projectsmirrors: remove kernel mirror ftp.riken.jp

This mirror has not been synchronized for over a year.

Signed-off-by: Shiji Yang <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19509
Signed-off-by: Hauke Mehrtens <[email protected]>

projectsmirrors: update Apache mirror apache.cs.utah.edu

The original URL no longer works.

Signed-off-by: Shiji Yang <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19509
Signed-off-by: Hauke Mehrtens <[email protected]>

projectsmirrors: remove some GNOME mirrors

These removed sites no longer provide GNOME mirror services.

Signed-off-by: Shiji Yang <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19509
Signed-off-by: Hauke Mehrtens <[email protected]>

kernel: bump 6.12 to 6.12.40

Changelog: https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.12.40
Removed upstreamed patches:
   generic/pending-6.12/680-net-fix-TCP-UDP-fraglist-GRO.patch
   generic/pending-6.12/802-nvmem-u-boot-env-align-endianness-of-crc32-values.patch
1- https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.12.40&id=7c532f222361191fe228e54c5d3e0026fef8a5a0
2- https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.12.40&id=c29a2328af96338d327cd851803338423c6f07a1
All other patches auto-refreshed.

Signed-off-by: Leo Barsky <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19514
Signed-off-by: Hauke Mehrtens <[email protected]>

kernel: bump 6.6 to 6.6.100

Changelog: https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.6.100
Manually refreshed:
    823-v6.12-0003-nvmem-layouts-add-U-Boot-env-layout.patch
All other patches auto-refreshed.

Signed-off-by: Leo Barsky <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19514
Signed-off-by: Hauke Mehrtens <[email protected]>

realtek: actually enable 2500Base-X

The SerDes setup function needs to be called to make 2500Base-X work.

Signed-off-by: Jan Hoffmann <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19517
Signed-off-by: Hauke Mehrtens <[email protected]>

tools: firmware-utils: update to Git HEAD (2025-07-24)

075cdc0 iptime-crc32: add support for ipTIME AX3000Q
48ababa iptime-crc32: add support for ipTIME AX3000SM
f29de74 iptime-crc32: Add device support for ipTIME AX6000M

Link: https://github.com/openwrt/openwrt/pull/19520
Signed-off-by: Hauke Mehrtens <[email protected]>

build: call ipkg-remove using xargs if #args>=512

The wildcard call to clean up luci package (luci*) can pick up over
2,300 files when the full tree is built. Running make package/luci/clean
or a second run of make package/luci/compile would fail with an
'Argument list too long' error.

To avoid that, a maybe_use_xargs function was created that runs the
command straight as usual if the number of arguments is < 512, or saves
the list in a temporary file and feeds it to xargs otherwise.

This is an update to current file names and resubmission of
https://lists.openwrt.org/pipermail/openwrt-devel/2020-February/027525.html

Fixes: https://github.com/openwrt/openwrt/issues/19510
Fixes: https://github.com/openwrt/luci/issues/7869
Authored-by: Eneas U de Queiroz <[email protected]>
Signed-off-by: Kuan-Yi Li <[email protected]>
Signed-off-by: Eric Fahlgren <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19516
Signed-off-by: Nick Hainke <[email protected]>

realtek: phy: add basic RTL8218B setup

On some devices (like ZyXEL GS1920) the phys are not initialized and patched
by the bootloader. This is done through the vendor SDK when the software
starts. To make these devices usable too, provide the most basic setup
sequence for the RTL8218B.

Signed-off-by: Markus Stockhausen <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19491
Signed-off-by: Hauke Mehrtens <[email protected]>

realtek: dsa: relax capability checks for 2.5G modes

The driver currently uses two checks to verify the capabilities. These
are ..._phylink_get_caps() and ..._pcs_validate(). For RTL930x these
must allow 2.5G modes. Enhance that as follows:

Add 2500BASEX to phylink_get_caps(). Sort the interfaces alphabetically
and rename the function to the new prefix. IMPORTANT REMARK! Until now
this function allowed the XGMII mode (10G only parallel interface) that
was somehow mixed with the Realtek proprietary mode XSGMII (10G SGMII).
Remove it to avoid further confusion.

