This adds a tiny binary that is hooked into SSH client config via
ProxyCommand and which simply connects to an AF_UNIX or AF_VSOCK socket
of choice.
The syntax is as simple as this:
ssh unix/some/path # (this connects to AF_UNIX socket /some/path)
or:
ssh vsock/4711
I used "/" as separator of the protocol ID and the value since ":" is
already taken by SSH itself when doing sftp. And "@" is already taken
for separating the user name.
Without `--root` or `--image`, the `security` command inspects all currently
loaded service units if no unit name is specified. But with `--root` or
`--image` with `--offline=true`, the `security` command exits silently if no
unit name is specified.
Also, fixed description of `--root` and `--image` in the man page, and added
missing `--unit` option to help text.
man:
- `verify` requires an argument
- `security` does not require an argument
- `fdstore` requires an argument
- `image-policy` requires an argument
`--help` text:
- missing `image-policy` command
- `cat-config` requires NAME or PATH
This uses openssh 9.4's -W support for AF_UNIX. Unfortunately older versions
don't work with this, and I couldn#t figure a way that would work for
older versions too, would not be racy and where we'd still could keep
track of the forked off ssh process.
Unfortunately, on older versions -W will just hang (because it tries to
resolve the AF_UNIX path as regular host name), which sucks, but hopefully this
issue will go away sooner or later on its own, as distributions update.
Fedora is still stuck at 9.3 at the time of posting this (even on
Fedora), even though 9.4, 9.5, 9.6 have all already been released by
now.
Example:
varlinkctl call -j ssh:root@somehost:/run/systemd/io.systemd.Credentials io.systemd.Credentials.Encrypt '{"text":"foobar"}'
To me this is the last major basic functionality that couldn't be
configured via credentials: the network.
We do not invent any new format for this, but simply copy relevant creds
1:1 into /run/systemd/network/ to open up the full functionality of
networkd to VM hosts.
We already have specifiers that resolve to $XDG_STATE_HOME, and
$XDG_CONFIG_HOME. $XDG_DATA_HOME is in a similar vein.
It allows units belonging to the user service manager to correctly look
into ~/.local/share. I imagine this would be most useful inside of
condition checks (i.e. only run a service on session startup if some
data is not found in ~/.local/share) or in the inotify monitoring of a
.path unit
cryptenroll accepts only PKCS#11 URIs that match both a certificate and a private key in a token.
This patch allows users to provide a PKCS#11 URI that points to a certificate only, and makes possible to use output of some PKCS#11 tools directly.
Internally the patch changes 'type=cert' in the provided PKCS#11 URI to 'type=private' before storing in a LUKS2 header.
Fixes: #23479
This is what it is after all: encryption with a NULL key. This is more
descriptive, but also relevant since we want to use this kind of
credentials in a different context soon: for carrying pcrlock data into
a UKI. In that case we don#t want encryption, since the pcrlock data is
intended to help unlocking secrets, hence should not be a secret itself.
This only changes the code labels and the way this is labelled in the
output. We retain compat with the old name.
session-status automatically uses "auto" if no ID is specified,
but show-session shows the manager's properties. Let's document
these special values so that users of show-session can benefit too.
This commit introduces new D-Bus API, StartAuxiliaryScope(). It may be
used by services as part of the restart procedure. Service sends an
array of PID file descriptors corresponding to processes that are part
of the service and must continue running also after service restarts,
i.e. they haven't finished the job why they were spawned in the first
place (e.g. long running video transcoding job). Systemd creates new
scope unit for these processes and migrates them into it. Cgroup
properties of scope are copied from the service so it retains same
cgroup settings and limits as service had.
Distributions apparently only compile a subset of TPM2 drivers into the
kernel. For those not compiled it but provided as kmod we need a
synchronization point: we must wait before the first TPM2 interaction
until the driver is available and accessible.
This adds a tpm2.target unit as such a synchronization point. It's
ordered after /dev/tpmrm0, and is pulled in by a generator whenever we
detect that the kernel reported a TPM2 to exist but we have no device
for it yet.
