This fixes the sysupgrade image generation for the LBR20 as before
updating resulted in a brick which needed to be recovered using
nmrpflash.
UART log of a bricked unit:
Loading DNI firmware for checking...
Loading firmware 1 ...
NAND read: device 0 offset 0xa600000, size 0x20000
131072 bytes read: OK
NAND read: device 0 offset 0xa600000, size 0x380000
3670016 bytes read: OK
rootfs imge header corrupted !
Loading firmware 2 ...
NAND read: device 0 offset 0xa600000, size 0x20000
131072 bytes read: OK
NAND read: device 0 offset 0xa600000, size 0x380000
3670016 bytes read: OK
rootfs imge header corrupted !
Comparing the partition contents of a bricked and a working unit showed
that on the bricked one the fake uImage header was missing. The UBI
partition also showed significant changes. Both are fixed when the
base DniImage receipt is used.
Signed-off-by: Christoph Krapp <achterin@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/20558
Signed-off-by: Robert Marko <robimarko@gmail.com>
This aligns the LED behaviour with other Orbi devices. Orbi devices have
multiple RBG LEDs at the top of the device and two status LEDs at the
back next to the barrel jack.
The current behaviour of other Orbi devices is to use the multi-color
LEDs at the top for status indication and the green/red LEDs at the back
for running/panic-indication. This matches the vendor behaviour except
the color choice.
Other devices use green as running, blue on upgrade, red on failsafe and
white on bootup, so this aligns the LBR20 behaviour to the rest.
Signed-off-by: Christoph Krapp <achterin@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/20558
Signed-off-by: Robert Marko <robimarko@gmail.com>
Some SoC might use the Serdes for the second USB port as a 3rd PCIe
line (with the SSTR register correctly setup).
Add the node for the 3rd PCIe card and enable for the eMMC RFB board.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Fix some warnings and compile errors generated by the new
fortify-headers when compiling some applications like strace.
Fixes: 6268692bd2 ("toolchain: fortify-headers: Update to version 2.3.3")
Link: https://github.com/openwrt/openwrt/pull/20561
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Update to recent fortify-headers version. This version also supports
_FORTIFY_SOURCE=3.
Changes: https://github.com/jvoisin/fortify-headers/compare/1.1...2.3.3
A patch similar to the removed one was applied upstream:
b40f6d8748
The root file system is getting a bit bigger.
Without this commit:
5250380 bin/targets/armsr/armv8/openwrt-armsr-armv8-generic-rootfs.tar.gz
4919997 bin/targets/malta/be/openwrt-malta-be-default-rootfs.tar.gz
With this commit:
5263407 bin/targets/armsr/armv8/openwrt-armsr-armv8-generic-rootfs.tar.gz
4930011 bin/targets/malta/be/openwrt-malta-be-default-rootfs.tar.gz
Link: https://github.com/openwrt/openwrt/pull/20313
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Fixes the following error by backporting upstream update:
```
scripts/dtc/pylibfdt/libfdt_wrap.c: In function ‘_wrap_fdt_next_node’:
scripts/dtc/pylibfdt/libfdt_wrap.c:5581:17: error: too few arguments to function ‘SWIG_Python_AppendOutput’
5581 | resultobj = SWIG_Python_AppendOutput(resultobj, val);
| ^~~~~~~~~~~~~~~~~~~~~~~~
```
This is the same fix as applied in 4e68103c4e ("uboot-rockchip: fix build with swig 4.3.0")
Link: https://github.com/openwrt/openwrt/pull/20538
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Backport a patch fixing a kernel panic on airoha_hw_init fail. This
should better handle scenario with NPU load Probe deferring.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Backport cleanup patch for Airoha Ethernet patch to permit easier
backport in the future.
Automatically refresh all affected patch.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Add pending patch to make address some workaround needed to make the
Aeonsemi AS21xxx PHY working on the Airoha AN7581/AN7583 board.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Airoha AN7581 eMMC RFB board mount 2 Aeonsemi AS21xxx PHY. One is
usually connected to GDM4 and the other is optionally connected to GDM2.