Looking upstream pcs_validate() is used less and less. There are only
2 consumers left in 6.16 and the calling location reads:

/* Validate the link parameters with the PCS */
if (pcs->ops->pcs_validate) {
ret = pcs->ops->pcs_validate(pcs, supported, state);
if (ret < 0 || phylink_is_empty_linkmode(supported))
return -EINVAL;

/* Ensure the advertising mask is a subset of the
 * supported mask.
 */
linkmode_and(state->advertising, state->advertising,
     supported);
}

There is no need for this additional check. Drop the functions.

Tested-by: Jan Hoffmann <[email protected]>
Signed-off-by: Markus Stockhausen <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19429
Signed-off-by: Hauke Mehrtens <[email protected]>

mbedtls: restore previous configuration behavior

Fallback to default mbedtls configurations in case of the package is
not configured. It is possible for some reasons it get built even if
it's unselected because of build system bugs or other build-only
dependencies. In this case current behavior will comment out all
necessary configurations and lead build errors.

Fixes: 5359639c2b74 ("mbedtls: Apply configuration in Configure instead of Prepare")
Signed-off-by: Tianling Shen <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19495
Signed-off-by: Hauke Mehrtens <[email protected]>

ucode: add padding to uc_resource_ext_t

Fixes misaligned allocations for ext resources

Signed-off-by: Felix Fietkau <[email protected]>

udebug: update to Git HEAD (2025-07-23)

5d10084ea885 lib-ucode.c: add #define _GNU_SOURCE
a95364b41d52 udebug-cli: fix terminating uloop
c00eb9b685a8 ucode: use FILE handle for pcap output
4265167cb6e8 ucode: add error reporting to pcap_write
4a908ee731a6 udebug-cli: stop event loop on write failure
6e04f4187231 ucode: use ucv_resource_create_ex for remote rings
c297f04e1852 ucode: drop use ucv_resource_create
f207d37a1055 ucode: add support for specifying ring format
98683a94bcdd ucode: support appending array data, similar to socket.send()
a7ecd483ed38 ucode: allow calling udebug.init() multiple times
d4a4c788c416 ucode: fix allocation size of local ring meta
184706abaf50 ucode: add timestamp argument to foreach()
8442c948c193 ucode: add function for getting ring information
f4958a4c591a ucode: add const entries for enum udebug_format
14d4fec36993 udebug-cli: add logstream command
6ed8536142bb ucode: fix entries/size confusion

Signed-off-by: Felix Fietkau <[email protected]>

libubox: update to Git HEAD (2025-07-23)

b7acc8e6fd5e remove compatibility code for older json-c versions
a1acd99f2eb7 udebug: remove obsolete debug message
49056d178f42 udebug: fix issue with snapshot of remote ring

Signed-off-by: Felix Fietkau <[email protected]>

toolchain/gcc: prevent the use of LDRD/STRD on ARMv5TE

These instructions are for 64-bit load/store. On ARMv5TE, the CPU
requires addresses to be aligned to 64-bit. When misaligned, behavior is
undefined (effectively either loads the same word twice on LDRD, or
corrupts surrounding memory on STRD).

On ARMv6 and newer, unaligned access is safe.

Removing these instructions for ARMv5TE is necessary, because GCC
ignores alignment information in pointers and does unsafe optimizations
that have shown up as bugs in various places.

This patch was originally added more than 11 years ago in commit b050f87d13b5,
but got lost 6 years ago, when gcc 9.1 was added in 88c07c655262.

This primarily affects the kirkwood and ixp4xx targets

Signed-off-by: Felix Fietkau <[email protected]>

mediatek: Remove modules.builtin file

This file should be generated automatically at runtime by the kernel
build system.

