initrd-parse-etc: override argv[0] to avoid dracut issue
Quoting https://github.com/systemd/systemd/pull/24054#issuecomment-1210501631:
> this would need a patch in dracut, specifically adding the
> systemd-sysroot-fstab-check to the list of installed stuff:
> https://github.com/dracutdevs/dracut/blob/fe8fa2b0cadbb33e27c8dd8b5851548dcd65835c/modules.d/00systemd/module-setup.sh#L47.
>
> I could do this manually in the CI (and I guess I'd have to do it anyway even
> if the patch lands in upstream, since it won't be available in C8S), but it
> should get there first before merging this PR, otherwise it's going to break
> Rawhide.
Luca Boccassi [Wed, 17 Aug 2022 18:43:01 +0000 (19:43 +0100)]
test: expect libdevmapper failure in TEST-50-DISSECT
libdevmapper/device mapper driver can return semi-random failures when
opening verity devices, and we have fallback code to deal with it.
But the test was not expecting the fallback path, so it became unreliable.
For whatever reason, nowadays we add entries that are quite long, so
we were using line continuation for most of new entries. This is hard
to read and edit. So let's just make the table wide enough to accomodate
all the names without line splits.
tpm2-util: credit TPM2 RNG entropy only once per boot
Acquiring random data from the TPM is not precisely quick, let's speed
things up by doing this at most once per boot. For that, let's maintain
a flag file in /run/.
Yu Watanabe [Wed, 17 Aug 2022 05:44:27 +0000 (14:44 +0900)]
udev: rename various validate() -> should_reload() for builtin commands
Previously, true by validate() means several configs are outdated and we
need to reload configs. That's not intuitive for me. Let's rename the
functions.
Yu Watanabe [Mon, 8 Aug 2022 17:02:02 +0000 (02:02 +0900)]
backlight: fix issue on multiple graphics cards system
If a system has multiple graphics cards, then we cannot associate
platform backlight devices to backlight devices under PCI bus.
Previously, in such case, vaidate_device() for a raw backlight device
might erroneously detect a platform device and return false. So, users
could not save/load backlight level.
This makes validate_device() give up to associate platform devices on
non-PCI bus with raw backlight devices. That may cause unwanted
backlight level save or restore by systemd-backlight@.service, but users
can workaround that by masking specific instances of the service.
manager: allow assignment of properties on target/swap/device units
E.g. Documentation or Markers could apply to any unit type. This already worked
partially, because a direct dbus call could be made:
After rebuild with the patch, but before the manager has been restarted:
$ build/systemctl --user set-property dev-zram0.swap Markers=+needs-restart
$ build/systemctl --user show -p Markers dev-zram0.swap
Markers=needs-restart
I noticed that that the rpm unit restart helper was throwing errors for target
units. We should just let the Markers be set for those too, even if it doesn't
do anything in the end. This way we don't need to special-case by unit type.
Kai Lueke [Tue, 16 Aug 2022 10:29:12 +0000 (12:29 +0200)]
man: Correct information on sysext masking
While I had tested that a symlink to /dev/null works to "mask" a sysext
I must have gotten something wrong and thus the instructions in 519c2f0d6b343d140f7e08e3eb0f46708c023b4a don't work. What works,
at least at the moment, is to instead have an empty directory with the
extension name under /etc/extensions/.
Correct the info in the man page and add a test for it.
Avram Lubkin [Tue, 16 Aug 2022 12:51:21 +0000 (08:51 -0400)]
sysusers: add fsync for passwd (#24324)
https://github.com/systemd/systemd/pull/6636 added `fsync()` when
temporary shadow, group, and gshadow files are created, but it was
not added for passwd. As far as I can tell, this seems to have been
an oversight. I'm seeing real world issues where a blank /etc/passwd
file is being created if a machine loses power early in the boot process.
TPM2 knows two types of policy sessions: "real" ones and "trial" ones. The
latter allow you to calculate a policy hash without this enforcing any
policy, which the former do. Typically you want to use the "trial" ones
when enrolling, and you have to use the "real" ones for unlocking. So
far we used "real" ones for both cases. Which works fine – as long as
the policy put together matches the current reality (e.g. the PCR values
included in the policy are the ones currently in place in the TPM).
