundant.
We hence replace machine_name_is_valid() by a macro that simply maps it
to hostname_is_valid() but sets the allow_trailing_dot parameter to
false. We also move this this call to hostname-util.h, to the same place
as the hostname_is_valid() declaration.
Whenever one of our calls is invoked with a non-NULL, writable
sd_bus_error parameter, let's fill in some valid error on failure. We
previously only filled in remote errors, but never local errors, which is
hard to handle by users. Hence, let's clean this up to always fill in
the error.
This introduces a new bus_assert_return() macro that works like
assert_return() but optionally also initializes a bus_error struct.
Fixes#224.
Based on a patch by Umut Tezduyar.
We should not fall back to dbus-1 and connect to the proxy when kdbus
returns an error that indicates that kdbus is running but just does not
accept new connections because of quota limits or something similar.
Using is_kdbus_available() in libsystemd/ requires it to move from
shared/ to libsystemd/.
Based on a patch from David Herrmann:
https://github.com/systemd/systemd/pull/886
sd_bus_flush_close_unref() is a call that simply combines sd_bus_flush()
(which writes all unwritten messages out) + sd_bus_close() (which
terminates the connection, releasing all unread messages) +
sd_bus_unref() (which frees the connection).
The combination of this call is used pretty frequently in systemd tools
right before exiting, and should also be relevant for most external
clients, and is hence useful to cover in a call of its own.
Previously the combination of the three calls was already done in the
_cleanup_bus_close_unref_ macro, but this was only available internally.
Also see #327
In kdbus we still have to support org.freedesktop.DBus matches even though
there is no real bus driver. The reason is that bus-control.c turns
NameOwnerChanged matches into proper kdbus matches. If we drop DRIVER
matches early, we will never match on name-changes for kdbus.
Two ways to fix this:
1) Install DRIVER matches on kdbus (which is the simple way our and which
is what this patch does).
2) Properly fix the scope-detection to let NameOwnerChanged matches
through (or better: block anything with Member!=NameOwnerChanged).
./configure --enable/disable-kdbus can be used to set the default
behavior regarding kdbus.
If no kdbus kernel support is available, dbus-dameon will be used.
With --enable-kdbus, the kernel command line option "kdbus=0" can
be used to disable kdbus.
With --disable-kdbus, the kernel command line option "kdbus=1" is
required to enable kdbus support.
Matches that can only match against messages from the
org.freedesktop.DBus.Local service (or the local interfaces or path)
should never be installed server side, suppress them hence.
Similar, on kdbus matches that can only match driver messages shouldn't
be passed to the kernel.
This appears to be the right time to do it for SOCK_STREAM
unix sockets.
Also: condition bus_get_owner_creds_dbus1 was reversed. Split
it out to a separate variable for clarity and fix.
https://bugzilla.redhat.com/show_bug.cgi?id=1224211
With the v221 release these APIs should be public, stable APIs, hence
let's install their headers by default now, and add their symbols to the
.sym file.
If NULL is specified for the bus it is now automatically derived from
the passed in message.
This commit also changes a number of invocations of sd_bus_send() to
make use of this.
This should simplify the prototype a bit. The bus parameter is redundant
in most cases, and in the few where it matters it can be derived from
the message via sd_bus_message_get_bus().
If we receive an sd_bus_message from the kernel, m->kdbus will contain
additional items that cannot be used when sending a message. Therefore,
always remarshal the message if it is used again.
This patch removes includes that are not used. The removals were found with
include-what-you-use which checks if any of the symbols from a header is
in use.
Most of our client tools want to set this bit for all their method
calls, even though it defaults to off in sd-bus, and rightfully so.
Hence, to simplify thing, introduce a per sd_bus-object flag that sets
the default value for all messages created on the connection.
include-what-you-use automatically does this and it makes finding
unnecessary harder to spot. The only content of poll.h is a include
of sys/poll.h so should be harmless.
After all it is now much more like strjoin() than strappend(). At the
same time, add support for NULL sentinels, even if they are normally not
necessary.
Pretty much everywhere else we use the generic term "machine" when
referring to containers in API, so let's do though in sd-bus too. In
particular, since the concept of a "container" exists in sd-bus too, but
as part of the marshalling system.
The ID returned really doesn't identify the owner, but the bus instance,
hence fix this misnaming.
Also, update "busctl status" to show the ID in its output.
Also:
- adds support for euid, suid, fsuid, egid, sgid, fsgid fields.
- makes augmentation of creds with data from /proc explicitly
controllable to give apps better control over this, given that this is
racy.
