OpenWrt changes:
- rearrange configure options and variables
- provide variables for both configure and build stages (binutils has more than one ./configure script in source tree)
OpenWrt changes:
- adjust patch manually:
- 500-Change-default-emulation-for-mips64-linux.patch
- rearrange configure options and variables
- provide variables for both configure and build stages (binutils has more than one ./configure script in source tree)
Konstantin Demin [Fri, 17 Apr 2026 20:47:11 +0000 (23:47 +0300)]
toolchain: binutils: simplify patch management
Take in account only first two version components to lookup patch directory.
Hovewer, computed "BASE_VERSION" may be overridden (if necessary).
This change should prevent further issues with binutils being unpatched, see commits adad973a9c34 and 525a1e94b343.
Also drop obsolete "BIN_VERSION" variable (not used anywhere).
Nick Hainke [Fri, 17 Apr 2026 21:03:40 +0000 (23:03 +0200)]
x86: add support for DFI ADN553
The DFI ADN553 is a 3.5" SBC based on Intel Atom Alder Lake-N
processors with three Intel I226V 2.5GbE ports.
Specs:
* CPU: Intel Atom x7425E (4C, 12W) / x7213E (2C, 10W) /
x7211E (2C, 6W)
* RAM: 1x DDR5 SO-DIMM, up to 16GB
* Storage: 1x M.2 M key 2242/2280 (PCIe Gen3 x1/SATA3),
1x SATA 3.0
* Ethernet: 3x 2.5GbE RJ-45 (Intel I226V)
* USB: 4x USB 3.2 (rear), 2x USB 2.0 (internal)
* Expansion: 1x M.2 B key 3052 (USB3/USB2, opt. PCIe x1, SIM),
1x M.2 E key 2230 (USB/PCIe x1, CNVi)
* Display: 1x HDMI, 1x Type-C DP Alt. Mode, 1x LVDS/eDP
* Power: 9-36V DC wide range input
* TPM: dTPM 2.0 (NPCT750AADYX)
* Form factor: 3.5" SBC (146mm x 102mm)
Installation:
1. Write the combined-efi.img to a USB drive:
dd if=combined-efi.img of=/dev/sdX conv=fdatasync
2. Boot the ADN553 from the USB drive via the UEFI boot menu.
3. For permanent installation, write the image to the M.2 or
SATA storage device.
The board uses "Default string" as DMI sys_vendor and product_name
placeholders, so board detection is fixed by filtering these out and
falling through to board_vendor (DFI Inc.) and board_name (ADN553).
The three I226V NICs are pinned to their PCIe paths to ensure
consistent interface ordering matching the physical left-to-right
port layout. eth0 is assigned as WAN and eth1/eth2 as LAN.
Robert Marko [Wed, 15 Apr 2026 10:58:58 +0000 (12:58 +0200)]
generic: 6.12: backport EMC2305 changes
Backport the Microchip EMC2305 driver changes up to current linux-next.
These are required in order to actually be able to use the driver with
DTS support to drive fans as cooling devices.
Replace the bcm27xx RPi patch with the one from RPi kernel 7.0 that was
already adapted to all of these backports.
Zoltan HERPAI [Sun, 29 Mar 2026 14:47:12 +0000 (14:47 +0000)]
kernel/sifiveu: restore files for v6.12
This is an automatically generated commit which aids following Kernel patch
history, as git will see the move and copy as a rename thus defeating the
purpose.
For the original discussion see:
https://lists.openwrt.org/pipermail/openwrt-devel/2023-October/041673.html
Daniel Golle [Fri, 10 Apr 2026 20:00:29 +0000 (21:00 +0100)]
mediatek: build driver for built-in 2.5GE PHY as module
The by-now-upstream driver for the built-in 2.5GE PHY of the MediaTek
MT7988 and MT7987 SoC loads firmware at probe time.
Build the driver as a module in order to make sure the driver only
attempts to load the firmware by the time the rootfs with the firmware
file has become available.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Daniel Golle [Fri, 10 Apr 2026 19:58:57 +0000 (20:58 +0100)]
mediatek: fix Arioha AN8855 DSA driver for Linux 6.18
Upstream changes to phylink require to make some small changes to our
downstream Airoha AN8855 DSA driver, so it build with Linux 6.18.
