As shown above, the SPI master controller shares its pin with GPIO 8, 9,
10, 11, 12. In some upcoming devices (like the Plasma Cloud PSX28/ESX28),
they will be used for SFP cage signaling. These pins must therefore be
switched manually to the GPIO mode.
The SPI_CTRL0 register provides all necessary configuration to enforce the
GPIO mode of the pins. And until more requirements (and a correct driver)
for the SPI master controller arise, it is therefore possible to use
pinctrl-single to configure it using the devicetree.
Previously the ethernet driver did configure the SPI master controller for
31.25 MHz. It is unknown for which kind of device this was originally made
and what was actually connected there. But this manual write to the
register conflicts potentially with the write of the pinctrl driver to the
same register. Luckily, we don't need this SPI speed configuration in the
ethernet driver. Still, to allow this device an easy migration, the
`spi0-31mhz` configuration was already prepared.
mac80211: ath: improve ath10k "failed to flush transmit queue" errors
Currently, in busy environments, ath10k logs "failed to flush transmit
queue" errors and have a spiking CPU usage, making the wireless barely
usable.
With this patch, taken from https://patchwork.kernel.org/project/linux-wireless/patch/20250806070005.1429-1-hujy652@gmail.com/
this does not occur in normal operation and the wifi is much more stable
Felix Fietkau [Wed, 1 Oct 2025 08:08:44 +0000 (10:08 +0200)]
netifd: update to Git HEAD (2025-09-30)
c3bf8fd913a4 interface: fix reload for devices that point to vlan aliases 22216cac7c94 bridge: fix reload when ports refer to aliased vlans on another bridge ecca21ca07dd system: add logging wrappers for basic system functions
Sven Eckelmann [Tue, 30 Sep 2025 07:06:57 +0000 (09:06 +0200)]
realtek: Work around missing 10g-qxgmii PHY mode
The current SerDes implementation for RTL931x handles 10G-QXGMII via the
"usxgmii" PHY mode. This is not 100% correct because it is not a single
port with 10G (max) but 4 ports with 2.5G each.
To allow setting of the "10g-qxgmii" phy mode, just change the code for now
to use the same codepaths as USXGMII. This has to be cleaned up further
during the SerDes driver rewrites.
Suggested-by: Markus Stockhausen <markus.stockhausen@gmx.de> Signed-off-by: Sven Eckelmann <se@simonwunderlich.de> Link: https://github.com/openwrt/openwrt/pull/20239 Signed-off-by: Robert Marko <robimarko@gmail.com>
Sven Eckelmann [Wed, 30 Jul 2025 05:39:32 +0000 (07:39 +0200)]
realtek: rtl931x: Enable REALTEK_PHY for RTL8224 support
The Plasma Cloud PSX28 and ESX28 are using RTL8224 as ethernet PHY. This
phy works perfectly fine on PSX8/PSX10 (RTL930x) but failed to establish a link
on rtl931x because the upstream realtek phy driver was not enabled.
uboot-mediatek: rax3000m: use ddr3-1866mhz built in the dependency
The u-boot of cmcc_rax3000m_*-ddr3 declares a dependency on
trusted-firmware-a-mt7981-*-ddr3-1866mhz, while actually packages
the normal BL2. This causes build failure if the the non-1866m
variant isn't built.
According to the previous commit, it's intended to use the lower
frequency variant to fix device stability issues. Correct
BL2_DDRTYPE as intended.
Paul Spooren [Tue, 30 Sep 2025 12:17:11 +0000 (14:17 +0200)]
include: make APK packing mtime reproducible
APK kindly stores the mtime of each containing file in created packages,
breaking reproducibility. As a fix, touch all files of the package with the
timestamp of PKGSOURCE_DATE_EPOCH, which contains the timestamp based on the
last package modification.
Over at OPKG, something similar is done by setting mtime in the tar command,
see the `ipkg-build` script.
To tackle this in APK directly, some changes are suggested. However until this
is merged, we should fix it downstream.
https://gitlab.alpinelinux.org/alpine/apk-tools/-/merge_requests/348
realtek: rtl931x: set hash_msb based on VLAN ID when adding a new L2 entry
During testing, we discovered that when adding a new offload FDB rule
on certain VLANs and then delete it, does not work as expected.
