Eric Fahlgren [Fri, 2 Jan 2026 02:02:30 +0000 (18:02 -0800)]
treewide: add explicit default variant
Some packages with variants did not specify the default among the
alternatives, so were left without any apk 'provider_priority'
for that package. This caused the apk solver to select the wrong
variant, silently changing the requested package list.
Notable among these were busybox, procd and the hostapd/wpad suite.
This behavior presented in the imagebuilders when creating the
image as follows, silently replacing packages even when explicitly
requested:
George Sapkin [Sun, 28 Dec 2025 15:22:41 +0000 (17:22 +0200)]
build: provide virtual self in kmods
Add an implicit self-provide to kmods. apk can't handle self provides,
be it versioned or virtual, so opt for a prefix and a suffix instead.
Package name without a prefix/suffix is too generic and might conflict
with other packages, e.g. wireguard. This allows several variants to
provide the same virtual package without adding extra provides to the
default one, e.g. r8169 implicitly provides kmod-r8169-any and is marked
as default, so r8125 can explicitly provide @kmod-r8169-any as well.
George Sapkin [Thu, 25 Dec 2025 13:32:29 +0000 (15:32 +0200)]
build: add support for virtual provides
Allow defining virtual provides using the PROVIDES field by prefixing
them with @, e.g.:
PROVIDES:=@ca-certs
Virtual provides don't own the provided name and multiple packages with
the same virtual provides can be installed side-by-side. Packages must
still take care not to override each other's files.
Add an implicit self-provide to packages. apk can't handle self
provides, be it versioned or virtual, so opt for a suffix instead. This
allows several variants to provide the same virtual package without
adding extra provides to the default one, e.g. wget implicitly provides
wget-any and is marked as default, so wget-ssl can explicitly provide
@wget-any as well.
Filter out virtual provides when generating metadata.
Filter out virtual provides prefix and self provide where appropriate.
Daniel Golle [Tue, 23 Dec 2025 14:49:48 +0000 (14:49 +0000)]
mediatek: filogic: asus,rt-ax57m: remove unused pinctrl groups
Remove unused pinctrl groups defined in the device tree, probably as a
result of importing the vendor device tree which is based on the
reference board from MediaTek's SDK.
2. Connect the PC via LAN to one of the yellow router ports and wait
until your PC to get a DHCP lease.
3. Browse to http://192.168.1.1 (or whatever your router IP is configured as)
4. If your router is brand new, finish the setup process and log into
the Web-UI.
5. Navigate to Administration -> Firmware Upgrade and upload the
downloaded OpenWrt initramfs.trx image.
6. Wait for OpenWrt to boot. Open the default OpenWrt IP in a web browser and perform a system upgrade using the sysupgrade image as instructed in the top banner.
---------------------------------------------------------------------------
Revert to stock firmware:
1: Install package facinstall -> https://forum.openwrt.org/t/facinstall-package-for-easy-installation-factory-images/177587
2. Install Asus stock trx image via OpenWRT LuCI interface.
Eric Fahlgren [Mon, 3 Nov 2025 20:34:37 +0000 (12:34 -0800)]
mediatek: filogic: fix supported_devices list for gl-mt2500
The SUPPORTED_DEVICES sets for both Maxlinear (v1) and Airoha (v2)
devices were identical, so sysupgrade was unable to detect when an
incorrect image was being installed. This caused "soft bricking" of
devices when a v1 image was installed on a v2 device, and vice versa.
Fix this by making the supported_devices distinct for each device
version, by renaming the devices with a version-specific name.
This is reflected in the file name and the image metadata.
Hauke Mehrtens [Sun, 28 Dec 2025 22:03:32 +0000 (23:03 +0100)]
wifi-scripts: use OWE + country for 6GHz
Wifi on the 6GHz band does not support open networks, configure owe by
default. 6GHz wifi also needs a country code, configure a broken country
code '00' by default to hint the user to change it.
Nick Hainke [Sat, 27 Dec 2025 12:43:47 +0000 (13:43 +0100)]
wifi-scripts: fix race in wireless interface creation
When wdev interfaces and hostapd interfaces are mixed, race conditions
can occur. Sometimes all interfaces are started correctly, sometimes only
the wdev interface and some of the hostapd interfaces, and sometimes only
the wdev interface.
Fix this by creating the wdev interfaces first, before processing
wpa_supplicant or hostapd interfaces.
