* Load container images descending by size
Loading container images using docker load seems to require more space
at load time (which gets freed after loading). Loading the largest
container first avoids running out of space.
* Disable real-time scheduling
It seems that Linux' cgroup v2 currenlty does not support RT scheduling.
* Remove Supervisor RT support flag
With CGroups v2 we can no longer support CPU resource allocation for
realtime scheduling.
* Bump OS Agent to 1.3.0 for CGroups v2 support
* Use upstream Linux driver for Bluetooth on ASUS Tinker
* Drop unnecessary Bluetooth initialization systemd service
Bluetooth is now entirely handled by the kernel.
It seems that the GitHub container registry sometimes returns 503
service unavailable temporarily ("Error fetching tags list: invalid status
code from registry 503"). Use skopeo's retry mechanism to try up to 5
times before failing.
The Google Gasket driver has been removed from the main kernels staging
tree between 5.10 and 5.15 development window. Readd Google's
out-of-tree driver to continiue support Google Coral devices.
* Replace bluetooth-bcm43xx with pi-bluetooth Buildroot package
The new pi-bluetooth packages the scripts and systemd service from
the Raspberry distribution package directly:
https://github.com/RPi-Distro/pi-bluetooth
* Update to latest pi-bluetooth service files
* Update busybox configuration to 1.35.0
The new/deleted configurations are generated automatically, no actual
change in this patch.
* Enable busybox xxd command
The xxd tool is useful for conversion in scripts.
* Prevent start erros on Compute Module 4 without WiFi/Bluetooth
* Add tempio host package
tempio is a template helper using Go's template engine and sprig
functions.
* Use tempio to generate rauc manifest
* Use tempio to generate rauc system.conf
* Linux: Update kernel 5.15.25
Use highest available kernel version in Buildroot 2021.08 (5.13)
* Update Hardkernel patches to Linux 5.15
* Update generic-x86-64/ova kernel config/patches for 5.15
* Drop Intel e1000e Sourceforge driver
The driver has been discontinued sometime last year. The main reason the
out-of-tree kernel has been enabled was for support for the i219-V
network chips which meanwhile are supported in mainline.
* updated generic_raw_uart to latest version which comes with dualcopro
support for the HmIP-RFUSB usb rf-sticks by eQ3/ELV.
* remove 99-hmip-rfusb.rules to keep a HmIP-RFUSB device free from being
occupied by the cp210x driver but use the new generic_raw_uart support
instead allowing for advanced dualcopro support for HomeMatic/BidCos-RF
and homematicIP support in parallel.
* Avoid race condition when fetching containers during build
So far only a single builder was active for each architecture. This
toghether with the naming scheme to include architecture/machine name
made sure that an image could only be fetched or used by a single
builder.
However, since most systems are now aarch64, multiple runners are now
active for a single architecture. This makes it necessary to lock
fetching/coping of container images to avoid race conditions.
Use rtl8812au driver provided by buildroot. This uses a newer verison of
the v5.6.4.2 branch which works with newer kernel and seems to be the
recommended branch.
Note: It seems that our buildroot package currently fails to properly
deploy the 88XXau.ko kernel module. Instead of fixing our version, just
move to the buildroot version.
* Bump to Buildroot 2021.08.1
Move to Buildroot 2021.08.1 using the 2021.08.x-haos branch. Some
patches on the previous branch 2021.02.x-haos have been applied upstream
meanwhile. Others required rather trivial rebasing.
This latest Buildroot release brings new versions of the following
components:
- glibc 2.33
- systemd 249.3
- Networkmanager 1.32.2
- BlueZ 5.60
- Docker 20.10.8
The patch "Fix dhcp client" seems not to be necessary anymore. The
directory /var/lib/dhcp seems not in use when NetworkManager invokes
dhclient. It seems the leases which are typically stored in that
directory are managed inside NetworkManager.
* buildroot 2021.08.1..2021.08.x-haos (6)
> package/rpi-firmware: bump version to 1.20210805
> package/rpi-wifi-firmware: bump version to 883b726
> package/linux-firmware: add rtl8761b/rtl8761bu firmware
> package/docker-proxy: bump version to 64b7a4574d14
> package/rpi-firmware: Allow to deploy multiple firmware files
> network-manager: wpa_supplicant
* Bump Raspberry Pi Bluetooth helper scripts
With the update to Buildroot 2021.08.1, the bthelper fails with an error
org.bluez.Error.Busy when trying to power off the device. Presumably this
is a race condition which surfaced due to a change in Bluez 5.60:
348feb005a
Oct 11 14:32:21 homeassistant systemd[1]: Reached target Bluetooth Support.
...
Oct 11 14:32:21 homeassistant bluetoothd[412]: Bluetooth management interface 1.18 initialized
Oct 11 14:32:21 homeassistant systemd[1]: Started Raspberry Pi bluetooth helper.
