- fixes: unrecognized option '-Wl,--no-warn-rwx-segment'
- add a patch to atf that removes the unrecognized linker options
Signed-off-by: Martin Schmiedel <Martin.Schmiedel@tq-group.com>
Follow conventional `if-then-else-fi` constructions by not having `else` followed by semicolon. Hush is accepting, but other commonly used shells do not accept this.
- UUID boot does not seem to work on PocketBeagle 2.
- So specify the block device path instead.
- Also without rootwait, the boot will fail.
Signed-off-by: Ayush Singh <ayush@beagleboard.org>
This is tested on a tinkerboard 2s.
The u-boot dts for tb2 contains other fixes, and the major ones are:
- add missing basic gpio-fan node
- usb-c altmode fix, dts part
- similar to e9708b8c33, but for the Blade3
- this reverts d048673c00 just for the Blade3
- there is little to be gained from .10 anyway on those machines (IMHO)
- it is still vendor 2017.09 after all, dunno why Radxa calls them "2024.x"
- UMS mode (use the black USB port)
- Boot order: SD -> NVMe -> SATA -> USB -> eMMC
- Flash LEDs in preboot
- Networking stuff
- Lots of FS's and compression methods
* helios4: fix wake-on-lan (wol)
- added ethtool package
- enable wol on all ethernet interfaces
- support common systemd.net-naming-schemes (ethX/endX/enoX)
* Update helios4-wol.service
Removed test code.
* Update helios4-wol.service
Fix bug after retesting.
From the Kconfig:
```
Preload default EDID (Extended Display Identification Data)
branded by Linux Foundation that exposes display modes up
to 4k@30Hz, which have best compatibility with HDMI transmitters.
Enabling this option is recommended for a non-production use-cases.
It will make driver usable out-of-the-box.
For a higher display modes you will need to load customized EDID
from userspace using v4l2-ctl tool or by other means.
Without enabling this option driver will be practically
non-functional until EDID will be loaded from userspace.
Which is a wanted behavior when using this driver in a
commercial product that should utilize own branded EDID.
```
On v2025.01 and v2025.04 I get the following error
starting USB...
Bus usb@ff500000: dwc3_meson_gxl_get_phys: usb2 ports: 1
probe failed, error -114
No USB controllers found
So revert back to v2024.10 until sorted.
Signed-off-by: Patrick Yavitz <pyavitz@gmail.com>
ATF is currently built in Debug mode for K3 family of devices. This is
due to the fact that K3 Armbian images have been experimental so far.
However, there is work underway to make Armbian a production-capable
environment. Therefore, drop the DEBUG=1 flag from the ATF build
command.
Signed-off-by: Suhaas Joshi <s-joshi@ti.com>
Presently, in k3 devices, we are setting TEE=bl31.bin. However TEE
should be set to OPTEE, not ATF.
Therefore, do the following:
* clone and compile OPTEE
* add variables for OPTEE platform and `make` arguments in all k3
board configs
* use the compiled binary as TEE to compile U-Boot
Signed-off-by: Suhaas Joshi <s-joshi@ti.com>
Currently `current` picks 6.6.y kernel baseline, and `edge` picks
upstream's 6.12.y tree.
Make `current` pick the current SDK release (which, at the moment, is at
11.00.09 baseline), and make `edge` pick the latest un-released RC's
from TI's tree.
`current` is what most people will use, therefore it should be the same
as our official and tested release. `edge` can be either cutting-edge
(latest RC) or bleeding-edge (latest commit on cicd branch). For now,
keep it cutting-edge.
Signed-off-by: Suhaas Joshi <s-joshi@ti.com>
- Added FDT resize/trim to allow big DT and overlays and still use
configured load addresses in case `setexpr` not available
- Moved FDT resize and next load address alignment
- Added filename information to load instructions
- Moved all "global" variable determinations to the top of the
bootscript
https://github.com/armbian/build/issues/8178
- Add switch to turn avoidance on/off
- Increment base address before alignment to resolve the oboe
- oboe observed in U-Boot v2021.04 and in particular with DT loading
- Updated names to reflect function better.
- Followed coding style of rest of the file, e.g. `test` instead of `itest` and `if ... ; then`.
- If `setexpr` fails, use configured addresses - either the ones defined in boot.cmd or set by armbianEnv.txt.
Changes to align with standard beagle images.
- extlinux.conf is much more flexible than uEnv.txt, specially when
having multiple kernel for development - menu entries are valuable for recovery
- typical beagleboard Debian images have been using extlinux.conf for
a while.
Signed-off-by: Ayush Singh <ayush@beagleboard.org>
Co-authored-by: Andrei Aldea <7153954+Grippy98@users.noreply.github.com>
- **CHANGE**: This version tries to boot SD first (thus SD -> USB -> NVMe -> eMMC); before it would boot USB before SD
- as Armbian u-boot is possibly deployed to SPI, try to make as compatible as possible; eg enable squashfs support
- enable `setexpr` and `fileenv`, used by HAOS
- HAOS already has support for booting VIM3 from NVMe: https://github.com/home-assistant/operating-system/pull/3784
- borrowed `fileenv` patch from v2025.01 to .04
- Similar treatment that `odroidm1` got in https://github.com/armbian/build/pull/7769
- note: I will be leaving my VIM3 behind in the other side of the globe, so this is my last chance to test
* Update rock-4se.csc
Update the version of U-boot used with rock 4 se.