RTL930x devices have highmem starting address at 0x20000000.
The Linksys LGS328C highmem definition is wrongly shared with
the larger LGS352C RTL931x model and starts at 0x90000000.
Fix it by splitting the definition.
Nick Hainke [Fri, 26 Dec 2025 17:35:01 +0000 (18:35 +0100)]
wifi-scripts: fix mcast_rate type (number, not array)
The multicast rate (mcast_rate) is defined as a single numeric value,
not an array. The ucode scripts already expect a single number, and
the OpenWrt documentation specifies this option as an integer [0].
Adjust the wireless schema accordingly. This also fixes the wdev.uc
call path for the iw fallback mesh setup.
Nick Hainke [Thu, 25 Dec 2025 17:41:10 +0000 (18:41 +0100)]
wifi-scripts: fix wdev fallback for mesh interfaces
The previous wdev.uc invocation failed for mesh interfaces.
Rewrite it as a loop so the mesh interface is created correctly
when no wpa_supplicant mesh support is installed.
Nick Hainke [Thu, 25 Dec 2025 19:32:28 +0000 (20:32 +0100)]
wifi-scripts: ucode: fix wpa_supplicant mesh
Mesh interfaces may not define an encryption setting. This results in
a ucode runtime error when parse_encryption() is called:
Reference error: left-hand side expression is null
In parse_encryption(), file /usr/share/ucode/wifi/iface.uc, line 20, byte 32:
called from function setup_sta (/usr/share/ucode/wifi/supplicant.uc:59:31)
called from function generate (/usr/share/ucode/wifi/supplicant.uc:243:41)
called from function setup (./mac80211.sh:254:61)
called from anonymous function (./mac80211.sh:342:14)
` config.auth_type = encryption[0] ?? 'none';`
Near here ------------------------^
Dominic Germain [Fri, 19 Dec 2025 14:46:57 +0000 (09:46 -0500)]
kernel: add option for USB FunctionFS gadget
Add menuconfig option 'kmod-usb-gadget-fs' for the USB FunctionFS gadget
module
OpenWRT is perfect for USB gadgets that are network-enabled, since it
provides everything needed to easily manage the device. Having support
for the FunctionFS gadget enables the use of any custom USB devices
that does not have a specific module, like Media Transfer Protocol.
George Sapkin [Wed, 24 Dec 2025 02:32:01 +0000 (04:32 +0200)]
toolchain: fix libgcc ABI for libc
libc depends on knowing libgcc's ABI, so it needs to be evaluated first.
Otherwise libc will depend on an ABI-less libgcc causing the rest of the
toolchain to fail.
Building package index...
ERROR: unable to select packages:
libgcc (virtual):
note: please select one of the 'provided by'
packages explicitly
provided by: libgcc1
required by: world[libgcc]
George Sapkin [Wed, 19 Nov 2025 23:50:03 +0000 (01:50 +0200)]
build: fix apk packaging and ABI-versioning
The updated logic for the APK dependencies and provides is as follows:
- If ABI version is defined:
- package is named `package_name-ABI_version`
- package implicitly provides
`package_name-ABI_version=package_version`
this implies that only one version of a package per ABI can be
installed at the same time
- additionally provide `package_name` so multiple packages can be
looked up by its base name
- for each `provides`, provide `provide-ABI_version=package_version`
this implies that only one version of a provide can be installed at
the same time
- else if ABI version is _not_ defined
- package is named `package_name`
- package implicitly provides `package_name=package_version`
this implies that only one version of a package can be installed at
the same time
- if `alternatives` is defined
- for each `provides`, provide `provide`
this implies that multiple versions of a provide can be installed
at the same time
- else if `alternatives` is _not_ defined
- for each `provides`, provide `provide=package_version`
this implies that only one version of a provide can be installed
at the same time
Both cases a package can be looked up by its base name.
ABI version `alternatives`, `conffiles`, `conffiles_static`, `list` and
`rusers` files so multiple versions of the same ABI package can be
installed side by side, and so they don't overwrite each other's
packaging files.
ABI version `EXTRA_DEPENDS` so dependencies can be correctly looked up
using the existing OpenWrt semantics without the ABI specified. This is
needed since ABI-versioned libraries no longer provide
`package_name=package_version`, so that they can be installed side by
side.