The efuse driver is upstream by now and can be dropped.
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Daniel Golle [Fri, 10 Apr 2026 19:58:19 +0000 (20:58 +0100)]
mediatek: fix patches for Linux 6.18
Delete patches merged upstream and refresh the remaining patches.
Import pending patchset to fix read-out-of-bounds bug in clk driver
while at it, as it makes maintaining the downstream clk driver for
MT7987 easier.
Replace downstream patch for Fidelix FM35X1GA SPI-NAND chip with
upstream commit replacing it (Dosilicon DS35Q1GA is indentical).
Several additional fixes were needed for MT7623N, especially to get
HDMI and the Mali-450 GPU working again...
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Daniel Golle [Fri, 10 Apr 2026 13:03:10 +0000 (14:03 +0100)]
kernel/mediatek: restore files for v6.12
This is an automatically generated commit which aids following Kernel patch
history, as git will see the move and copy as a rename thus defeating the
purpose.
For the original discussion see:
https://lists.openwrt.org/pipermail/openwrt-devel/2023-October/041673.html
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Jonas Jelonek [Tue, 14 Apr 2026 11:17:43 +0000 (11:17 +0000)]
realtek: pcs: rtl93xx: extend USXGMII config
Our USXGMII config only covers one single register of which kind there
are actually more. We only set a value for 'CFG_QHSG_TXCFG_MAC_CH0' but
there are additional registers for CH1-CH3. Those refer to the 4
USXGMII 'channels'. While the RTL930x part of the SDK doesn't set them
explicitly, from RTK setup SerDes dumps we can see they are usually
similarly set.
The RTL931x part of the SDK actually writes those register explicitly
during USXGMII. We just haven't implemented that so far. Thus, add this
to the USXGMII config for both RTL93xx.
Jonas Jelonek [Tue, 14 Apr 2026 11:10:38 +0000 (11:10 +0000)]
realtek: pcs: rtl931x: use generic USXGMII config
Since the USXGMII config was made generic before, we can use it now for
the RTL931x setup too. This brings it closer to what the SDK configures
(see _dal_mango_construct_init_usxgmii), fixes some deviations in our
previous settings and makes clearer what is actually done.
Jonas Jelonek [Tue, 14 Apr 2026 11:05:16 +0000 (11:05 +0000)]
realtek: pcs: rtl93xx: make USXGMII config generic
The USXGMII config setting specific options currently is only
implemented for RTL930x. Looking at the SDK one can see that RTL931x
shares a lot of these configuration options. Additionally, several dumps
from RTK setups have shown that values which aren't set yet by us for
RTL930x but for RTL931x, also apply for RTL930x. Thus, both can be
merged. Start this by making the current function generic for RTL93xx.
USXGMII autoneg setting is currently hardcoded for RTL930x and not even
handled explicitly for RTL931x. This should be handled via neg_mode from
phylink subsystem too. Thus, move it over to rtpcs_93xx_sds_set_autoneg
as generic implementation for RTL93xx and set it depending on the
provided negotiation mode.
1. Prepare a TFTP server with 192.168.11.2
2. Rename initramfs image to "linux.ubi-recovery" and put it to the TFTP
directory
3. Hold the "AOSS" button and power on WSR-3000AX4P, release after 7~
seconds
4. The bootloader automatically downloads the initramfs image and boots
with it
5. After booting, upload a sysupgrade image to the device and perform
sysupgrade with it
6. Wait ~100 seconds to complete flashing
Reverting to stock image:
1. Download a official firmware and decrypt it by buffalo-enc
INAGAKI Hiroshi [Sun, 22 Mar 2026 13:15:04 +0000 (22:15 +0900)]
mediatek: filogic: enable ascii-env NVMEM driver
Enable ascii-env NVMEM driver (CONFIG_NVMEM_LAYOUT_ASCII_ENV) on
mediatek/filogic subtarget, to handle ascii-based MAC addresses
in key-value format stored to the mtd partition.
