With certain fstabs we may propagate ENXIO from the $SYSTEMD_SYSFS_CHECK
check all the way up, making fstab-generator exit with a non-zero EC and
without any helpful message, which is really confusing.
The confext concept is an extension of the existing sysext concept and
allows to extend the host's filesystem or a unit's filesystem with signed
images that add new files to the /etc/ directory using OverlayFS.
The release file that accompanies the confext images needs to be
host compatible to be able to be merged into the host /etc/ directory.
This commit checks for version compatibility between the image file and
the host file.
Adds a new image type called IMAGE_CONFEXT which is similar to IMAGE_SYSEXT but works
for the /etc/ directory instead of /usr/ and /opt/. This commit also adds the ability to
parse the release file that is present with the confext image in /etc/confext-release.d/
directory.
- Remove useless serial console configs
- Add CONFIG_SERIAL_8250_PCI so that qemu's virtconsole works properly
- Add CONFIG_DMI and CONFIG_DMI_SYSFS so that SMBIOS credentials work
- Add CONFIG_IKCONFIG and CONFIG_IKCONFIG_PROC so we can read the
current kernel config from /proc/config.gz
- Add CONFIG_IOSCHED_BFQ to get rid of a udev error on boot on Fedora
- Add CONFIG_HW_RANDOM_VIRTIO to get hw rng in qemu
- Add CONFIG_SECURITY_YAMA to get rid of warning from sysctl on boot
- Add CONFIG_VSOCKETS and CONFIG_VIRTIO_VSOCKETS to enable vsock support
This reverts commit be266f49d6b7e3f021e2a07f937d39d1f14a283d.
Turns out we installed this because it's required by the bpf selftests
so let's keep it intact.
We have some missing coverage in the CI, all builds enable these features,
but there are often changes and they cover a lot of code. Do one build
without them to ensure we don't break builds.
Fixes:
- Comment style
- Alignment style
- cleanup macro usage
- incorrect error message[1]
1. Thanks to tempusfugit991@gmail.com for pointing out the error
message mistake.
Signed-off-by: William Roberts <william.c.roberts@intel.com>
If a root directory is specified, and e.g. /var under the root directory
is a symlink to the host's /var, then we wrongly read host's machine ID,
even if O_NOFOLLOW is set.
Let's chase the path with CHASE_NOFOLLOW to refuse such case.
Also, refuse null ID, otherwise we may setup machine ID with NULL.
Previously, when the NULL (all zero) machine ID is configured in the
container, nspawn refused to execute.
Now id128_get_machine() is used, so NULL machine ID is refused with
-ENOMEDIUM, and fallback to specified UUID or randomly generated one.
I was testing transient units and user@.service crashed. I restarted it, and
tried to create a transient unit. It failed because
/run/user/1000/systemd/transient/ remained after the previous aborted run:
Failed to start transient service unit: Unit run-u0.service was already loaded or has a fragment file.
Remove the directory during initial startup so we don't get confused by our own
files.
I was changing how some properties are appended to the StartTransientUnit call
and messed up the message contents. When something is wrong with how the
message is structed, we would return a very generic
"Failed to start transient service unit: No such device or address".
Mention that it was property setting that failed, and translate ENXIO to a
different message. bus_unit_set_properties() or any of the children it calls
may also return other errors, in particular EBADMSG or ENOMEM, but the error
message that is generated for those is understandable, so we don't need to
"translate" them explicitly.
bus_unit_set_properties() is called from two places, so it seems nicer to
generate the message internally, rather than ask the caller to do that. Also,
now bus_unit_set_properties() always sets <error>, which is nicer for the
callers.
Yu Watanabe
Frantisek Sumsal
Luca Boccassi
Daan De Meyer
maanyagoenka
Franck Bui
Ronan Pigott
Zbigniew Jędrzejewski-Szmek
Lennart Poettering
William Roberts
Thomas Blume