We have to supply it when using fu_device_bind_driver(), so we need to get the
old value for ->attach() to avoid guessing what driver was previously loaded.
This restores compatibility when running with a new daemon and old remote files
and properly fixes all combinations of the regression casued by the commit
2f49da7f4e which appeared in the 1.5.2 release.
We can't do this in the library as Ubuntu want to ship a new libfwupd with an
old daemon. The new daemon only understands jcat and does not know how to
determine the age of an .asc file.
This fixes a regression from 2f49da7f4e where we
tried to parse the GPG .asc file as a JcatFile even although the remote had
Keyring=gpg (which is the default for ODM accounts on the LVFS) which returned
the cryptic 'Invalid compressed data' message.
The workaround for 1.5.2 is to change the local remote.conf from 'Keyring=gpg'
to 'Keyring=jcat' and to save the file.
The plugin currently supports Fastboot and QMI/PDC methods, but given
that QMI/PDC is exclusively used for carrier config installations,
there is no need for QMI/PDC-only updates, so make sure that
combination is right away forbidden.
This validation will also be useful when adding additional update
methods, as we'll be able to clearly specify which are the
combinations expcted.
The upgrade process requires that the USB interface layout is not
changed between upgrades, something that we require for AT and we also
require now for QMI. We store the USB interface number of the QMI port
as soon as the device is probed, and we use it to match the interface
number once the Udev based device is created.
squash! modem-manager: make sure the correct interface number is used for QMI
The built-in udev monitoring in the ModemManager plugin is only
required for those upgrade methods where the device switches into a
fully different port/interface layout. So far, we only expect this to
happen when the Fastboot update method is requested by the module.
The GUdevClient is created when the ModemManager sysfs path is
inhibited; we need to make sure we destroy the object once
uninhibited, or we'll otherwise create a new one if we're running
an additional upgrade operation afterwards.
This would allow us to add other component types in the future, for instance a
'generic' type that adds information to the composite device.
Any generic components would need to have a requirement of 1.5.2 to avoid
showing a runtime warning when trying to get the local file details.
The snap-store intends to ship an updated libfwupd library but
to use it with whatever version daemon is on the host system.
This means that the library needs to still work with older metadata
signing types.
This fixes the following error in that scenario:
```Failed to update metadata for lvfs: Keyring kind jcat not supported```
This reverts commit b5eddee5f6.
Using pci function 0 works on my Lenovo P50 but not my Lenovo X1. Don't break
machines where we are shipping the chip rather than ones where we probably are
not, especially when it's probably a kernel bug somewhere.
Fixes https://github.com/fwupd/fwupd/issues/2608
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.
If a new dependency is added but not present in the container,
`debian.sh` will abort trying to use packaging, which will cause
`ubuntu.sh` to executed. Rather than fail the build, try to
install anything missing.
This helps with the split CI where the containers are re-generated
nightly.