This should solve the issue, but might create problems: if there are TPM
devices supported by firmware that we don't have Linux drivers for we'll
hang for a bit. Hence let's add a kernel cmdline switch to disable (or
alternatively force) this logic.
Fixes: #30164
Global resource (whole system or root cg's (e.g. in a container)) is
also a well-defined limit for memory and tasks, take it into account
when calculating effective limits.
Users become perplexed when they run their workload in a unit with no
explicit limits configured (moreover, listing the limit property would
even show it's infinity) but they experience unexpected resource
limitation.
The memory and pid limits come as the most visible, therefore add new
unit read-only properties:
- EffectiveMemoryMax=,
- EffectiveMemoryHigh=,
- EffectiveTasksMax=.
These properties represent the most stringent limit systemd is aware of
for the given unit -- and that is typically(*) the effective value.
Implement the properties by simply traversing all parents in the
leaf-slice tree and picking the minimum value. Note that effective
limits are thus defined even for units that don't enable explicit
accounting (because of the hierarchy).
(*) The evasive case is when systemd runs in a cgroupns and cannot
reason about outer setup. Complete solution would need kernel support.
This does what we do for system extension also for configuration
extension.
This is complicated by the fact that we previously looked for
<uki-binary>.d/*.raw for system extensions. We want to measure sysexts
and confexts to different PCRs (13 vs. 12) hence we must distinguish
them, but *.raw would match both kinds.
This commit solves this via the following mechanism: we'll load confexts
from *.confext.raw and sysexts from *.raw but will then enclude
*.confext.raw from the latter. This preserves compatibility but allows
us to somewhat reasonable distinguish both types of images.
The documentation is updated not going into this detail though, and
instead now claims that sysexts shall be *.sysext.raw and confexts
*.confext.raw even though we actually are more lenient than this. This
is simply to push people towards using the longer, more descriptive
suffixes.
I added an XML comment (<!-- … -->) about this to the docs, so that
whenever somebody notices the difference between code and docs
understands why and leaves it that way.
DocBook document model doesn't allow mixing of <refsection> with the
numbered variants (<refsect1> etc.). Therefore, any document that
included something from standard-conf.xml was invalid. Fortunately, all
the includes are at the 1st level, hence let's just change
standard-conf.xml to use <refsect1> to fix that.
The binaries are built and installed if HAVE_TPM2 is set, and ignore ENABLE_BOOTLOADER,
so do the same for the manpages.
For the sd-pcrlock case this also installs the manpage aliases for the units, which
are not installed with -Dbootloader=disabled, but there's no way to conditionalize
the aliases, so on balance it's better to have too much documentation rather than
too little.
Fixes https://github.com/systemd/systemd/issues/30588
This adds a new --background= switch that allows specifiying a
background color for the terminal while the tool runs.
It also teaches the tool when invoked as uid0 to tint the terminal in a
reddish hue when operating as root, and in a yellowish hue when
operating as any other user.
This should highlight nicely when the user is operating with elevated
privileges, or changed privileges.
This turns "systemd-run" into a multi-call binary. When invoked under
the name "uid0", then it behaves a bit more like traditional "sudo".
This mostly means defaults appropriuate for that, for example a PAM
stack, interactivity and more.
Fixes: #29199
This geneally makes sense as setting up a PAM session pretty much
defines what a login session is.
In context of #30547 this has the benefit that we can take benefit of
the SetLoginEnvironment= effect without having to set it explicitly,
thus retaining some compat of the uid0 client towards older systemd
service managers.
Users can currently pick specific versions of NIC naming, but that
does not guarantee that NIC names won't change after the kernel adds
a new sysfs attribute.
This patch allows for an allow/deny list of sysfs attributes
that could be used when composing the name.