Steps to Reproduce:
* Create VLAN 4094 on the port lan1:
bridge vlan add vid 4094 dev lan1 pvid
* Add a new FDB entry on port lan1 for VLAN 4094:
bridge fdb add 00:01:02:22:33:44 dev lan1 vlan 4094 master permanent
* Delete the new FDB entry on port lan1 for VLAN4094
bridge fdb del 00:01:02:22:33:44 dev lan1 vlan 4094 master permanent
Root Cause:
The failure occurs because the hash_msb flag is not set correctly
based on the VLAN ID when adding a new L2 entry.
Signed-off-by: Issam Hamdi <ih@simonwunderlich.de> Signed-off-by: Sven Eckelmann <se@simonwunderlich.de> Link: https://github.com/openwrt/openwrt/pull/20183 Signed-off-by: Robert Marko <robimarko@gmail.com>
Sven Eckelmann [Fri, 26 Sep 2025 09:57:20 +0000 (11:57 +0200)]
realtek: Switch booleans in rtl838x_l2_entry to single bits
In upstream kernel, it is not well received to use a lot of simple booleans
in structs. It is preferred to use 1-bit bitfields [1] and consolidate the
booleans together.
John Thomson [Thu, 16 May 2024 00:16:53 +0000 (10:16 +1000)]
image: adapt mikrotik yafut to allow switch to bootimage YAFFS path
Adjust the YAFFS file path written depending on the sysupgrade
filename.
Default to kernel (for ELF), switch to bootimage (for NPK) if image name
has v7.
Signed-off-by: John Thomson <git@johnthomson.fastmail.com.au>
John Thomson [Sun, 22 Jan 2023 11:34:43 +0000 (21:34 +1000)]
build: add Mikrotik NOR RouterBOOT v7 image build
Add build step 'kernel-pack-npk' which uses 'npk_pack_kernel' which is now
part of firmware-utils to enable wrapping the kernel inside a MikroTik NPK
package.
Signed-off-by: John Thomson <git@johnthomson.fastmail.com.au> Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Daniel Golle [Tue, 23 Sep 2025 22:30:03 +0000 (23:30 +0100)]
firmware-utils: update to git HEAD
7e6f69b npk_pack_kernel: add tool for creating MikroTik NPK kernel packages d3f8b6e zynsig: add new tool for creating images for the ZyXEL GS1920 series 9fa340d ptgen: create separate images for gpt data structure 78d8084 ptgen: allow image generation for a specified disk size 3f2d148 ptgen: allow non-default placement of gpt entry table f1f98b0 ptgen: do not create stub partition to fill a gap if gap caused by alignment 6ea8b6d ptgen: fix protective MBR partition size febfef7 ptgen: use long long instead of long for sizes 996dc48 ptgen: fix misprint and simplify calculation a bit 0725c3d build: require CMake >= 3.5 due to dropped legacy support
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
scripts/patch-kernel.sh: do not try to check files after patch
Since we are not using patch -b, *.orig files are only created when
there are conflicts, or never according to posix patch.
As such, it doesn't really make sense to always delete *.orig files
presuming they are patch backups, even if they are patch backups.
Doing so is both deleting potentially useful information for failed
patch applications and creating hard to diagnose bugs [1].
In a similar vein, checking for *.rej files does not add any value
since we're already checking the patch command's return code.
Felix Fietkau [Mon, 29 Sep 2025 07:45:45 +0000 (09:45 +0200)]
wpa_supplicant: MLO bringup order
Do not allow hostapd phy state update to bring up links as long between mld_set
and mld_start calls. Configuration on other PHYs could still be pending.
Felix Fietkau [Sun, 28 Sep 2025 18:55:57 +0000 (20:55 +0200)]
hostapd: fix an AP+STA corner case on MLO APs
Bring up AP interfaces, even if no frequency update was provided.
Fixes bringup when a MLO STA on the same radios connects to fewer links
than available, or to a non-MLD AP.
Add missing syscon compatible for scuclk node.
Fix the unit-address of the scuclk node.
This fixes the pcs driver error:
airoha-pcs 1fa08000.pcs: probe with driver airoha-pcs failed with error -22
airoha-pcs 1fa09000.pcs: probe with driver airoha-pcs failed with error -22
Fixes: c3d70b1 ("airoha: en7581: Add support for external PHY") Signed-off-by: Aleksander Jan Bajkowski <olek2@wp.pl> Signed-off-by: Chukun Pan <amadeus@jmu.edu.cn> Link: https://github.com/openwrt/openwrt/pull/20190 Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Chukun Pan [Fri, 26 Sep 2025 15:02:15 +0000 (23:02 +0800)]
airoha: fix pinctrl driver function bug
The pinctrl driver for airoha was expecting a function name that was
not a string, but was passed one. Removing #string fixed this issue.