Remove duplicate dependencies when `EXTRA_DEPENDS` specifies a versioned
one that is already in `DEPENDS`.
ABI is defined
------------------------------------------------------------------------
`libsqlite3` has `PROVIDES` set to `libfake` and has two different ABI
versions installed. `libfake` is just an example to demonstrate the
mechanics, as the library can already be depended upon using e.g.
`libsqlite3-0=3.51.0-r1`. Note the ABI-versioned lists.
root@OpenWrt:/tmp# ls -lh /usr/lib/libsqlite3.so.*
lrwxrwxrwx 1 root root 20 Nov 20 00:23 /usr/lib/libsqlite3.so.0 -> libsqlite3.so.3.51.0
lrwxrwxrwx 1 root root 20 Nov 20 00:27 /usr/lib/libsqlite3.so.1 -> libsqlite3.so.4.00.0
-rwxr-xr-x 1 root root 1.0M Nov 6 18:19 /usr/lib/libsqlite3.so.3.51.0
-rwxr-xr-x 1 root root 1.0M Nov 6 18:19 /usr/lib/libsqlite3.so.4.00.0
```
ABI is not defined
------------------------------------------------------------------------
Both `avahi-dbus-daemon` and `avahi-nodbus-daemon` provide `avahi-daemon`,
but have no ABI specified. This results in `avahi-daemon=0.8-r11` provides
for both packages and only one being able to be installed at the same time:
Provides and alternatives
------------------------------------------------------------------------
Both `uclient-fetch` and `wget-nossl` provide `wget` and specify
alternatives, so provides are not versioned and both packages can be
installed at the same time:
Interrupt vector for reo2host-status is wrongly assigned.
Status interrupts weren't received and a workaround was applied to
mac80211 to periodically poll the reo status ring. Therefore, the
workaround can be removed with the proper hardware interrupt line
assigned.
Radek Dostál [Sun, 14 Dec 2025 09:37:06 +0000 (10:37 +0100)]
ath79: fix build of big images for TP RE355v1, RE450v1 & v2
After merge of https://github.com/openwrt/openwrt/pull/20709 I noticed,
that https://mirror-03.infra.openwrt.org/snapshots/targets/ath79/tiny/
still does not contain the build for RE450.
I analyzed the problem and noticed, that only builds which fits into
original size can be build and if the image is larger, it fails with
the following error: "WARNING: Image file .../openwrt/build_dir/
target-mips_24kc_musl/linux-ath79_tiny/tmp/
openwrt-ath79-tiny-tplink_re450-v1-squashfs-sysupgrade.bin is too big:
> 7864320".
This error is quite misleading as the image is smaller than 7864320. The
reason for this error is that default tplink-safeloader has hardcoded
factory partition structure and fails to generate any file.
Rather then fixing tplink-safeloader I followed approach from commit ebd5e5fb5359 ("ramips: switched TP-Link RE305 v1 to new partition layout")
and switched away from "tplink-safeloader sysupgrade".
I did not include "IMAGES := sysupgrade.bin", because with tiny target it
is still possible to locally build even factory image.
Fixes commit e768731dc8e9 ("ath79: switch TP-Link RE355 v1, RE450 v1 and
RE450 v2 to mtd-concat").
mediatek: fix pwn fan settings for sinovoip bpi-r3 (v3)
Popular cheap PWM fans for this machine, like the ones coming in
heatsink+fan combos will not work properly at the currently defined
medium speed. Trying different pwm setting using a command
pwm1 value fan rotation speed cpu temperature notes
-----------------------------------------------------------------
0 maximal 31.5 Celsius too noisy
40 optimal 35.2 Celsius no noise hearable
95 minimal
above 95 does not rotate 55.5 Celsius
-----------------------------------------------------------------
Current cooling levels are:
cooling-levels = <255 96 0>;
Thus only cpu-active-high and cpu-active-low modes are usable.
This patch fixes cpu-active-medium settings for bpi-r3 board.
This patch may not be enough. Users may wants to tweak their thermal_zone0
trip points, thus tuning fan rotation speed depending on cpu temperature.