These lists can be supplied as an hwdb entry in the form of
/etc/udev/hwdb.d/50-net-naming-allowlist.hwdb
net:naming:drvirtio_net
ID_NET_NAME_ALLOW=0
ID_NET_NAME_ALLOW_ACPI_INDEX=1
ID_NET_NAME_ALLOW_ADDR_ASSIGN_TYPE=1
ID_NET_NAME_ALLOW_ADDRESS=1
ID_NET_NAME_ALLOW_ARI_ENABLED=1
ID_NET_NAME_ALLOW_DEV_PORT=1
ID_NET_NAME_ALLOW_FUNCTION_ID=1
ID_NET_NAME_ALLOW_IFLINK=1
ID_NET_NAME_ALLOW_INDEX=1
ID_NET_NAME_ALLOW_LABEL=1
ID_NET_NAME_ALLOW_PHYS_PORT_NAME=1
ID_NET_NAME_ALLOW_TYPE=1
Since EC keys doesn't support encryption directly, we use ECDH protocol.
We generate a pair of EC keys in the same EC group, then derive a shared secret using the generated private key and the public key in the token.
The derived shared secret is used as a volume key. The generated public key is stored in the LUKS2 JSON token header area. The generated private key is erased.
To unlock a volume, we derive the shared secret with the stored public key and a private key in the token.
Co-authored-by: MkfsSion <mkfssion@mkfssion.com>
This extends what systemd-firstboot does and runs on first boots only
and either processes user records passed in via credentials to create,
or asks the user interactively to create one (only if no regular user
exists yet).
With <para><filename>…</filename></para>, we get a separate "paragraph" for
each line, i.e. entries separated by empty lines. This uses up a lot of space
and was only done because docbook makes it hard to insert a newline. In some
other places, <literallayout> was used, but then we cannot indent the source
text (because the whitespace would end up in the final page). We can get the
desired result with <simplelist>.
With <simplelist> the items are indented in roff output, but not in html
output. In some places this looks better then no indentation, and in others it
would probably be better to have no indent. But this is a minor issue and we
cannot control that.
(I didn't convert all spots. There's a bunch of other man pages which have two
lines, e.g. an executable and service file, and it doesn't matter there so
much.)
So far we created the target directory, and the source for bind mounts,
but not workdir/upperdir for overlays, so it has to be done separately
and strictly before the unit is started, which is annoying. Check the
options when creating directories, and if upper/work directories are
specified, create them.
cryptsetup reads a bunch of credentials now but we don't pass import
those in any service units yet. Let's pass through all cryptsetup
prefixed credentials to the systemd-cryptsetup@root instance.
Request with Range header like 'entries=<cursor>:' (with a colon at the end,
invalid syntax per the doc), is now rejected with error 400 Bad Request.
fix#4883
This adds a new ProtectSystem= setting that mirrors the option of the
same of services, but in a more restrictive way. If enabled will remount
/usr/ to read-only, very early at boot. Takes a special value "auto"
(which is the default) which is equivalent to true in the initrd, and
false otherwise.
Unlike the per-service option we don't support full/strict modes, but
the door is open to eventually support that too if it makes sense. It's
not entirely trivial though as we have very little mounted this early,
and hence the mechanism might not apply 1:1. Hence in this PR is a
conservative first step.
My primary goal with this is to lock down initrds a bit, since they
conceptually are mostly immutable, but they are unpacked into a mutable
tmpfs. let's tighten the screws a bit on that, and at least make /usr/
immutable.
This is particularly nice on USIs (i.e. Unified System Images, that pack
a whole OS into a UKI without transitioning out of it), such as
diskomator.
This is a fancy wrapper around "cat <<EOF", but:
- the user doesn't need to figure out the file name,
- parent directories are created automatically,
- daemon-reload is implied,
so it's a convenient way to create units or drop-ins.
Closes https://github.com/systemd/systemd/issues/21862.
I see myself frequently typing in "journalctl --file=…". That's a lot to
type. Let's add a short switch too. Unfortunately, "-f" and "-F" are
already taken, hence let's pick "-i".
The header and keyfile are necessary only for opening the device, not
for closing, so it is not necessary to deactivate the generated
cryptsetup unit when the header or keyfile backing store are removed.
