This new setting allows unsharing the pid namespace in a unit. Because
you have to fork to get a process into a pid namespace, we fork in
systemd-executor to get into the new pid namespace. The parent then
sends the pid of the child process back to the manager and exits while
the child process continues on with the rest of exec_invoke() and then
executes the actual payload.
Communicating the child pid is done via a new pidref socket pair that is
set up on manager startup.
We unshare the PID namespace right before the mount namespace so we
mount procfs correctly. Note PrivatePIDs=yes always implies MountAPIVFS=yes
to mount procfs.
When running unprivileged in a user session, user namespace is set up first
to allow for PID namespace to be unshared. However, when running in
privileged mode, we unshare the user namespace last to ensure the user
namespace does not own the PID namespace and cannot break out of the sandbox.
Note we disallow Type=forking services from using PrivatePIDs=yes since the
init proess inside the PID namespace must not exit for other processes in
the namespace to exist.
Note Daan De Meyer did the original work for this commit with Ryan Wilson
addressing follow-ups.
Co-authored-by: Daan De Meyer <daan.j.demeyer@gmail.com>
Setting this flag is a noop without a corresponding call to
posix_spawnattr_setsigdefault.
If we call posix_spawnattr_setsigdefault with a full signal set,
it causes glibc's posix_spawn implementation to call sigaction 63 times,
once for each signal. That seems wasteful.
This feature is really only useful for signals which have their
disposition set to SIG_IGN. Otherwise the dispostion gets set to
SIG_DFL automatically, either by clone(CLONE_CLEAR_SIGHAND) or the
subsequent execve.
As far as I can tell, systemd does not have any signals set to SIG_IGN
under normal operating conditions.
Follow-up for 7ac58157ca
With the mentioned commit, iff E2BIG we'd retry pidfd_spawn()
with POSIX_SPAWN_SETCGROUP disabled. However, the same strategy
should actually apply to EOPNOTSUPP/ENOSYS/EPERM too -
they can mean two things here: no clone3() or no CLONE_PIDFD.
Therefore, let's first try clone() + CLONE_PIDFD, and fall further back
to plain clone() (posix_spawn()) only as last resort. Plus, record
the fact so that we don't unnecessarily retry every single time
if CLONE_PIDFD is the one that's unavailable.
In some kernels (specifically, 5.4) even though the clone3 syscall is
supported, setting CLONE_INTO_CGROUP is not. The error message returned
in this case is E2BIG.
If posix_spawn_wrapper encounters this error, it does not retry, and
cannot spawn any programs in said kernels.
This commit adds a check for the E2BIG error and retries pidfd_spawn()
without the POSIX_SPAWN_SETCGROUP flag.
If we encounter an E2BIG error, and the pidfd_spawn() succeeds after
removing the POSIX_SPAWN_SETCGROUP flag, then we cache the result so
that we do not retry every time.
Originally, this issue was reported in https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=1077204.
Signed-off-by: Kornilios Kourtis <kornilios@gmail.com>
Before this commit, the "Cannot raise nice level" branch
is rather confusing, as we're actually lowering the nice.
Also, it's better to log about the final nice value
for both cases, no matter whether we need to set to limit
or not.
Prompted by #32259
We already have this check in exec_invoke(), i.e. child.
But if CLONE_INTO_CGROUP is used, the failure would
occur on parent's side, so do the check there too.
This is useful for situations where an array of FDs is to be passed into
a child process (i.e. by passing it through safe_fork). This function
can be called in the child (before calling exec) to pack the FDs to all
be next to each-other starting from SD_LISTEN_FDS_START (i.e. 3)
The personality() syscall returns a 32-bit value where the top three
bytes are reserved for flags that emulate historical or architectural
quirks, and only the least significant byte reflects the actual
personality we're interested in (in opinionated_personality()).
Use the newly defined mask in the corresponding test as well, otherwise
the test fails on some more "exotic" architectures that set some of the
"quirk" flags:
~# uname -m
armv7l
~# build/test-seccomp
...
/* test_lock_personality */
current personality=0x0
safe_personality(PERSONALITY_INVALID)=0x800000
Assertion '(unsigned long) safe_personality(current) == current' failed at src/test/test-seccomp.c:970, function test_lock_personality(). Aborting.
lockpersonalityseccomp terminated by signal ABRT.
Assertion 'wait_for_terminate_and_check("lockpersonalityseccomp", pid, WAIT_LOG) == EXIT_SUCCESS' failed at src/test/test-seccomp.c:996, function test_lock_personality(). Aborting.
Aborted (core dumped)
See: personality(2) and comments in sys/personality.h
posix_spawnattr_setflags() doesn't OR the input to the current set of flags,
it overwrites them, so we are currently losing POSIX_SPAWN_SETSIGDEF.
Follow-up for: 6ecdfe7d10
This combines safe_fork() with pidref_set_pid().
Eventually we really should switch this to use CLONE_PIDFD, but as that
is not wrapped by glibc yet, it's hard. But this is not crucial anyway,
as a child we just forked off can always safely be referenced also by
PID, given the reaping is under our own control.
A simple test case is added in a follow-up commit.
Sometimes it makes sense to hard kill a client if we die. Let's hence
add a third FORK_DEATHSIG flag for this purpose: FORK_DEATHSIG_SIGKILL.
To make things less confusing this also renames FORK_DEATHSIG to
FORK_DEATHSIG_SIGTERM to make clear it sends SIGTERM. We already had
FORK_DEATHSIG_SIGINT, hence this makes things nicely symmetric.
A bunch of users are switched over for FORK_DEATHSIG_SIGKILL where we
know it's safe to abort things abruptly. This should make some kernel
cases more robust, since we cannot get confused by signal masks or such.
While we are at it, also fix a bunch of bugs where we didn't take
FORK_DEATHSIG_SIGINT into account in safe_fork()
For a given PID and namespace type, this helper function gives the PID
of the leader of the namespace containing the given PID. Use this in
systemd-coredump instead of using the existing get_mount_namespace_leader.
This helper will be used again in a later commit.
glibc does not provide clone() on ia64, only clone2. But only as a
symbol in the shared library, there's no prototype in the gblic
headers, so we have to define it, copied from the manpage.
This wraps glibc's clone() but deals with the 'stack' parameter in a
sensible way. Only supports invocations without CLONE_VM, i.e. when
child is a CoW copy of parent.
Let's be more accurate about what this function does: it checks whether
the underlying reported inode is the same. Internally, this already uses
a better named stat_inode_same() call, hence let's similarly name the
wrapping function following the same logic.
Similar for files_same_at() and path_equal_or_same_files().
No code changes, just some renaming.