That can be done on the base of the following example:
=== example =========
# cpu temperature below 25 Celsius degrees, no rotation
echo 25000 > /sys/class/thermal/thermal_zone0/trip_point_4_temp
# cpu temperature in [25..32] Celsius degrees, normal rotation speed
echo 32000 > /sys/class/thermal/thermal_zone0/trip_point_3_temp
# cpu temperature above 50 Celsius degrees, max rotation speed
echo 50000 > /sys/class/thermal/thermal_zone0/trip_point_2_temp
=====================
Changes v2:
* put patch to a proper directory
* updated patch description
* tested with latest openwrt based on linux-6.6
Changes v3:
* use upstream linux patch
* update patch description
0779ee287db5 dhcpv6-ia: reconfigure message length bug fix d69896b0ac7c all: remove dead code 3f6b3b4b9f8a odhcpd: remove fallback DNS search domain ff3f72fdd01d router: remove some log spam in send_router_advert() c1cc541e1065 router: improve send_router_advert() 901851b883da router: Modify relayed RA PIO P flag according to interface policy 179429ba477d router: Modify relayed RA PIO A flags according to interface policy
cb1b8770b782 config: add default value for dhcpv6_pd_min_len 20eb94d02dfe config: allow minimum PD len up to 64 7ebd96083971 Revert "router: optimize duplicated PIO comparison" 90ae6fc6e478 statefiles: don't consider no hostname as broken
Petr Štetiar [Tue, 16 Dec 2025 16:42:19 +0000 (16:42 +0000)]
dropbear: bump to 2025.89 (CVE-2025-14282, CVE-2019-6111)
- Security: Avoid privilege escalation via unix stream forwarding in Dropbear
server. Other programs on a system may authenticate unix sockets via
SO_PEERCRED, which would be root user for Dropbear forwarded connections,
allowing root privilege escalation.
Reported by Turistu, and thanks for advice on the fix.
This is tracked as CVE-2025-14282, and affects 2024.84 to 2025.88.
It is fixed by dropping privileges of the dropbear process after
authentication. Unix stream sockets are now disallowed when a
forced command is used, either with authorized_key restrictions or
"dropbear -c command".
In previous affected releases running with "dropbear -j" (will also disable
TCP fowarding) or building with localoptions.h/distrooptions.h
"#define DROPBEAR_SVR_LOCALSTREAMFWD 0" is a mitigation.
- Security: Include scp fix for CVE-2019-6111. This allowed
a malicious server to overwrite arbitrary local files.
The missing fix was reported by Ashish Kunwar.
- Server dropping privileges post-auth is enabled by default. This requires
setresgid() support, so some platforms such as netbsd or macos will have to
disable DROPBEAR_SVR_DROP_PRIVS in localoptions.h. Unix stream forwarding is
not available if DROPBEAR_SVR_DROP_PRIVS is disabled.
Remote server TCP socket forwarding will now use OS privileged port
restrictions rather than having a fixed "allow >=1024 for non-root" rule.
A future release may implement privilege dropping for netbsd/macos.
- Fix a regression in 2025.87 when RSA and DSS are not built. This would lead
to a crash at startup with bad_bufptr().
Reported by Dani Schmitt and Sebastian Priebe.
- Don't limit channel window to 500MB. That is could cause stuck connections
if peers advise a large window and don't send an increment within 500MB.
Affects SSH.NET https://github.com/sshnet/SSH.NET/issues/1671
Reported by Rob Hague.
- Ignore -g -s when passwords arent enabled. Patch from Norbert Lange.
Ignore -m (disable MOTD), -j/-k (tcp forwarding) when not enabled.
- Report SIGBUS and SIGTRAP signals. Patch from Loïc Mangeonjean.
- Fix incorrect server auth delay. Was meant to be 250-350ms, it was actually
150-350ms or possibly negative (zero). Reported by pickaxprograms.
- Fix building without public key options. Thanks to Konstantin Demin
- Fix building with proxycmd but without netcat. Thanks to Konstantin Demin
- Fix incorrect path documentation for distrooptions, thanks to Todd Zullinger
- Fix SO_REUSEADDR for TCP tests, reported by vt-alt.