Add the relevant nodes to support the one connected to GDM4.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Pack the BL2 and BL31+U-Boot artifacts as Airoha AN7581 is currently
supported in upstream U-Boot and bootloader files can be used for
unfused boards.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
As mkits.sh only generates the relevant DTS for mkimage and is
specific for the building image, we can move it outside the locked
section as it doesn't do any operation that can be used concurrently by
others.
This won't have any real impact but clean the code making it clear what
needs to be protected and what can be executed concurrently.
Link: https://github.com/openwrt/openwrt/pull/20492
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
It's expected the mkits.sh script to generate only the relevant DTS
entry and have all the blob already prepared to use for mkimage.
This is not the case for the RootFS case where the script generates a
.pagesync with the dd command.
To better handle this, drop the dd command and instead error out if the
.pagesync blob is not found if RootFS is used.
Adapt the generic fit build in image-commands.mk to call the dd for
.pagesync right before mkits.sh.
Link: https://github.com/openwrt/openwrt/pull/20492
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
With further investigation it was found a race in generating .itb images
that include a RootFS caused by the mkits.sh and the mkimage.
Due to the fact that mkits.sh generates a .pagesync image of the passed
rootfs, it can happen that, concurrently, mkimage can be called at the
same time mkits.sh is creating another .pagesync for the same rootfs.
This cause mkimage to use an half made rootfs.pagesync creating a
corrupted image.
To address this, also protect the mkimage with the same lock used for
mkits.sh preventing any kind of concurrent usage/generation of the
rootfs.pagesync blob
Fixes: 52cc9d82f1 ("kernel: rework Initramfs locking logic")
Link: https://github.com/openwrt/openwrt/pull/20492
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Add pending patch that disable libmagic support as this cause
reproducibility issue due to timing of presence of libmagic library at
times e2fsprogs gets compiled.
The new patch adds control on libmagic inclusion and we add the
--without-libmagic to the configure args by default.
Reported-by: Paul Spooren <mail@aparcar.org>
Link: https://github.com/openwrt/openwrt/pull/20499
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Drop redundant patch for host build as this has been fixed in 1.44.1
but for some reason it wasn't notice with the patch directly refreshed
(with the fix right up)
Fixes: 198172c048 ("package/utils/e2fsprogs: update to 1.44.1")
Link: https://github.com/openwrt/openwrt/pull/20499
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Backport additional patch required for NPU support of Airoha AN7583.
These are specific for the NPU module with some minor fixes and to adds
upport for loading the specific Airoha AN7583 NPU firmware.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Add relevant nodes for NPU support for Airoha AN7583 and enable the
kernel config to enable support in the Airoha Ethernet driver.
NPU firmware is scheduled to be upstreamed to linux-firmware and will be
added once present upstream.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Fix the current image Makefile for Airoha AN7583 by dropping the
artifacts entry for bootloader and add default PKG for the PHY mounted
on the RFB boards.
The artifacts for the bootloader will be readded later once ATF-A
support will be pushed mainline.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
While adding support for Airoha AN7583, it was forget to push the EMMC
DTS while defining the variant in the image Makefile.
Add the missing DTS to fix image creation error.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
On the Airoha AN7583 EVB there are 2 Aeonsemi AS21xxx PHY but only the
one connected to Serdes Ethernet is usable as the other connected to
Serdes PON is not actually HW connected (as the serdes is connected to
the BOSA port)
Add comments and disable the relevant PHY.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
This commit changes the Meraki MR33 and MR74 device trees to use nvmem
for ART calibration.
The WiFi BDF was moved from insect-common.dtsi to the respective
device files in preparation for additional insect-family devices being added.
Signed-off-by: Hal Martin <hal.martin@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/20474
Signed-off-by: Robert Marko <robimarko@gmail.com>