Zoltan HERPAI [Tue, 31 Mar 2026 12:36:53 +0000 (14:36 +0200)]
kernel/pistachio: restore files for v6.12
This is an automatically generated commit which aids following Kernel patch
history, as git will see the move and copy as a rename thus defeating the
purpose.
For the original discussion see:
https://lists.openwrt.org/pipermail/openwrt-devel/2023-October/041673.html
Jörg Seitz [Sun, 24 Aug 2025 04:51:11 +0000 (06:51 +0200)]
mediatek: filogic: Add new Router model ZBT-Z8106AX-T
Device support for zbt-z8106ax-t
Vendor Zbtlink advertizes this device as model Z8106AX-M2-T
on their website www.zbtlink.com. Device label sticked on
enclosure however states this is model Z8106AX version -T.
I made firmware selector to show this device as
- ZBT-Z8106AX-T to match information printed on the label and
- ZBT-Z8106AX-M2-T to match information found on vendors web pages.
Specifications:
SoC: MediaTek MT7981B
RAM: 256MiB
Flash: Winbond SPI-NAND 128 MiB
Switch: 1 WAN, 4 LAN (Gigabit) MediaTek MT7531
Buttons: Reset
Power: DC 12V-32V 1A
WiFi: MT7981B 2.4Ghz & 5Ghz
USB 3
M2 slot to hold LTE modem
2x nano SIM slots (user controllable)
Router comes in a flat metal box with all antennas detachable.
- 4 antennas for LTE 4G/5G communication
- 2 antennas for Wifi 2.4 GHz
- 2 antennas for Wifi 5 GHz
Power supply could be between 12V and 32V.
This serves both cars equipped with 12V batteries
and trucks equipped with 24V batteries.
Led Layout:
Power (green, user controllable, default set to OpenWrt Status)
Mobile (green, user controllable)
WLAN 2.4G (green, user controllable)
WLAN 5G (green, user controllable)
WAN (amber, user controllable, set to show eth1)
LAN1 (amber, hardware controlled)
LAN2 (amber, hardware controlled)
LAN3 (amber, hardware controlled)
LAN4 (amber, hardware controlled)
Slot SIM2 is set as default and matches label on Router enclosure
---
Installation:
A. Through U-Boot menu:
- Prepare your connecting computer to use a static IP in
network 192.168.1.0/24 like
a) 192.168.1.10 netmask 255.255.255.0 (legacy notation)
b) 192.168.1.10/24 (CIDR notation)
- Power down the router and hold in the Reset button.
- While holding in the button power up the router again.
- Hold the button in for 10 seconds and then release.
- Use your browser to go to 192.168.1.1
- If you see a GUI allowing for flashing firmware then you got the right spot.
- Upload the **Factory** image file.
Note: U-Boot GUI it can be used to recover from an incorrect firmware flash.
B. Through OpenWrt Dashboard:
If your router comes with OpenWrt preinstalled (modified by vendor),
you can easily upgrade by going to the dashboard (192.168.1.1) and
then navigate to "System" -> "Backup/Flash firmware"
Flash OpenWRT firmware and take care to deselect (untick) option
"keep settings". Settings done by vendor are incompatible with
versions 24.10 or 25.12.
Device features a GPIO controlled hardware watchdog.
Verfied by removing procd controlled watchdog and
seeing device rebooting.
---
Notes:
The zbt-z8106ax-t could be ordered from vendor with a variety of modems.
Mine came with a Quectel RM520N-GL. Quectel firmware was at RM520NGLAAR01A07M4G.
This level of firmware made some trouble connecting with some of my
SIM cardproviders.
Newer firmware level RM520NGLAAR01A08M4G_01.205.01.205 was available searching
github repositories. Upgrading my RM520-GL allowed to get successful connects
that did fail with older Quectel firmware.
Modem communication is set to ethernet control mode (ECM) by vendor.
Vendor takes advantage of ECM by wiring modem to internal switch port WWAN.