Dropped:
* 050-dropbear-multihop-fix.patch as its included in the release 5cc0127000db5f
* 051-fix-pubkey-options.patch as its included in the release 1d4c4a542cd5df
* 052-fix-missing-depends-for-sntrup761x25519-sha512.patch as its included
in the release 1a2c1e649a1824
* 053-Don-t-limit-channel-window-to-500MB.patch as its included in the release a8610f7b98ad
Manually rebased:
* 110-change_user.patch
Fixes: CVE-2025-14282, CVE-2019-6111 Reviewed-by: Hauke Mehrtens <hauke@hauke-m.de> Reviewed-by: Konstantin Demin <rockdrilla@gmail.com> Tested-by: Konstantin Demin <rockdrilla@gmail.com> [mediatek/filogic (GL.iNet GL-MT6000)] Link: https://github.com/openwrt/openwrt/pull/21186 Signed-off-by: Petr Štetiar <ynezz@true.cz>
(cherry picked from commit 70a4da1cebc87fc7b3cfa3f634b5bdf3db12ccd5)
This fixes a boot hang on realtek rtl838x switches.
This is the last printed message:
Inode-cache hash table entries: 8192 (order: 3, 32768 bytes, linear)
Thread on mips mailing list:
https://lore.kernel.org/linux-mips/b35fe4b3-8f42-49f4-a6bf-9f0e56d4050c@hauke-m.de/T/#u
Rany Hany [Mon, 15 Dec 2025 17:11:24 +0000 (19:11 +0200)]
wifi-scripts: update phys after rename_phy_by_name call
This fixes a failed bring up of the radio on bootup
if the model defines a rename of phy in its /etc/board.json.
This specifically impacts Redmi AX6S and any router that does so
in its /etc/board.json. The fix fortunately is simple, just update
phy name in phys after rename.
The entry that specifically causes this issue is the following:
The issue is that after rename, referenced phy in config is going to be
wl0 but in phys array it is still phy0; and so it fails to find phy
and does not bring up radio.
Jonas Jelonek [Tue, 16 Dec 2025 12:06:57 +0000 (12:06 +0000)]
realtek: mdio-serdes: use correct device table identifier
Use the correct identifier 'rtsds_of_match' instead of
'rtsds_mdio_of_match' because the latter doesn't exist.
This doesn't cause an error for 6.12. However, with 6.18 the
implementation of MODULE_DEVICE_TABLE has changed to use 'static' and
'used' [1] instead of 'extern' and 'unused' [2].
Jonas Jelonek [Thu, 11 Dec 2025 23:33:59 +0000 (23:33 +0000)]
realtek: pcs: move polarity into SerDes struct
As a first real usage of the new SerDes struct, move the polarity
configuration there. It was previously located in the global rtpcs_ctrl
struct as an array, indexed by SerDes id. Because this is per-SerDes
information, the new SerDes struct is the correct place to live in.
Jonas Jelonek [Sat, 13 Dec 2025 10:48:52 +0000 (10:48 +0000)]
realtek: pcs: drop unneeded SerDes number range checks
By using references to pre-initiated SerDes instances instead of plain
SerDes number, there is no need to check for the range anymore in
various places. During driver/pcs init it is ensured that only valid
SerDes will reach the configuration functions.
Jonas Jelonek [Sat, 13 Dec 2025 20:12:13 +0000 (20:12 +0000)]
realtek: pcs: make use of SerDes struct in set_autoneg
Also switch set_autoneg (and related helper rtpcs_sds_modify) to the
SerDes struct instead of the plain SerDes id by using just the reference
to the SerDes instance instead of (ctrl, sds_id) tuple. This completes
the transition.
Jonas Jelonek [Thu, 11 Dec 2025 21:42:38 +0000 (21:42 +0000)]
realtek: pcs: make use of SerDes struct in SerDes setup
Make use of the previously added SerDes struct in SerDes setup and all
functions in its call path by removing (ctrl, sds_num) being passed to
every function call and instead just pass the reference to the
corresponding SerDes instance.
Various SerDes calculations for even, odd and neighbor are unified by
switching to previously introduced helpers.
Jonas Jelonek [Sat, 13 Dec 2025 20:01:03 +0000 (20:01 +0000)]
realtek: pcs: switch to id from SerDes struct
Drop usage of the to-be-phased-out SerDes id stored in rtpcs_link and
use the reference to the SerDes instance to use the embedded id in
rtpcs_serdes instead.
Jonas Jelonek [Sat, 13 Dec 2025 10:29:56 +0000 (10:29 +0000)]
realtek: pcs: assign SerDes reference upon PCS creation
Upon creation of a phylink_pcs instance by calling rtpcs_create, assign
a reference to the corresponding SerDes to the link structure. In the
next step, this should be used everywhere instead of the plain SerDes
number.