- enables augmentation for kdbus connections (previously we only did it
for dbus1). This is useful since with recent kdbus versions it is
possible for clients to control the metadata they want to send.
- changes sd_bus_query_sender_privilege() to take the euid of the client
into consideration, if known
- when we don't have permissions to read augmentation data from /proc,
don't fail, just don't add the data in
kdbus has seen a larger update than expected lately, most notably with
kdbusfs, a file system to expose the kdbus control files:
* Each time a file system of this type is mounted, a new kdbus
domain is created.
* The layout inside each mount point is the same as before, except
that domains are not hierarchically nested anymore.
* Domains are therefore also unnamed now.
* Unmounting a kdbusfs will automatically also detroy the
associated domain.
* Hence, the action of creating a kdbus domain is now as
privileged as mounting a filesystem.
* This way, we can get around creating dev nodes for everything,
which is last but not least something that is not limited by
20-bit minor numbers.
The kdbus specific bits in nspawn have all been dropped now, as nspawn
can rely on the container OS to set up its own kdbus domain, simply by
mounting a new instance.
A new set of mounts has been added to mount things *after* the kernel
modules have been loaded. For now, only kdbus is in this set, which is
invoked with mount_setup_late().
In kdbus a "server id" is mostly a misnomer, as there isn't any "server"
involved anymore. Let's rename this to "owner" id hence, since it is an
ID that is picked by the owner of a bus or direct connection. This
matches nicely the sd_bus_get_owner_creds() call we already have.
This reverts commit b1543c4c93.
We check b->address anyway, no need to check the return value,
especially given that the other #ifdef branch don't get the same
treatment.
kdbus learned a new ioctl to tell userspace about a bus creator's
credentials, which is what we need to implement sd_bus_get_owner_creds() for
kdbus.
Move the function from sd-bus.c to bus-control.c to be able to reuse
the bus_populate_creds_from_items() helper.
Clean up the function namespace by renaming the following:
sd_bus_get_owner_uid() → sd_bus_get_name_creds_uid()
sd_bus_get_owner_machine_id() → sd_bus_get_name_machine_id()
sd_bus_get_peer_creds() → sd_bus_get_owner_creds()
sd_bus_message_get_errno can currently return either a number of
different poitive errno values (from bus-error-mapping), or a negative
EINVAL if passed null as parameter.
The check for null parameter was introduced in 40ca29a137
at the same as the function was renamed from bus_message_to_errno and
made public API. Before becoming public the function used to return
only negative values.
It is weird to have a function return both positive and negative errno
and it generally looks like a mistake. The function is guarded by the
--enable-kdbus flags so I wonder if we still have time to fix it up?
It does not have any documentation yet. However, except for a few details
it is just a convenient way to call sd_bus_error_get_errno which is documented
to return only positive errno.
This patch makes it return only positive errno and fixes up the two
calls to the function that tried to cope with both positive and negative
values.
It is redundant to store 'hash' and 'compare' function pointers in
struct Hashmap separately. The functions always comprise a pair.
Store a single pointer to struct hash_ops instead.
systemd keeps hundreds of hashmaps, so this saves a little bit of
memory.
Actually unref the buscreds when we are not going to return a
pointer to them. As when bus_creds_add_more fails we immediately
return the error code otherwise and leak the new buscreds.
Found with coverity. Fixes: CID#1237761
We must not access slot->floating after we possible dropped the last
reference to it. Fix all callback-invocations to first check
slot->floating and possible disconnect the slot, then release the last
reference.
On systems without properly setup systemd, cg_get_root_path returns
-ENOENT. This means that busctl doesn't display much information.
busctl monitor also fails whenever it intercepts messages.
This fix fakes creates a fake "/" root cgroup which lets busctl work
on such systems.
systemctl -H root@foobar:waldi
will now show a list of services running on container "waldi" on host
"foobar", using "root" for authenticating at "foobar".
Since entereing a container requires priviliges, this will only work
correctly for root logins.
When a caller drops all references to a bus and its messages while the
messages where still queue, this causes the bus to reference the
messages, and the messages to reference the bus, without anybody else
keeping a reference, which is something we so far considered a leak, and
tried to fix with a GC logic that would recognize cases like this, and
drop the reference.
This GC logic has been broken sofar, and remained unfixed. This commit
removes it altogther, replacing it with nothing. The rationale is that
simply because all refs to the bus have been dropped its queued messages
should *still* be written to the bus, even if the caller doesn't retain
any reference to either bus nor message. This means it was actually
wrong to attempt to clean up the bus in this case.
The proper way how applications should handle this is by explicitly
invoking sd_bus_close(), when they want busses to go away. This is
probably want they want to do anyway to avoid getting spurious
callbacks after they stopped using a bus.