OpenWRT network configuration wants to define two network interfaces
- Network interface covering USB0 set with high metric
- Network interface covering WWAN set with low metric
Network interface covering WWAN would be preferred default route.
Please take note that internal switch port wired to LTE modem is named LAN5
in vendor provided firmwares. OpenWRT however names port as WWAN to better
describe purpose of port. WWAN is suggested to be assigned to firewall zone WAN.
Did use package qmodem from github repository FUjR/QModem to manage RM520N-GL LTE modem.
layerscape: armv8_64b: traverse ten64-mtd fix ASU support
The profiles.json[1] in generated and used during ASU sysupgrades takes
DEVICE_NAME as profile name which break ASU sysupgrades, use BOARD_NAME
which serves the same goal as overwriting DEVICE_NAME, allowing
sysupgrade-tar's checks be succesfully passed using the old name used in
older installations. Without affecting the generated profiles.json, thus
recovering ASU support.
*Keep the non-stardand device name with suffix "-mtd"(replaced underscore
by hyphen) to remark that this device is also supported in the armsr
(armvirt) target as originally intented[2].
*I think I have fully checked the sysupgrade path to confirm this change
does not have side effects.
Works:
------
- 2 SFP+ cages either 1G/10G or 2.5G
- 8*2.5G Ethernet ports
- Switch function
- LEDs
- Boot from flash
- Assigning MAC addresses from flash
Ethernet ports:
---------------
The 8x 2.5Gbps ethernet ports are provided by two RTL8224 chips. The
ports are supported by the upstream realtek PHY driver plus a local
initialization patch for the 10g-qxgmii mode.
Installation:
-------------
This device uses ZyNOS instead of Linux which makes the installation a
bit cumbersome. OpenWRT will be installed on the slot of the second
firmware image, thus the switch original firmware can be booted with a
little change. The serial console is required.
1. Set the switch to boot from the first image. This is required to
be sure not to be blocked in the middle of the installation
procedure.
2. Connect to the switch using the serial adapter and interrupt the boot
process by pressing "Enter" repeatedly.
3. Load the OpenWRT initramfs image using xmodem. From bootext console
(use ATHE to get the list of commands):
6. Reboot the switch and from the stock firmware set the configuration
to boot from the second image (Maintenance > Firmware upgrade > Boot
image).
7. Reboot again and enjoy OpenWRT.
Recovery/Return to stock:
-------------------------
1. Connect the switch using the serial adapter and interrupt the boot
process during the "DRAM POST: Testing:" sequence by pressing '$'
until the "XMG1915$" prompt appears.
2. Start an ymodem upgrade using the following command and use an xmodem
upload tool to send the .bin file provided in an OEM upgrade package.
XMG1915$ upgradeY image2 81000000 115200
## Ready for binary (ymodem) download to 0x081000000 at 115200 bps...
3. Wait for the upgrade process to finish and reboot the switch.
realtek: eth: normalize ring counter/size register access
This refactoring is a little bit of all ...
For access to the ring counter and size registers there exist currently
separate helpers. These are only for the register number but not the
bit position inside the register. Register definitions still use the
old prefix. Resolve this mix by
- Changing the prefixes of the register defines
- replacing the helpers with the register addresses
- just calculating register & offset manually where needed.
This might be a little step back but at least aligns with the rest of
the register access code. A future commit might bring an even better
version.
While we are here convert the main consumer rteth_93xx_hw_reset() to
regmap and fix a bug. Until now this function only clears the first
counter in a register because of a missing bit shift during writeback.
Brian Norris [Thu, 16 Apr 2026 01:58:59 +0000 (18:58 -0700)]
chromium: Add #{address,size}-cells to /firmware
Commit b4d7263bc3b6 ("kernel: of: avoid some unnecessary bad cell count
warnings") backported Linux commit 6e5773d52f4a ("of/address: Fix WARN
when attempting translating non-translatable addresses"), which started
requiring #address-cells for a device's parent if we want to use the
reg resource in a device node.
Many Chromium devices use a /firmware/coreboot device node that is
patched in by the boot firmware. These structures look something like:
There are currently two header files for the dsa driver. rtl83xx.h
and rtl838x.h. It is unclear for what reason they are separated.