Tested locally on my board after building an armbian image.
ubuntu@rock-4se:~$ strings /dev/mmcblk1 | grep U-Boot
U-Boot SPL 2025.04-armbian-2025.04-S3482-P4a0b-H3d34-Va9af-Bb703-R448a (Apr 22 2025 - 10:05:17 +0000)
U-Boot phase
U-Boot
FIT image for U-Boot with bl31 (TF-A)
* Use BOOTCONFIG="rock-4se-rk3399_defconfig" and remove BOOTPATCHDIR
* Removed old comment
* Update rock-4se.csc
Revert BOOTCONFIG and update BOOT_FDT_FILE="rk3399-rock-4se.dtd"
* Update rock-4se.csc
Set BOOTCONFIG="rock-4se-rk3399_defconfig"
When RTL8152 USB network is enabled as build-in, it will throw out warnings such as:
W: Possible missing firmware /lib/firmware/rtl_nic/rtl8156b-2.fw for built-in driver r8152
W: Possible missing firmware /lib/firmware/rtl_nic/rtl8156a-2.fw for built-in driver r8152
W: Possible missing firmware /lib/firmware/rtl_nic/rtl8153c-1.fw for built-in driver r8152
W: Possible missing firmware /lib/firmware/rtl_nic/rtl8153b-2.fw for built-in driver r8152
W: Possible missing firmware /lib/firmware/rtl_nic/rtl8153a-4.fw for built-in driver r8152
W: Possible missing firmware /lib/firmware/rtl_nic/rtl8153a-3.fw for built-in driver r8152
W: Possible missing firmware /lib/firmware/rtl_nic/rtl8153a-2.fw for built-in driver r8152
This should prevent this while keep driver as module.
- Fix dram, mmc for h616.
- The prefix is missing if we use OF_UPSTREAM
- upstream: arm64: sun50i-h616: sync with sunxi-6.12 kernel
- upstream: arm64: Add sun50i-h618-bananapi-m4-berry.dts
- change name of /etc/kernel/postinst.d/z51-raspi-firmware and check for distribution in runtime. This will prevent conflict with raspi-firmware on Debian, while Ubuntu don't have that.
- install packages depending on Distribution. Ubuntu gets what it was, while Debian gets different packages, mirroed from official Rpi bookwork repo: raspi-firmware raspberrypi-sys-mods bluez-firmware bluez
- bump EDGE kernel to 6.14.y
At this stage, the desktops are functioning normally.
The Ethernet is working; module dwmac-sun8i.
WiFi\Bluetooth - the rtl8821cu module is loaded and
can be configured.
This adds fixed +16Gb of resize, which doesn't make image any bigger, but when importing this as is, it might be better to not start without any free space. Now its 16Gb - cca. 600Mb.
add some missing configs and fix update-initramfs
problems by switching to build-in for some modules
- USB_ONBOARD_DEV now build-in
- ARM_IMX_BUS_DEVFREQ now build-in
Signed-off-by: Martin Schmiedel <Martin.Schmiedel@tq-group.com>
- this reverts d048673c00 just for the Meko's
- there is little to be gained from .10 anyway on those machines (IMHO)
- it is still vendor 2017.09 after all, dunno why Radxa calls them "2024.x"
- some boards have u-boot auto-detect the board variant, and thus we unset `BOOT_FDT_FILE`
- but grub needs to know...
- so introduce a `GRUB_FDT_FILE` that is checked in case `BOOT_FDT_FILE` is not set
- this way we can keep u-boot autodetecting and grub happy
- preserve BOOT_FDT_FILE into GRUB_FDT_FILE for `nanopct6`/`nanopct6-lts`
software-properties-common was dropped from Debian Unstable
and may or may not come back according to its maintainer.
It provides the add-apt-repository binary.
c.f. https://github.com/wimpysworld/deb-get/issues/1215
Drop policykit-1 from common packages to include, it is only a
transitional package to pkexec and polkitd and has been dropped from
Debian Testing/Unstable and Ubuntu Plucky. Add pkexec and polkitd as
replacement.
it was dropped from debian unstable and hence trixie and ubuntu plucky. The
kerneloops.org website has been gone for about 5 years already so while the
package is still available for example in jammy, it is completely useless
in 2025.
- https://bugs.debian.org/953172
- https://bugs.launchpad.net/ubuntu/+bug/1930877
* Create rk3588-orangepi-5-ultra.dts
* Create orangepi5-ultra.csc
* fixups
* add notes
* fix u-boot compilation
* Update to v7 patch series
https://patchwork.kernel.org/project/linux-arm-kernel/cover/20250222193332.1761-1-honyuenkwun@gmail.com/
* Update rk3588-orangepi-5-ultra.dts
While I don't fully understand either the `kernel-dtb` command nor the dt syntax itself for some reason the `nophandles` version stripped more than actually phandles. Lets see what happens this way.
* Delete patch/kernel/integrate-6.15/0001-tools-Makefile-remove-pci-target.patch
* disable OPTEE
c81f89ca90
- Replace `armbian.list` with `armbian.sources`. This holds the same
information in a newer format, deb822.
- Replace HTTP with HTTPS for Armbian repositories.
- as Armbian u-boot is possibly deployed to SPI (replacing Petitboot),
try to make as compatible as possible; eg enable squashfs support
- in this case, `setexpr` and `fileenv` are used by HAOS
- found `fileenv` here 049e1a1949
Signed-off-by: Ricardo Pardini <ricardo@pardini.net>
