This adds support for optionally using a UNIX domain socket where a
D-Bus daemon may not be running.
To use this, launch the daemon and clients with something like
`FWUPD_DBUS_SOCKET=/var/run/fwupd.sock fwupdmgr get-devices`
The initial name of the varianle has been changed upon review in
[nrf_desktop DFU PR](https://github.com/nrfconnect/sdk-nrf/pull/6343)
This commit align the naming with NRF desktop upstream and fix testing
images URLs for device testing.
This allows us to replace the 'dfu-tool replace' with a much more
controlled 'fwupdtool firmware-patch' command that patches at a set of
specific offsets.
Drop in the new functionality into the existing firmware ->write()
vfuncs so that the image headers, footers and/or checksums are correct.
Added the test for nRF52840-DK board with MCUBoot variant.
Images based on MCUBoot may be used for test automation since this
bootloader type does not have version dependencies.
To prepare the board for testing it should be flashed with MCUBoot
image variant with `nrfjprog` initially.
Install or downgrade firmware on all devices to make the system match
a well known set. This allows two things:
* Factory recovery where a device in the field has been upgraded
* Ensuring a consistent set of tested firmware for a specific workload
A tag is assigned either during upload or added post-upload on the LVFS
which is included in the metadata. A single firmware can be marked with
multiple tags, and tags can be duplicated for different firmwares.
This allows us to override the location we load data files from, which
allows us to do more kinds of installed tests in the future.
Also, move the global data/tests content into the place that it is used
as it was getting impossible to manage.
Now two plugins are using hardcoded SPI constants for various CFI chips,
and it makes sense to have some common quirk data that can be used by
both.
Add a FuSpiChip helper object that can be used by FuDevice subclasses
to get the specific SPI commands to use for each flash ID.
99.9999% of users are consuming firmware updates from the LVFS or
another trusted remote. It's far too easy to get a user to enter the
password to install an untrusted firmware, where the security
consequences are pretty dire.
Provide an escape-hatch for firmware engineers, but it does mean
editing a file in /etc as root. This seems like an acceptable level of
inconvenience.
This avoids clashing with real-world devices that have the same VID/PID.
Many thanks to Ricky WU <ricky_wu@realtek.com> for all the help.
Fixes: https://github.com/fwupd/fwupd/issues/3835
This functionality broke a number of releases ago as part of
implementing device inhibition and was just noticed now.
Instead of fixing it, the preference seems to be to remove the
functionality as it exists today as inhibitions can happen for
a number of reasons.
To still allow people to override these power warnings (such as during
development) add a new daemon configuration item that can be used.
Fixes: #3778
This allows us to do a few things:
* Remove the runtime dep on Python 3, which is tricky for ChromeOS
* Test composite devices more efficiently, only writing once per test
* Automatically upload signed reports for successful device tests.