Fixes: c5b12fc ("airoha: Introduce support for Airoha AN7583 SoC") Signed-off-by: Chukun Pan <amadeus@jmu.edu.cn> Link: https://github.com/openwrt/openwrt/pull/20190 Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Chukun Pan [Thu, 18 Sep 2025 14:30:18 +0000 (22:30 +0800)]
airoha: remove duplicate AN7581 PHY patch
This patch already exists in generic/backport-6.12.
Fixes: 122135b ("airoha: an7581: add support for kernel 6.12") Signed-off-by: Chukun Pan <amadeus@jmu.edu.cn> Link: https://github.com/openwrt/openwrt/pull/20190 Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Chukun Pan [Tue, 9 Sep 2025 15:06:02 +0000 (23:06 +0800)]
airoha: set default DEVICE_DTS/DEVICE_DTS_DIR
All devices under airoha use dts directory under the target,
so update the default DEVICE_DTS_DIR. Also set the default
DEVICE_DTS based on the SoC name for non-dev boards.
Add support for swapping the Serdes TX line on RTL8261N PHYs.
This is used on an Arcadyan Mozart board where the Serdes TX is swapped
on the PHY (instead of on the Soc) to permit support of SFP module by
using toggling the integrated MUX.
Felix Fietkau [Sun, 28 Sep 2025 14:25:01 +0000 (16:25 +0200)]
iwinfo: drop @!WIFI_SCRIPTS_UCODE dependency
Make the package empty if ucode wifi scripts are being used (since iwinfo is
provided by wifi-scripts in that case). This fixes packages that explicitly
select iwinfo
Do not always try to include iwinfo in the images when wpa supplicant or
Broadcom nas is also included. iwinfo is incompatible with current
default configuration.
iwinfo is only build when CONFIG_WIFI_SCRIPTS_UCODE is not set. If
CONFIG_WIFI_SCRIPTS_UCODE is not set kmod-cfg80211 depends on iwinfo,
so it should be included in all images with wifi drivers.
The CONFIG_WIFI_SCRIPTS_UCODE option was recently changed to be active
by default.
This should fix the current buildbot build failures.
This reverts commit 6435b8bb27ea ("build: include iwinfo by default
if nas or wpad(-mini) is selected")
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
All changes to the DTS have been applied upstream, remove our DTS copy.
This change effectively drops the workaround for the 2500Base-x
interface which was broken until kernel 5.15 and removes redundant
address-/size-cells.
boot: Introduce support for U-Boot support for Airoha EN7581/AN7583
Introduce support for U-Boot for Airoha EN7581/AN7583. For EN7581
initial patch are already in U-Boot mainline and doesn't require
backport, for AN7583 some patch are still pending but already posted
upstream.
Also add for now, precompiled binary for ATF BL2 and BL31. Support for
ATF is planned and will come later.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Introduce initial support for Airoha AN7583 SoC and add all the required
patch for basic functionality of the SoC.
Airoha AN7583 is based on Airoha EN7581 SoC with some major changes on
the PHY handling and Serdes. It can be see as a lower spec of EN7581
with modern and simplified implementations.
All the patch are sent upstream and are pending revision. Support for
PCIe and USB will come later as soon as DT structure is accepted
upstream.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
This add a pending version of the Airoha PCS driver to add support for
External PHY. The Airoha PCS driver will receive some minor modification
once we the PCS subsystem will be defined upstream.
Add all the required node for GDM2 and GDM4 and enable the PCS config.
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Jan Hoffmann [Thu, 24 Jul 2025 17:39:28 +0000 (19:39 +0200)]
realtek: support configuring SerDes auto-negotiation on RTL93xx
There are SFP modules which only work if auto-negotiation is disabled,
like some "OEM SFP-2.5G-T" modules. This also seems to be necessary for
RTL8226/RTL8221B PHYs when using 2500Base-X.
However, currently, it is always enabled, so add support for configuring
it to make these SFP modules and PHYs work.
This also adds locking which should be useful for future extension of
the PCS driver.
realtek: evaluate pcs-handle instead of sds property
In the Realtek dts the pcs-handle property at the switch port is the
successor of the sds property at the phy. Rearrange the MDIO and DSA
driver so they always look at the new attribute.
Remark! This code can be dropped completely if the new PCS driver
is fully featured. But this will take some time.
projects / thirdparty / openwrt.git / log