Rename the field used to hold the SerDes number from 'sds' to 'sds_num'
and name the new field 'sds' to make clear what is what.
Jonas Jelonek [Sat, 13 Dec 2025 10:16:21 +0000 (10:16 +0000)]
realtek: pcs: add helpers for even, odd, neighbor SerDes
Add dedicated helpers to get references to even, odd and neigbor SerDes
if needed. This should replace the various calculations scattered
throughout the code, providing a unified way to work with adjacent
SerDes.
Jonas Jelonek [Sat, 13 Dec 2025 10:12:35 +0000 (10:12 +0000)]
realtek: pcs: add separate SerDes struct
Add a separate structure for a SerDes. This is needed to appropriately
store per-SerDes information, which in turn is needed for future work.
Additionally, it's intended to reduce boilerplate and several
inconsistencies.
Jonas Jelonek [Sat, 13 Dec 2025 09:10:55 +0000 (09:10 +0000)]
realtek: pcs: use per-variant SerDes count
Use a separate configuration field for the number of SerDes for each
variant of the Realtek Otto family. Add this field to the config
structure, assign it and use it during driver probe. This narrows
possible error cases and is needed for upcoming extensions.
Jonas Jelonek [Thu, 11 Dec 2025 19:17:26 +0000 (19:17 +0000)]
realtek: pcs: add dedicated enum for SerDes modes
The Realtek SerDes mode capabilities do not map 1:1 to the
PHY_INTERFACE_MODE_* modes used in the kernel and passed to the PCS.
For example, some PHY chips use the proprietary XSGMII mode for which
there isn't an equivalent in the kernel, or HSGMII.
In the past, this led to problems and confusion using kernel's XGMII to
handle the XSGMII mode, and needed a downstream patch for HSGMII. They
have been solved/worked around for now, but XSGMII is currently not
implemented at all. And who knows what might come in the future.
To make our life easier, introduce a dedicated internal representation
of SerDes modes which differs from kernel's PHY_INTERFACE_MODE_*. This
allows us to map "external" modes to different internal modes as needed
instead of carrying the PHY_INTERFACE_MODE_* through the whole SerDes
configuration code. The PCS driver needs to map PHY_INTERFACE_MODE_* to
RTPCS_SDS_MODE_* in pcs_config, and the latter should be used as the
only one.
Jonas Jelonek [Fri, 12 Dec 2025 20:05:52 +0000 (20:05 +0000)]
realtek: pcs: rtl930x: drop unused and broken function
Drop the unused and broken function rtpcs_930x_sds_clock_wait from the
PCS driver. The proper working variant is already some lines above and
called rtpcs_930x_sds_wait_clock_ready.
Jonas Jelonek [Thu, 11 Dec 2025 22:37:33 +0000 (22:37 +0000)]
realtek: pcs: drop PCS creation without SerDes ref
Since the beginning, the PCS driver had the ability to call its
rtpcs_create without a reference to a valid PCS node. A comment in the
code mentions that this is done for RTL838X and its built-in octa-PHY
which is connected directly instead of via a SerDes. Further
explanations are not provided.
Drop this ability and make the rtpcs_create call in the dsa driver
conditional. As the built-in PHY of RTL838X isn't attached to a SerDes,
there is no obvious point of having the PCS driver in that chain. The
ports are marked as internal and have no pcs-handle, thus no phylink_pcs
instance should be created.
Jonas Jelonek [Tue, 16 Dec 2025 10:20:13 +0000 (10:20 +0000)]
realtek: mdio-serdes: improve debugfs creation
Commit 3c073b5cb2 cleaned up the debugfs creation in
mdio-realtek-otto-serdes driver to not explicitly check if the root
directory already exists. This is fine because kernel handles the case
properly so there's no need to check anymore.
However, this pollutes the boot log with:
[..] debugfs: 'realtek_otto_serdes' already exists in '/'
[..] debugfs: 'realtek_otto_serdes' already exists in '/'
[..] debugfs: 'realtek_otto_serdes' already exists in '/'
[..] debugfs: 'realtek_otto_serdes' already exists in '/'
[..] debugfs: 'realtek_otto_serdes' already exists in '/'
[..] debugfs: 'realtek_otto_serdes' already exists in '/'
[..] debugfs: 'realtek_otto_serdes' already exists in '/'
[..] debugfs: 'realtek_otto_serdes' already exists in '/'
[..] debugfs: 'realtek_otto_serdes' already exists in '/'
[..] debugfs: 'realtek_otto_serdes' already exists in '/'
[..] debugfs: 'realtek_otto_serdes' already exists in '/'
Now, the root directory creation is attempted multiple times, causing
the kernel to print an error message because the directory already
exists.
