Saving the quirks in the GResource section worked well, but it made the build
system very complicated and also meant the .data section was duplicated in
both `fwupd` and `fwupdtool` -- negating a lot of the hard-fought savings.
Simplify this feature so that we just `cat` all the quirk files together, then
gzip them into a single file. This means that at startup fwupd only needs to
check the mtime of one file, and weirdly it's actually faster to load a smaller
compressed file from disk that it is to load multiple uncompressed files.
tristate features will automatically disable if dependencies marked
as required are missing.
Packagers can manually override using `auto_features`.
Link: https://mesonbuild.com/Build-options.html#features
Until gi-docgen is declared stable support either of them.
This effectively means that hand builds and CI builds will use
gi-docgen, but distro builds use gtk-doc-tools.
We happily fallback with an unset -Defi_os_dir and it gets confusing explaining
why the EFI dir needs to be set on a non-supported system.
In practice they'll probably have the same eventual value on most supported
distributions.
Fixes https://github.com/fwupd/fwupd/issues/2873
The Secure Boot Advanced Targeting (SBAT) [0] is a Generation Number Based
Revocation mechanism that is meant to replace the DBX revocation file list.
Binaries must contain a .sbat data section that has a set entries, each of
them consisting of UTF-8 strings as comma separated values. Allow to embed
this information into the fwupd EFI binary at build time.
The SBAT metadata must contain at least two entries. One that defines the
SBAT version used and another one that defines the component generation.
Downstream users can add additional entries if have changes that make them
diverge from the upstream code and potentially add other vulnerabilities.
[0]: https://github.com/rhboot/shim/blob/sbat/SBAT.md
The former drags on glib-networking and then gsettings-desktop-schemas, which
add over 5Mb to the minimal IoT and CoreOS composes. Everything already uses
libcurl (even NetworkManager!) and so this is an easy way to reduce image size.
Getting the version string from git means the commit version changes each time
we commit any patch, which means we need to use --force to install firmware
when building fwupd against a version that should be compatible.
It is also very inconvenient not bumping the release version for git snapshots
as firmware can no longer depend on the "planned" release triplet.
tl;dr: A good idea for Flashrom, not so awesome for me.
This is inspired by a change in flashrom to read the version string for meson
dynamically.
No need for "post release version bump", this happens automatically from git
now by there being a dirty commit.
- Moved version discovery routine to PKGBUILD
- Set PKGEXT to .pkg.tar to avoid the package being compressed
- Added --needed to pacman arguments when installing the dependencies to
avoid reinstalling packages
Signed-off-by: Filipe Laíns <lains@archlinux.org>
Using the library instead of the command line tools provides a more
stable interface. This implementation only fetches PCR 0 for all
available hash algorithms since this is the only PCR that is actually
used in fwupd.
This also means we now include a flashrom subproject as no distro currently has
a flashrom new enough to build the plugin.
Signed-off-by: Richard Hughes <richard@hughsie.com>
Signed-off-by: Artur Raglis <artur.raglis@3mdeb.com>
Signed-off-by: Maciej Pijanowski <maciej.pijanowski@3mdeb.com>
Some firmware has a different on-device checksum to the hash of the firmware
file itself. This may be because:
* The content is not a binary file, e.g. Intel HEX or SREC
* Only part of the firmware is flashed, e.g. ignoring the bootloader section
* The device checksum is calculated using another method entirely, e.g. PCR0
It's also made complicated as there may be more than one 'correct' device
checksum in some cases, but nothing that a union query can't solve.
We can't actually access the UEFI ROM from userspace, but the PCR0 is a hash
built from the ROM itself. We could use this value to ensure the firmware has
been written correctly, and that the PCR0 matches the expected value specified
in the metadata.
The libxmlb library is much faster to query, and does not require the daemon
to parse the XML metadata at startup. It's a zero-copy mmap design that is more
modern and less clunky.
RSS has reduced from 3Mb (peak 3.61Mb) to 1Mb (peak 1.07Mb) and the startup
time has gone from 280ms to 250ms.
In Fedora the only user of libfwupdate is fwupd and the fwupdate command line
tool. It makes sense to absorb the libfwupdate library interface into the
uefi plugin in fwupd. Benefits I can see include:
* fwupd and fwupdate are very similar names; a lot of OEMs are confused
* fwupd already depends on efivar for other things
* We are maintaining an artificial library interface
* The CI and translation hooks are already in place for fwupd
* We don't need to check for features or versions in fwupd, we can just develop
the feature (e.g. BGRT) all in one place.
Requiring colord to be built before fwupd makes it hard to build packages.
The HID-based flashing protocol is stable and documented, so there's no need
to use an external library for this now.
There have been some problems with changes in systemd that will require
other changes to be made in umockdev. These changes are being sorted
out upstream. For now this makes Travis CI fail every time on the Arch
job. Until we know that's fixed upstream, disabling the test suite in
Arch at least makes CI useful again.
Switching to build in -werror in 4016c839a8
was not working properly. Adjust the CI builds to use --werror instead.
Also while doing this, enable -werror in clang builds.
It's not super useful. If this has no bad effects for a couple of releases we
can either remove it completely or move the functionality to the test plugin.
* Use sed to adjust package version while running script rather than `makepkg`
* Run script as root so it can modify `/build`
* Run `makepkg` as nobody rather than "build"
* Separate build/install steps
