When Secure Boot is not available (unsupported or disabled), Lanzaboote
will attempt to boot kernels and initrds even when they fail the hash
verification. Previously, this would happen by falling back to use
LoadImage on the kernel, which fails if Secure Boot is available, as the
kernel is not signed.
The SecureBoot variable offers a more explicit way of checking whether
Secure Boot is available. If the firmware supports Secure Boot, it
initializes this variable to 1 if it is enabled, and to 0 if it is
disabled. Applications are not supposed to modify this variable, and in
particular, since only trusted applications are loaded when Secure Boot
is active, we can assume it is never changed to 0 or deleted if Secure
Boot is active.
Hence, we can be sure of Secure Boot being inactive if this variable is
absent or set to 0, and thus treat all hash verification errors as
non-fatal and proceed to boot arbitrary kernels and initrds (a warning
is still logged in this case). In all other cases, we treat all hash
verification failures as fatal security violations, as it must be done
in the case where Secure Boot is active (it is expected that this does
not lead to any false positives in practice, unless there are bigger
problems anyway).
This includes the options of all modules used in the evaluation, not
just the ones from `<nixpkgs/nixos>` in the local manual.
Right now this breaks with
error: attribute 'loader' missing
at /nix/store/wf59fvxch3l5s7x0pnpfv7b26q6y010x-source/nix/modules/lanzaboote.nix:26:17:
25| configurationLimit = mkOption {
26| default = config.boot.loader.systemd-boot.configurationLimit;
| ^
27| example = 120;
I'm not sure what's up with `config.boot.loader` (had the exact same
issue with `disko`), but using `defaultText` is the common workaround
for that.
The stubs on the ESP are now input-addressed, where the inputs are the
system toplevel and the public key used for signature. This way, it is
guaranteed that any stub at a given path will boot the desired system,
even in the presence of one of the two edge-cases where it was not
previously guaranteed:
* The latest generation was deleted at one point, and its generation
number was reused by a different system configuration. This is
detected because the toplevel will change.
* The secure boot signing key was rotated, so old stubs would not boot
at all any more. This is detected because the public key will change.
Avoiding these two cases will allow to skip reinstallation of stubs that
are already in place at the correct path.
Kernels and initrds on the ESP are now content-addressed. By definition,
it is impossible for two different kernels or initrds to ever end up at
the same place, even in the presence of changing initrd secrets or other
unreproducibility.
The basic advantage of this is that installing the kernel or initrd for
a generation can never break another generation. In turn, this enables
the following two improvements:
* All generations can be installed independently. In particular, the
installation can be performed in one pass, one generation at a time.
As a result, the code is significantly simplified, and memory usage
(due to the temporary files) does not grow with the number of
generations any more.
* Generations that already have their files in place on the ESP do not
need to be reinstalled. This will be taken advantage of in a
subsequent commit.
We will hard fail in case of encountering different architectures in bootspec.
This should still be compatible with cross-compiling systems in the future.
Instead of patching the derivation in-place via flags, we just have
a higher order function that takes the Rust package derivation and override it
into a Rustfmt / Clippy oriented derivation: it turns off checks and adds its
required dependencies.
This builds the stub and tool using `rustPlatform.buildRustPackage`
which features a stable Rust compiler, recent enough to support UEFI
targets.
In the future, it will rely on properly defined targets for UEFI in
nixpkgs.
A compile time feature is introduced that allows to build "fat" stubs
that can be used to build "fat" UKIs. "fat" here means that the actual
kernel and initrd are embedded in the PE binary, not only the file path
and hash. This brings us one step closer to feature partiy with
systemd-stub and thus one step closer to replacing it fully. Such a
"fat" or "real" UKI is also interesting for image-based deployments of
NixOS.
Bootspec has a mechanism called synthesis where you can synthesize
bootspecs if they are not present based on the generation link only.
This is useful for "vanilla bootspec" which does not contain any
extensions, as this is what we do right now.
If we need extensions, we can also implement our synthesis mechanism on
the top of it.
Enabling synthesis gives us the superpower to support non-bootspec
users. :-)
This commit adds settings key for configuring systemd-boot to the lanzaboot
nixos module. The are couple of the default values that are set from the usual
nixos boot.loader.systemd-boot options, they are merged with the user defined
configuration.
This commit modifies default loader.conf to boot into the latest nixos
generation by default, for when you have other operating systems installed.
Primary reason behind this PR is to allow extensible loader configuration.
Co-authored-by: Raito Bezarius <masterancpp@gmail.com>
The way the test was implemented previously did not make it fail if no
secret was appended to the initrd. Now it is implemented similary to the
initrd-secrets test in Nixpkgs and works correctly.
To minimize the number of arguments passed to `lzbt`, the loader config
is assembled outside `lzbt` and passed as a single argument.
Instead of reimplementing `consoleMode` under the `lanzaboote`
namespace, `config.loader.systemd-boot.consoleMode` is reused as is.
The test attributes and names are simplified and standardized. They now
roughly follow the same structure as the systemd-boot test in Nixpkgs.
Some comments are added and variable names changed to make it more clear
what they actually do.
The process of installing systemd-boot is "smarter" because it now
considers a a few conditions instead of doing nothing if there is a file
at the deistination path. systemd-boot is now forcibly installed (i.e.
overwriting any file at the destination) if (1) there is no file at the
destination, OR (2) a newer version of systemd-boot is available, OR (3)
the signature of the file at the destination could not be verified.
* lanzatool: add support for generations and correct naming of kerels and initrds
* test: use convert_to_esp(extract_bspec_attr(⋅)) for unsigned tests
* lanzatool: ryan is a B class engineer
Co-authored-by: nikstur@outlook.com
Julian Stecklina