Especially as one always includes the other. Merge those two files
into one.
ltq_atm.c: In function 'ltq_atm_probe':
ltq_atm.c:1840:36: error: passing argument 2 of 'platform_set_drvdata' discards 'const' qualifier from pointer target type [-Werror=discarded-qualifiers]
1840 | platform_set_drvdata(pdev, ops);
| ^~~
In file included from ltq_atm.c:40:
/workspaces/openwrt/build_dir/target-mips_24kc_musl/linux-lantiq_xway/linux-6.18.21/include/linux/platform_device.h:276:47: note: expected 'void *' but argument is of type 'const struct ltq_atm_ops *'
276 | void *data)
| ~~~~~~^~~~
Fixes: c1fa85f65931 ("treewide: use of_device_get_match_data") Signed-off-by: Shiji Yang <yangshiji66@outlook.com> Link: https://github.com/openwrt/openwrt/pull/22921 Signed-off-by: Robert Marko <robimarko@gmail.com>
This is an automatically generated commit which aids following Kernel patch
history, as git will see the move and copy as a rename thus defeating the
purpose.
For the original discussion see:
https://lists.openwrt.org/pipermail/openwrt-devel/2023-October/041673.html
This is an automatically generated commit which aids following Kernel patch
history, as git will see the move and copy as a rename thus defeating the
purpose.
For the original discussion see:
https://lists.openwrt.org/pipermail/openwrt-devel/2023-October/041673.html
Shiji Yang [Tue, 14 Apr 2026 13:52:50 +0000 (21:52 +0800)]
uboot-tools: fix libyaml linker error
When we cross build uboot-tools, the dtc tool is still compiled for
the host. Therefore, we should not attempt to link the OpenWrt system
libraries. CPU architecture mismatch can lead to build errors.
Fixes: https://github.com/openwrt/openwrt/issues/22924 Fixes: 55925650aabb ("uboot-tools: update to v2026.04") Signed-off-by: Shiji Yang <yangshiji66@outlook.com> Link: https://github.com/openwrt/openwrt/pull/22927 Signed-off-by: Robert Marko <robimarko@gmail.com>
Shiji Yang [Sat, 11 Apr 2026 00:34:00 +0000 (08:34 +0800)]
generic: phy: adm6996: fix build on 6.18 kernel
devm_gpio_request() was removed since kernel 6.17. Convert it to
devm_gpio_request_one() to fix:
target/linux/generic/files/drivers/net/phy/adm6996.c: In function 'adm6996_gpio_probe':
target/linux/generic/files/drivers/net/phy/adm6996.c:1183:15: error: implicit declaration of function 'devm_gpio_request'; did you mean 'devm_gpio_request_one'? [-Wimplicit-function-declaration]
1183 | ret = devm_gpio_request(&pdev->dev, priv->eecs, "adm_eecs");
| ^~~~~~~~~~~~~~~~~
| devm_gpio_request_one
Jonas Jelonek [Mon, 2 Feb 2026 22:30:05 +0000 (22:30 +0000)]
realtek: pcs: add media type for PCB
Previous commits carved out some TX configuration from different places
into a dedicated function. A cause of this is a behavioral change where
the TX amplifier values aren't applied anymore if it isn't a "fiber
mode". To restore this behavior and make the media/TX-RX configuration
more generic, we need a dedicated media type. The existing ones only
cover fiber and DACs, but not the other left case where a PHY is
attached to the SerDes. This now calls set_media for USXGMII and XSGMII
modes intentionally, both to prepare for future changes and to restore
previous behavior.
We do not have a reliable way to distinct between the actually used
media types, this is still a TODO. But several parts of the code already
have different values applied based on this information. Moreover, this
is especially needed for DAC cables to work properly. While this is
missing, we need to rely on inferring the media from the SerDes mode.
While at it, improve the call site of media handling since there's a
media type for all cases now. This allows to reduce the number of
function calls by moving it out and just have the media type in the
decision block.