Fix this by moving the SerDes loop into rtsds_debug_init and only try
to create the root debugfs directory once.
Christoph Krapp [Sat, 22 Nov 2025 21:34:04 +0000 (22:34 +0100)]
ipq40xx: convert Orbi led labels to function/color
The eight leds controlled by the LED controller are RGB leds themselves
but are flashing white by default. The color part is controlled by GPIOs
53 (green), 54 (red), 57 (blue) and 60 (white).
Therefore define the led nodes of the controller as white instead of RBG
as well as backlight as their function.
Christoph Krapp [Sat, 22 Nov 2025 19:42:33 +0000 (20:42 +0100)]
ipq40xx: fix second 5ghz radio on Netgear RBx40
When support was added for the RBR40 and RBS40 it was assumed that they
also share the same second 5ghz wifi chip as their bigger siblings.
Turns out that instead of QCA9984 (RBx50, SRx60) these devices use
QCA9886 like the RBx20 devices to.
They also load different boardfiles for the IPQ4019 chip.
This moves the wifi nodes from the orbi.dtsi to each device dts file and
change the RBx40 boardfile variants.
Christoph Krapp [Sat, 22 Nov 2025 19:20:14 +0000 (20:20 +0100)]
ipq40xx: split orbi devices in router and satellite
Netgear Orbi devices are split into router and satellite units. Even
though the hardware is mostly the same, the network configuration is
different. Router units have a designated WAN port while satellite units
have all available ports labeled as "Ethernet".
This splits the device trees into both unit types and adjusts the port
labels.
sunxi: image: sync target profiles names with DT compatible
Following up with errors reported in the ASU repo, these bananapi cases
do not match the DT compatible "bpi", sync with dts sources.
Also some profiles were overwriting SUPPORTED_DEVICES.
Sysupgrade would be failing in SUPPORTED_DEVICES check since
the DT compatible(/tmp/sysinfo/board_name) is not in SUPPORTED_DEVICES.
This should also fix errors when using ASU sysupgrade clients.
- Sync profile makefile target names with DT compatibles.
- Fix overwrites of SUPPORTED_DEVICES instead of appending.
- Adapt the uboot-sunxi profiles accordingly.
*bpi-p2-zero dts is still not upstream.
V2:
- Include fixes for arm926ejs(ARM926EJ-S) subtarget (LicheePi Nano and
PopStick v1.1) (profile rename for correct default SUPPORTED_DEVICES)
Fixes: https://forum.openwrt.org/t/luci-attended-sysupgrade-support-thread/230552/246 Fixes: https://github.com/openwrt/asu/issues/486 Fixes: https://github.com/openwrt/asu/issues/524 Fixes: 9aa66b8ce730aebff76d353392151708a897a3a0 "sunxi: add support for Banana Pi M2 Berry" Fixes: d5f615bf2a0434c15d13943b566d46f25da579bb "sunxi: add support for Sinovoip Banana Pi M2 Plus" Fixes: 3819c1638a9d300840d0f869628891f9696be112 "sunxi: Add support for Banana Pi M2 Ultra" Fixes: 6bf8193b25a147abfe3720104e63af890c1ca2b8 "sunxi: add support for Bananapi P2 Zero" Fixes: 80edfaf675364835e6d2e17d97ebec6afc6b2103 "sunxi: add support for Banana Pi M3" Fixes: 3c24a1d423a6052b101c00cb7d94e70d72702639 "sunxi: add support for NanoPi NEO Plus2 board" Fixes: a689307c970e37c247c6452ba3963be8b109eb4f "sunxi: build image/uboot for the NanoPi NEO2" Fixes: fde68cb80941a60be93ece75e808b5b407d11cc8 "sunxi: add support for FriendlyARM NanoPi R1S H5" Fixes: 3ec468ff4fa3106d459ae58ec4bc4833715118c6 "sunxi: add F1C100 (arm926ej-s) support" Signed-off-by: Mario Andrés Pérez <mapb_@outlook.com> Link: https://github.com/openwrt/openwrt/pull/21095 Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit d871e95e7f836435e8323074e3abab2c7608e016)
Martin Nadvornik [Mon, 15 Dec 2025 15:54:52 +0000 (16:54 +0100)]
mediatek: fix IPv4 address missing on interface in failsafe mode for cudy ap3000-v1
cudy ap3000-v1 did not work correctly in failsafe mode
because the address 192.168.1.1 was missing on the eth0 inteface.
it was reachable via it's IPv6 link-local address however.
this commit fixes the issue.
