Let's start moving towards a more involved partitioning setup to
test our stuff more when using mkosi.
The root partition is generated on boot with systemd-repart.
CentOS supports neither erofs nor btrfs so we use squashfs and xfs
instead.
We also enable SecureBoot= locally for additional coverage. This
and the use of verity means users need to run `mkosi genkey` once
to generate the keys necessary to do secure boot and verity.
If we're making a /usr only image, we still want to populate /etc
fully on first boot. To make that possible, let's copy /etc to
/usr/share/factory/mkosi in a finalize script, which runs after
all changes to the image have been made. Let's also add a tmpfiles
snippet that merges /usr/share/factory/mkosi with /etc on boot to
populate /etc.
This is only useful to be able to remove generated files using rm -rf
instead of mkosi clean and has some implications for the generated
rootfs so let's not enable it by default.
This allows setting it on the kernel cmdline and having it work
automatically without having to write any dropins or such.
Also enable the option in mkosi so that we can debug the initrd
properly with a locked root account.
These are all services that valid to be run in the initrd, so let's
make sure they have the appropriate dependencies on
initrd-switch-root.target so that they are stopped when we're about
to switch root.
We already have the systemd.tty.xxx kernel cmdline arguments for
configuring tty's for services, let's make sure the term cmdline
argument applies to pid1 as well.
We default to quiet operation everywhere except for repart, where
we disable quiet and have the mkfs tools write to stdout.
We also make sure --quiet or equivalent is implemented for all mkfs
tools.
If we switch root and can't execute an init program afterwards, we're
completely stuck as we can't go back to the initramfs to start
emergency.service as it will have been completely removed by the switch
root operation.
To prevent leaving users with a completely undebuggable system, let's
at least check before we switch root whether at least one of the init
programs we might want to execute actually exist, and fail early if
none of them exists.
When running in a VM, we now support propagating the exit status
via a vsock notify socket, so drop the restrictions on propagating
an exit status when not in a container to make sure this works
properly.
The systemd-cryptenroll man page states:
Takes a comma separated list of numeric slot indexes, or the special
strings ..., or any combination of these strings or numeric
indexes, in which case all slots matching either are wiped.
but we'd allow only one special string at any given time as the value
was not ORed when assigning. So, for example, --wipe=recovery,password
would actually become --wipe=password, etc.
Let's make sure we log more of what mkosi's doing so we can debug
issues better. Note this also makes mkosi set SYSTEMD_LOG_LEVEL=debug
when running programs so we'll get all the systemd debug logging as
well.
We run the build as a regular user and create-log-dirs requires to
run as root so let's disable the option to avoid error noise during
the install phase.
Daan De Meyer
Luca Boccassi
drosdeck
Frantisek Sumsal
Yu Watanabe
dependabot[bot]