This is especially useful in the case of softreboot, when the new
mount root is setup under /run/nextroot/ but we don't want to close
the cryptsetup devices for encrypted /var/ or so, and we simply
mount it directly on /run/nextroot/var/ before the soft-reboot.
This is the equivalent of RequiresMountsFor=, but adds Wants= instead
of Requires=. It will be useful for example for the autogenerated
systemd-cryptsetup units.
Fixes https://github.com/systemd/systemd/issues/11646
Let's put the section name at the beginning of each sentence. This way we
can avoid awkward constructs like "kernel is looked for in the .linux section".
Also, since the paragraph above says that this is a list of "PE sections", we
can just say "section". In other places, it is often useful to say "PE section"
to avoid ambiguity.
Also fix the off-by-one in the count of sections.
The only reason to recommend this would be if people had multiple commands
with the same name in the search path. This probably was never the best idea,
and it happens rarely anyway. Since the patch that dropped requirement for full
paths was introduced, we have dropped support for unmerged-usr and we're planning
to drop support for split-bin at some point too. Many people effectively have just
one directory in the search path, so there is even less reason to use an absolute
path. So let's recommend just using the command name, which makes the unit file
shorter and nicer to read.
An event loop is not "synchronous". So we shouldn't say that the client must
"synchrounusly poll" if we recommend an event loop. Let's just say it should
poll, and say that "sd_bus_wait()" is blocking, which is clearer and more
correct than "synchronous".
AllowList= and DenyList= filter only DHCPOFFER messages. So, if
RapidCommit= is enabled, then networkd unconditionally accepts a rapid
ACK message even if its sender is filtered out by the lists.
As AllowList=/DenyList= implemented earlier than RapidCommit=, so
enabling RapidCommit= unconditionally by default may break existing
setups that use AllowList=/DenyList=.
Let's disable RapidCommit= by default when AllowList=/DenyList= is
enabled. Still the setting can be enabled by setting explicitly even
AllowList=/DenyList= is also specified.
This footgun should at least be documented, if there's not going
to be a shortcut setting to establish the async `journalctl
--follow` equivalent.
Fixes: https://github.com/systemd/systemd/issues/2815
With the introduction of SetDisplay() method in 4885d7490b the Display
property emits a "properties changed" message every time the display is
changed using this method, so mark it appropriately.
Caught by systemd/systemd#30029:
systemd-logind[1366]: Got message type=method_call sender=:1.165 destination=org.freedesktop.login1 path=/org/freedesktop/login1/session/_310 interface=org.freedesktop.login1.Session member=SetDisplay cookie=8 reply_cookie=0 signature=s error-name=n/a error-message=n/a
systemd-logind[1366]: Assertion 'v->vtable->flags & SD_BUS_VTABLE_PROPERTY_EMITS_CHANGE || v->vtable->flags & SD_BUS_VTABLE_PROPERTY_EMITS_INVALIDATION' failed at src/libsystemd/sd-bus/bus-objects.c:2141, function emit_properties_changed_on_interface(). Ignoring.
Follow-up to 4885d7490b.
This seems to work as expected. In the issue, doubts were raised whether it
works fine with daemon-reload/daemon-reexec, and it seems to work fine.
(The property cannot be set via set-property, the dbus property is 'const'.
We could relax this, but that'd be a separate feature.)
Closes#29574.
When an interface is being reconfigured with different bridge vlan
settings, this makes old vlan IDs on the interface removed.
This also makes the PVID= setting support negative boolean value, e.g. "no",
in which case, the currently assigned PVID (typically, assigned by the
kernel when an interface is joined to a bridge) is dropped.
This feature is requested by #15291.
Note, if a .network file has no settings about bridge vlan, networkd
keeps the currently assigned vlan IDs. That's intended, to make not
break existing setups.
When a .network file has only PVID=no line in [BridgeVLAN] section, then
all assigned vlan IDs are removed.
Fixes#29975.
Closes#15291.
In systemctl-show we only show current swap if ever swapped or non-zero. This
reduces the noise on swapless systems, that would otherwise always show a swap
value that never has the chance to become non-zero. It further reduces the
noise for services that never swapped.