Rany Hany [Sat, 29 Nov 2025 00:04:09 +0000 (00:04 +0000)]
wifi-scripts: add support for using list for iface in wifi-station/vlan
This is a trivial change to allow users to use 'list' on iface.
Old wifi-scripts already implements this, so this just ensures
that shell-based and ucode wifi-scripts are on-par with each other.
Rany Hany [Sat, 29 Nov 2025 10:37:22 +0000 (12:37 +0200)]
wifi-scripts: fix broken match all case for wifi-vlan
When iface is omitted, wifi-vlan will apply to all interfaces.
However, netifd.set_vlan call is not successful as it assumes
that every wifi-vlan section corresponds to one VIF.
For this reason in the wifi-vlan case (cur_type == "vlan")
we create a composite key in the form `${vif.name}/${vlan.name}`
allowing the same vlan section to correspond to multiple VAPs.
`/` was decided as a delimiter as it is an invalid character
for a network interface name and UCI identifier; so it is
impossible for it to cause conflicts.
It was verified that the `ubus call network.wireless status`
works as expected with this change. Moreover, wifi-station
is not susceptible to this problem.
This also means that it is now possible for wifi-vlan
to support `list` iface similar to old shell-based wifi-scripts.
This will be done in a follow-up commit.
David Woodhouse [Sat, 29 Nov 2025 16:18:26 +0000 (16:18 +0000)]
image: add CONFIG_EXTRA_IMAGE_NAME
This allows an optional tag to be put in the .config file which is
included in the filename of the resulting images, so it's easier to
build images with different functionality for the same target hardware.
David Härdeman [Tue, 9 Dec 2025 22:49:48 +0000 (23:49 +0100)]
base-files: always generate default DUID
The previous logic was copied from 12_network-generate-ula, but fails to
account for upgrades where the "auto" value isn't set (it is set by
base-files/files/bin/config_generate). Fix this to always set the
default duid if it isn't set.
Also, rename the file to better reflect what it does.
Add the DDR4_4BG_MODE option, which supports 4GB DDR4 RAM
for the MT7987 and 8GB DDR4 RAM for the MT7988. If this mode
is not enabled, bl2 can only recognize half the size of RAM.
Based on 35f6d79, which introduced Watchguard Firebox T10 support.
The T10 and T15 are identical hardware, with the exception of the T15
having twice the flash and RAM size.
The T10-W and T15-W models have their Mini-PCIe slot populated with an ath9
(AR9582) based WiFi card. The slot is either unpopulated or empty for
non-WiFi models. All required drivers are present by default on the mpc85xx
target, so T10/T10-W resp. T15/T15-W can use the same OpenWrt image.
This commit also introduces the zImage loader from 7d768a9 to boot the
kernel. This is required, since the U-Boot version used in these devices
appears to have a hard limit of 16MB for the kernel size it can handle. The
current kernel size is around 17MB, though, due to kernel page alignment
required for memory protection.
Installation (replaces previous instructions for T10):
1. If the U-Boot password is known, proceed with step 2.
If the U-Boot password is unknown, dump the NOR flash using a SPI
programmer and patch the unknown password to a known one. You can use
blocktrron's Python script:
https://github.com/blocktrron/t10-uboot-patcher/
This script will patch the password to '1234' (without quotes).
Alternatively, you can search for the hashed password in the NOR dump
yourself and overwrite it with a known one. The SHA1 hash is:
6. The device should now boot OpenWrt from NAND flash. Enjoy.
Back to stock:
Use the vendor recovery procedure.
Stock recovery might also be necessary in case you have accidentally used
the fw_setenv command from within OpenWrt without using saveenv in U-Boot
first.
In order to use the vendor firmware recovery procedure, the NAND partitions
mtd3 to mtd6 must remain intact. Make sure not to overwrite them, or keep
dumps of them for later recovery.
projects / thirdparty / openwrt.git / log