Let's switch to using trial sessions for enrolling. First of all this is
preparation for later work to implement further policy extensions (for
example, policies binding to literally specified PCR values, instead of
the once currently measured). But from my perspective more importantly
it actually is cleaner, as it communicates more clearly what we are
actually doing here.
The lifetime of variables decalred in the main switch/case body is just
weird. Let's not rely on it, but just open a new {} block, like
everywhere else in similar cases.
Kai Lueke [Tue, 9 Aug 2022 15:07:20 +0000 (17:07 +0200)]
sysext: introduce ARCHITECTURE field to match host architecture
When an extension image has binaries they should match the host
architecture. Currently there is no way to specify this requirement.
Introduce an ARCHITECTURE field in the extension's release file that
may be set to prevent loading on the wrong host architecture.
Since this new field is introduced late, we don't want to make
specifying it mandatory as it would break existing sysext images.
See https://github.com/systemd/systemd/issues/24061
Kai Lueke [Tue, 9 Aug 2022 11:34:18 +0000 (13:34 +0200)]
sysext: support distribution-independent extensions using ID=_any
A sysext image that merely contains static binaries has no dependency
on the host distribution and should be able to be used anywhere.
Support the special '_any' value for the ID field in the extension to
opt-out of ID and VERSION_ID/SYSEXT_LEVEL matching.
See https://github.com/systemd/systemd/issues/24061
Yu Watanabe [Fri, 12 Aug 2022 19:47:54 +0000 (04:47 +0900)]
dhcp6: gracefully handle NoBinding error
When we receive NoBinding status code, the requesting binding (address or
any other information) does not exist anymore in the server. Hence,
resending the request is meaningless. Let's restart the transaction from
the beginning in that case.
hwdb: Apply existing accel orientation quirk to all Chromebooks
The cros-ec-accel and cros-ec-accel-legacy kernel modules internally
correct for the board-specific accelerometer mounting orientations.
Their sensor outputs are in a standard reference frame consistent across
different boards, so the orientation matrix already added for a number
of devices should apply to every device using cros-ec accelerometers.
The different matrix for the 'Nocturne' board seems to be an error.
Replace the existing hwdb rules for select Chromebooks with generic
rules that apply to all Chromebooks.
Yu Watanabe [Thu, 11 Aug 2022 16:54:47 +0000 (01:54 +0900)]
sd-device: allow to create sd-device object through a symlink outside of /sys
For example, /proc/device-tree is a symlink to /sys/firmware/devicetree/base,
and the kernel documentation says the symlink should be used by userspace app.
See, https://www.kernel.org/doc/Documentation/ABI/testing/sysfs-firmware-ofw.
Hence, it is useful to make `sd_device_new_from_path()` support such symlink.
Sean Anderson [Tue, 9 Aug 2022 17:38:26 +0000 (13:38 -0400)]
udev: net_id: Use devicetree aliases when available
Devicetree firmware contains an "aliases" node, containing various
aliases for devices described by the firmware. For ethernet devices,
these are named "ethernet0", "ethernet1", etc. They provide a convenient
means of numbering ethernet devices, especially on systems with no other
stable number other than the address. In particular, U-Boot already uses
these aliases to name its ethernet devices.
Previously, there have been attempts (such as [1]) to add support for
these aliases to Linux. However, these patches have been rejected
because it is the maintainers' view that naming policy be left to
userspace. Well, systemd is userspace, so here we are.
In terms of implementation, apparently there can be multiple device
trees at once. I have decided to dodge this problem for now, and just
use /proc/device-tree. If it is desired to support multiple device trees
later, then the scheme can be modified to include the device tree's
index. For example, /sys/firmware/devicetree/base2/aliases/ethernet3
might be named enb2d3.
For the moment we only support "ethernetX" aliases. Future patches might
want to also handle "canX" and "wifiX".
It is common on boards with only one ethernet device to use an alias of
just "ethernet". In this case, the index is an implicit 0. In case the
author of the firmware made a mistake, we check to ensure that aliases
of "ethernet" and "ethernet0" do not both exist.