Jonas Jelonek [Sat, 31 Jan 2026 23:34:30 +0000 (23:34 +0000)]
realtek: pcs: rtl931x: start to carve out tx config
Proceed with the consolidation of TX/RX path config by carving out some
TX configuration out into a separate function. For now, this only
affects TX amplifier config. But it already slims down other polluted
call sites and brings some structure into the TX/RX path config.
While at it, already provide a similar function for RX configuration.
This does nothing yet but should be filled with RX path config later.
Jonas Jelonek [Wed, 4 Feb 2026 20:19:55 +0000 (20:19 +0000)]
realtek: pcs: rtl931x: demystify some setup magic
Shed some light into the darkness by giving a part of the setup a
dedicated name. This applies to the magic data blocks in the generic
setup entrypoint and some writes in the media handling. This was just
copied from the SDK which doesn't annotate this properly with
information. But we can clearly connect this with some other functions
of the SDK and extract meaningful information.
The mentioned magic blocks and register writes set coefficients of three
amplifiers in the TX path, a pre-amp, a main-amp and a post-amp. They
are used to tune the signal for different kinds of media. Generic values
are applied for all SerDes and further special values for different media
types depending on the needs.
Provide a dedicated function that sets these amplifier coefficients.
Replace the mystic register writes with the function call so one can see
at a glance what's happening. Also replace the magic TX values with the
coefficients organized into a separate struct. This might be extended
further and is also applicable for RTL930x.
Jonas Jelonek [Fri, 27 Mar 2026 11:22:00 +0000 (12:22 +0100)]
realtek: pcs: rtl931x: condense similar writes
Within RTL931x set_media, two writes are applied in general for all modes
but later again changed depending on the mode. This is a confusing
pattern. To fix this, move those generic writes into the switch-block as
a default case. To maintain behavior, pull the 10G guard into the fiber
media case.
Jonas Jelonek [Fri, 27 Mar 2026 11:16:14 +0000 (12:16 +0100)]
realtek: pcs: rtl931x: relocate analog reset
Relocate the analog reset code within set_media for RTL931X to have it
in one place at the very top of the function, running all reset actions
before further real configuration is done. Also drop a separate call to
rx_reset because a subsequent sequence already includes this
functionality.
eip93_hmac_setkey() allocates a temporary ahash transform for
computing HMAC ipad/opad key material. The allocation uses the
driver-specific cra_driver_name (e.g. "sha256-eip93") but passes
CRYPTO_ALG_ASYNC as the mask, which excludes async algorithms.
Since the EIP93 hash algorithms are the only ones registered
under those driver names and they are inherently async, the
lookup is self-contradictory and always fails with -ENOENT.
When called from the AEAD setkey path, this failure leaves the
SA record partially initialized with zeroed digest fields. A
subsequent crypto operation then dereferences a NULL pointer in
the request context, resulting in a kernel panic:
The reported symbol eip93_aead_handle_result+0xc8c is a
resolution artifact from static functions being merged under
the nearest exported symbol. Decoding the faulting sequence:
The faulting LDR at [X26, #0x8] is loading ctx->flags
(offset 8 in eip93_hash_ctx), where ctx has been resolved
to NULL from a partially initialized or unreachable
transform context following the failed setkey.
Fix this by dropping the CRYPTO_ALG_ASYNC mask from the
crypto_alloc_ahash() call. The code already handles async
completion correctly via crypto_wait_req(), so there is no
requirement to restrict the lookup to synchronous algorithms.
Note that hashing a single 64-byte block through the hardware
is likely slower than doing it in software due to the DMA
round-trip overhead, but offloading it may still spare CPU
cycles on the slower embedded cores where this IP is found.
Fixes: 9739f5f93b78 ("crypto: eip93 - Add Inside Secure SafeXcel EIP-93 crypto engine support") Signed-off-by: Aleksander Jan Bajkowski <olek2@wp.pl>
[Detailed investigation report of this bug] Signed-off-by: Kenneth Kasilag <kenneth@kasilag.me> Link: https://github.com/openwrt/openwrt/pull/22886 Signed-off-by: Robert Marko <robimarko@gmail.com>
projects / thirdparty / openwrt.git / log