This is pretty low-level functionality, hence placed in systemd-analyze.
This is useful for working with systemd-cryptenroll --tpm2-device-key=,
as it acquires the SRK without requiring the full tpm2-tss tool set.
We already save it in PEM format, also store it TPM2_PUBLIC format next
to it. This is useful for usage with systemd-cryptenroll's
--tpm2-device-key= switch.
This completes/corrects the documentation for the following fields:
COREDUMP_CGROUP= - doc where wrong, actually covered COREDUMP_PROC_CGROUP=
COREDUMP_CMDLINE= → undocumented so far
COREDUMP_PROC_CGROUP= → docs where there but incorrectly assigned to COREDUMP_CGROUP=
COREDUMP_PROC_AUXV= → undocumented so far
COREDUMP_SESSION= → undocumented so far
Fixes: #29832
Instead of enrolling the local TPM to a luks volume, use the public key from a
TPM to enroll it into the luks volume. This is useful when enrolling a TPM that
is not currently accessible, for example if the TPM is located on a different
system.
Let's make kernel-install a bit easier to use:
If the kernel version is not specified, let's imply "uname -r", so that
we regnerate the entries for the current kernel.
If the kernel image is not specified let's imply using
/usr/lib/modules/$version/vmlinuz, i.e. the location distros like Fedora
drop the kernel into, which we generally recommend people to use.
If the kernel is not found there, don't try to automatically pick the
kernel path, and fail, as before.
Until now, using any form of seccomp while being unprivileged (User=)
resulted in systemd enabling no_new_privs.
There's no need for doing this because:
* We trust the filters we apply
* If User= is set and a process wants to apply a new seccomp filter, it
will need to set no_new_privs itself
An example of application that might want seccomp + !no_new_privs is a
program that wants to run as an unprivileged user but uses file
capabilities to start a web server on a privileged port while
benefitting from a restrictive seccomp profile.
We now keep the privileges needed to do seccomp before calling
enforce_user() and drop them after the seccomp filters are applied.
If the syscall filter doesn't allow the needed syscalls to drop the
privileges, we keep the previous behavior by enabling no_new_privs.
<ulinks> contents are used to generate a SEE ALSO section at the
bottom of a man page, so we need to include the title, without additional
words, in the content.
As pointed out in https://github.com/systemd/systemd/issues/29814, we need to
use phrases are are meaningful on their own, because the man page formatter
creates a list at the bottom. With <ulink>see docs</ulink>, we end up with:
NOTES:
1. see docs
https://some.url/page
2. see docs
https://some.url/page2
which is not very useful :(
Also, the text inside the tag should not include punctuation.
Python helper:
from xml_helper import xml_parse
for p in glob.glob('../man/*.xml'):
t = xml_parse(p)
ulinks = t.iterfind('.//ulink')
for ulink in ulinks:
if ulink.text is None: continue
text = ' '.join(ulink.text.split())
print(f'{p}: {text}')
Linux's Control Group v2 interfaces exposes memory.peak, which contains the
"max memory usage recorded for the cgroup and its descendants since the
creation of the cgroup."
This commit adds a new property "MemoryPeak" for units and makes "systemctl
show" display this value if it is available.
Fixes#29878.
Signed-off-by: Florian Schmaus <flo@geekplace.eu>
In mkosi, we can't use kernel-install directly but we do want to
mimick its behavior as much as possible. Let's make that easier by
making it easy to fetch data from kernel-install as JSON.
To get all the necessary data, we extend the inspect verb to also
allow passing in a kernel version and initrds, to mimick the "add"
verb. The kernel version is used to determine the "Entry Directory",
and in absence of auto-detection of kernel version in kernel-install
we have to allow users to pass it.
We also add --no-pager while we're at it.
This changes how the "extra" command line arguments passed to vmspawn
are handled.
Previously they were appended to the QEMU command line directly.
Now they are appended to the kernel command line using SMBIOS instead.