Oct 11 14:32:21 homeassistant bthelper[417]: Raspberry Pi BDADDR already set
Oct 11 14:32:21 homeassistant bthelper[426]: [58B blob data]
Oct 11 14:32:21 homeassistant bthelper[426]: [59B blob data]
Oct 11 14:32:21 homeassistant bthelper[426]: Failed to set power off: org.bluez.Error.Busy
Oct 11 14:32:21 homeassistant systemd[1]: bthelper@hci0.service: Main process exited, code=exited, status=1/FAILURE
Oct 11 14:32:21 homeassistant systemd[1]: bthelper@hci0.service: Failed with result 'exit-code'.
The latest version of the pi-bluetooth package introduced a sleep before
powering off the device, however, presumably for a different reason:
ae2efdeee8 (diff-609c8a23261988c47afd40be9b012feb1d167de8761c1301e44e1864635c19e3)
Anyways, this latest version seems to also fix the above mentioned race
condition.
Sometimes the first command after starting the Docker daemon container
fails, presumably because the container did not start yet. Wait until
the Docker daemon is ready.
* Use skopeo to download container images
Separate container download from image build. This will allow to share
the downloaded images between multiple builds.
We won't store the Supervisor container with the version tag, just with
the latest tag. This allows to simplify the procedure a bit. It seems
there is no downside to this approach.
* Use official Docker in Docker images to build data partition
Instead of building our own Debian based image let's use the official
Docker in Docker image. This avoids building an image for the hassio
data partition and speeds up build as well.
This calls mount commands using sudo to mount the data partition as part
of the buildroot build now. This is not much different from before as
mount has been called as root inside the container, essentially equates
to the same "isolation" level.
* Use image digest as part of the file name
The landing page has no version information in the tag. To avoid
potentially source caching issues, use the digest as part of the file
name.
Add buildroot utils/check-package check to the pr-checks.yml workflow.
It checks for common errors/mistakes when creating own buildroot
packages. Also fixed all warnings this utility output for our existing packages.
* Enable some useful kernel configurations
* Add xe-guest-utilities for better Xen support
Add guest utilities and make sure the Xen guest daemon gets started
when running under Xen virtualization.
* add eq3_char_loop package (eQ-3 char loopback kernel module)
* add generic_raw_uart package (low-latency raw UART kernel driver)
* add rpi-rf-mod package
* add device tree overlay support for RPI-RF-MOD/HM-MOD-RPI-PCB on Raspberry Pi
* enable GPIOLIB and GPIO_SYSFS required for RPI-RF-MOD/HM-MOD-RPI-PCB support.
* add basic RPI-RF-MOD/HM-MOD-RPI-PCB support for ASUS Tinker Board
* add device tree overlay support for ASUS Tinker Board and add
haos-config.txt loading support to U-Boot boot script
The latest version of OS Agent sets haos.wipe=1 as kernel argument to
trigger a device wipe. Let systemd pickup this kernel command line
argument and start haos-wipe.service.
This rather complex architecture allows to add other triggers in the
future, e.g. a button read in the boot loader.
* Disable systemd-logind support for udisks2
Currently udisks2 uses systemd-logind to prevent the system from
rebooting or similar operations while udisks operations are ongoing.
Unfortunately this stops us from using udisks2 during early boot since
systemd-logind is not ready at this point. Make the dependency
configureable so we can opt-out of using systemd-logind.
* Make dbus.service/socket and udisks2.service/socket available early
Disable default dependencies. This avoids those services to be ordered
after sysinit.target, and makes them available before local-fs.target
is reached. All mounts like mnt-data.mount are ordered before
local-fs.target, so breaking this dependency allows to use D-Bus before
mounting local file systems.
This seems fine when using the system bus directly from /run (instead of
/var/run, which is anyway a symlink to /run normally). It seems that
udisks misses /var/lib/udisks2 but it seems not to be required for the
features used so far.
* Add udisks2 package
Add latest release of udisks2 as a package. Also disable polkit to avoid
excessive dependencies.
* Add udisks2 and os-agent to Home Assistant OS
* Bump OS Agent to latest version with udisks support
This matches the 1.2-4+rpt8 release of Raspberry Pi OS' bluez-firmware
package. It addresses mainly addresses Spectra fix for CYW43455
(CVE-2020-10370).
Also update the Bluetooth start scripts with CM4 support and some
minor improvements.
* Improve ASUS Tinker Board support for 5.10
Remove patches which are unnecessary. Revert DMA for UART as it seems to
cause more problems (its also what Armbian is doing). With that
Bluetooth firmware seems to load without errors when loaded before the
bluetooth daemon is running!
Note: It seems that the board overheats quite quickly. With Armbian,
without load, that seems not to be a big deal, but HAOS does quite a
bunch at startup, leading the CPU to reach the 90°C trip point. Maybe it
was related to the rather closed shelf I have the ASUS Tinker board
running, but only after using a fan the board behaved for me.
* Use hardware flow control explicitly
The rtk_hciattach program uses hardware flow control by default (judging
from tty settings after starting the program). Just to be sure,
explicitly request 115200 and hardware flow control.