This makes callback behaviour more like sd-event or sd-resolve, and
creates proper object for unregistering callbacks.
Taking the refernce to the slot is optional. If not taken life time of
the slot will be bound to the underlying bus object (or in the case of
an async call until the reply has been recieved).
If you allocate a message with bus==NULL and then unref the main bus,
it will free your message underneath and your program will go boom!
To fix that, we really need to figure out what the semantics for
self-references (m->bus) should be and when/where/what accesses are
actually allowed.
Same is true for the pseudo-thread-safety we employ..
In sd_bus_unref() we check for self-reference loops and destruct our
queues in case we're the only reference holders. However, we do _not_
modify our own ref-count, thus effectively causing the
message-destructions to enter with the same reference count as we did.
The only reason this doesn't cause an endless recursion (or trigger
assert(m->n_ref > 0) in sd_bus_message_unref()) is the fact that we
decrease queue-counters _before_ calling _unref(). That's not obvious at
all, so add a big fat note in bus_reset_queues() to everyone touching that
code.
This makes sure we don't mishandle if developers specificy a different
AcceptFileDescriptors= setting in .busname units then they set for the
bus connection in the activated program.
safe_close() automatically becomes a NOP when a negative fd is passed,
and returns -1 unconditionally. This makes it easy to write lines like
this:
fd = safe_close(fd);
Which will close an fd if it is open, and reset the fd variable
correctly.
By making use of this new scheme we can drop a > 200 lines of code that
was required to test for non-negative fds or to reset the closed fd
variable afterwards.
This broke hashtable lookups for the message cookies on s390x, which is
a 64bit BE machine where accessing 32bit values as 64bit and vice versa
will explode.
Also, while we are at it, be a bit more careful when dealing with the
64bit cookies we expose and the 32bit serial numbers dbus uses in its
payload.
Problem identified by Fridrich Strba.
The new calls work similarly, but enforce a that a common, fixed bus
path prefix is used.
This follows discussions with Simon McVittie on IRC that it should be a
good idea to make sure that people don't use the escaping applied here
too wildly as anything other than the last label of a bus path.
This is primarily useful for services that need to track clients which
reference certain objects they maintain, or which explicitly want to
subscribe to certain events. Something like this is done in a large
number of services, and not trivial to do. Hence, let's unify this at
one place.
This also ports over PID 1 to use this to ensure that subscriptions to
job and manager events are correctly tracked. As a side-effect this
makes sure we properly serialize and restore the track list across
daemon reexec/reload, which didn't work correctly before.
This also simplifies how we distribute messages to broadcast to the
direct busses: we only track subscriptions for the API bus and
implicitly assume that all direct busses are subscribed. This should be
a pretty OK simplification since clients connected via direct bus
connections are shortlived anyway.
This partially reverts 41a55c46ab
Some specifications we want to stay compatibility actually document
/var/run, not /run, and we should stay compatible with that. In order to
make sure our D-Bus implementation works on any system, regardless if
running systemd or not, we should always use /var/run which is the
only path mandated by the D-Bus spec.
Similar, glibc hardcodes the utmp location to /var/run, and this is
exposed in _UTMP_PATH in limits.h, hence let's stay in sync with this
public API, too.
We simply do not support systems where /var/run is not a symlink → /run.
Hence both are equivalent. Staying compatible with upstream
specifications hence weighs more than cleaning up superficial
appearance.
../src/shared/unit-name.c:462: error: undefined reference to 'sd_bus_label_escape'
../src/shared/unit-name.c:477: error: undefined reference to 'sd_bus_label_unescape'
collect2: error: ld returned 1 exit status
Previously the returned object of constructor functions where sometimes
returned as last, sometimes as first and sometimes as second parameter.
Let's clean this up a bit. Here are the new rules:
1. The object the new object is derived from is put first, if there is any
2. The object we are creating will be returned in the next arguments
3. This is followed by any additional arguments
Rationale:
For functions that operate on an object we always put that object first.
Constructors should probably not be too different in this regard. Also,
if the additional parameters might want to use varargs which suggests to
put them last.
Note that this new scheme only applies to constructor functions, not to
all other functions. We do give a lot of freedom for those.
Note that this commit only changes the order of the new functions we
added, for old ones we accept the wrong order and leave it like that.
Add new calls sd_bus_open() and sd_bus_default() for connecting to the
starter bus a service was invoked for, or -- if the process is not a
bus-activated service -- the appropriate bus for the scope the process
has been started in.
Zbigniew Jędrzejewski-Szmek