In order to check for matching policy entries at message transfers, we
have to consider the following:
* check the currently owned names of both the sending and the receiving
peer. If the sending peer is connected via kdbus, the currently owned
names are already attached to the message. If it was originated by the
connection we're proxying for, we store the owned names in our own strv
so we can check against them.
* Walk the list of names to check which name would allow the message to
pass, and explicitly use that name as destination of the message. If the
destination is on kdbus, store both the connection's unique name and the
chosen well-known-name in the message. That way, the kernel will make sure
the supplied name is owned by the supplied unique name, at the time of
sending, and return -EREMCHG otherwise.
* Make the policy checks optional by retrieving the bus owner creds, and
when the uid matches the current user's uid and is non-null, don't check
the bus policy.
We need to figure out which of the possible names satisfied the policy,
so we cannot do the iteration in check_policy_item() but have to leave it
to the users.
Test cases amended accordingly.
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.
Catch up with some changes in kdbus.h:
* KDBUS_{ITEM,ATTACH}_CONN_NAME were renamed to
KDBUS_{ITEM,ATTACH}_CONN_DESCRIPTION, so the term 'name' is not
overloaded as much.
* The item types were re-ordered a little so they are lined up to the
order of the corresponding KDBUS_ATTACH flags
* A new item type KDBUS_ITEM_OWNED_NAME was introduced, designated to
store a struct kdbus_name in item->name. KDBUS_ITEM_NAME soley
stores data in item->str now
* Some kerneldoc fixes
We should use the data if we can (if stdin/stdout is an AF_UNIX socket),
but still work if we can't (if stdin/stdout are pipes, like in the SSH
case).
This effectively reverts 55534fb5e4
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()
In kdbus.h, the following details changed:
* All commands gained a 'kernel_flags' field to report the flags supported
by the driver. Before, this was done in the 'flags' field in a
bidirectional way, which turned out to be a problem for the code in
sd-bus, as many parts of it reuse the same ioctl struct more than once
and consider them to be owned by userspace.
* Name listings are now returned by a new struct instead of reusing struct
kdbus_cmd_name for that matter. This way, we don't add more unneeded
fields to it and make the API cleaner.
* 'conn_flags' was renamed to 'flags' in struct kdbus_cmd_hello to make
the API a bit more unified.
'GetConnectionUnixProcessID', 'GetConnectionUnixUser' and
'GetConnectionSELinuxSecurityContext' methods should return
'NameHasNoOwner' error (if chosen name is not available on bus)
with more detailed description - like dbus-1:
Could not get PID of name 'org.freedesktop.test': no such name.
Could not get UID of name 'org.freedesktop.test': no such name.
Could not get security context of name 'org.freedesktop.test': no such name.
Otherwise we have only laconic message without proper dbus error:
Error System.Error.ENXIO: No such device or address
The KDBUS_CMD_FREE ioctl now uses a struct rather than a direct pointer
to the offset to free.
The KDBUS_CMD_MSG_CANCEL ioctl has also changes, but there's no user of
it yet in systemd.
In pty.c there was both an include of our pty.h and the system installed pty.h.
The latter contains only two functions openpty and forkpty. We use neither so
I assume it was a typo and removed it. We still compile and pass all tests.
The kdbus logic name registry logic was changed to transport the actual
name to acquire, release or report in a kdbus item.
This brings the name API a little more in line with other calls, and allows
for later augmentation.
Follow that change on the systemd side.
Since b5eca3a205 we don't attempt to GC
busses anymore when unsent messages remain that keep their reference,
when they otherwise are not referenced anymore. This means that if we
explicitly want connections to go away, we need to close them.
With this change we will no do so explicitly wherver we connect to the
bus from a main program (and thus know when the bus connection should go
away), or when we create a private bus connection, that really should go
away after our use.
This fixes connection leaks in the NSS and PAM modules.
getopt is usually good at printing out a nice error message when
commandline options are invalid. It distinguishes between an unknown
option and a known option with a missing arg. It is better to let it
do its job and not use opterr=0 unless we actually want to suppress
messages. So remove opterr=0 in the few places where it wasn't really
useful.
When an error in options is encountered, we should not print a lengthy
help() and overwhelm the user, when we know precisely what is wrong
with the commandline. In addition, since help() prints to stdout, it
should not be used except when requested with -h or --help.
Also, simplify things here and there.
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.
bus-proxyd is not only the bridge between legacy dbus clients and kdbus
but is also used to access remote dbus servers via ssh. Let's make sure
it actually works for that.
Either become uid/gid of the client we have been forked for, or become
the "systemd-bus-proxy" user if the client was root. We retain
CAP_IPC_OWNER so that we can tell kdbus we are actually our own client.
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).
Previously, AddMatch/RemoveMatch calls where processed exclusively in
the proxy. That's racy however, since subscribing to a signal might not
complete before the signal is sent due to some subsequent method call.
Hence, in order to expose the same ordering guarantees as dbus1 process
the AddMatch/RemoveMatch calls from the proxy, so that they are
dispatched synchronously to all following messages, thus fixing the
race.
Ultimately, we should probabably dissolve the driver entirely into the
proxy, as it is purely a compatibility feature anyway...
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.
In trying to track down a stupid linker bug, I noticed a bunch of
memset() calls that should be using memzero() to make it more "obvious"
that the options are correct (i.e. 0 is not the length, but the data to
set). So fix up all current calls to memset(foo, 0, length) to
memzero(foo, length).
It is nicer to predefine patterns using configure time check instead of
using casts everywhere.
Since we do not need to use any flags, include "%" in the format instead
of excluding it like PRI* macros.
gdm relies on the policy to deny its own requests to not deadlock. Given
that we currently do not enforce any policy in the dbus1 compat proxy
service this means that gdm will necessarily deadlock on our systems.
To work around this, enforce a fixed policy teomporarily, until we
interpret the legacy XML policy in full.
Booh, gdm, booh, for requring this and making me waste two days on
tracking this brokenness down.