Link: https://github.com/openwrt/openwrt/pull/19473
Signed-off-by: Hauke Mehrtens <[email protected]>

ramips: EAP615-Wall v1: fix bootloop by reducing LZMA dictionary

Fix bootloop on TP-Link EAP615-Wall v1 by reducing LZMA dictionary
size. Before this patch and after an upgrade to kernel 6.12 this
device couldn't boot a kernel because of an error: "lzma compressed:
uncompress error 1".

I have chosen -d22 as dictionary size as suggested by @namiltd.
The usual sizes for problematic devices are -d16, -d20, -d22. I
have confirmed with my tests that this device can boot with a value
up to -d27, but there is no size benefit from values above -d21,
therefore -d22 is good enough.

See also: https://github.com/openwrt/openwrt/issues/19403

Signed-off-by: Marcin FM <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19433
Signed-off-by: Hauke Mehrtens <[email protected]>

mediatek: filogic: add support for ipTIME AX3000Q

Specification
-------------
- SoC       : MediaTek MT7981BA dual-core ARM Cortex-A53 1.3GHz
- RAM       : DDR3 256Mbytes, ESMT M15T2G16128A
- Flash     : 128Mbytes NAND Flash, ESMT F50L1G41LB
- WLAN      : MediaTek MT7976CN dual-band Wi-Fi 6
  - 2.4GHz  : b/g/n/ax, MU-MIMO
  - 5GHz    : a/n/ac/ax, MU-MIMO
- Ethernet  : MediaTek MT7531AE
  - LAN     : 10/100/1000 Mbps x4
  - WAN     : 10/100/1000 Mbps x1
- UART      : 1x4 pin header on PCB
  - [J6] TX, RX, GND, 3.3V (115200, 8N1)
- Buttons   : WPS, Reset
- LEDs      : 1x CPU (Amber)
              1x Wi-Fi 5GHz (Amber)
              1x Wi-Fi 2.4GHz (Amber)
              1x WAN activity (Amber)
              4x LAN activity (Amber)
- Power     : 12VDC, 1A (Center positive polarity)

MAC address
-----------
+-----------+-------------------+-----------------------+
| Interface | MAC               | Algorithm             |
+-----------+-------------------+-----------------------+
| WLAN 2.4G | B0:38:6C:48:xx:xx | label                 |
| WLAN 5G   | B2:38:6C:48:xx:xx | label with LA Bit Set |
| WAN       | B0:38:6C:48:xx:xx | label + 1             |
| LAN       | B0:38:6C:48:xx:xx | label + 3             |
+-----------+-------------------+-----------------------+
The WLAN 2.4G MAC was found in 'Factory' partition, 0x4

Installation
------------
1. Download the OEM recovery software from the manufacturer's website
2. Download the *squashfs-factory.bin file from the OpenWrt website
3. Press a reset button, and power up the router(keep pressing the reset button)
4. Wait more than 10 seconds until the CPU LED stop blinking
5. Connect the router(LAN port) to the PC
6. Replace a file in the OEM recovery software with the file from step 2
7. Run the OEM recovery software and follow the instructions
8. Wait for the router to boot from *squashfs-factory.bin

Signed-off-by: Donghyun Ko <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19368
Signed-off-by: Hauke Mehrtens <[email protected]>

ca-certificates: Update to 20250419

upstream changelog:

[ Alexander Kanavin ]
* update-ca-certificates: add a --sysroot option

[ Julien Cristau ]
* Update Mozilla certificate authority bundle to version 2.74.
  The following certificate authorities were added (+):
  + D-TRUST BR Root CA 2 2023
  + D-TRUST EV Root CA 2 2023
  The following certificate authorities were removed (-):
  - Entrust Root Certification Authority - G4
  - SecureSign RootCA11
  - Security Communication RootCA3
  - SwissSign Silver CA - G2

Signed-off-by: Fengyu Wu <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19462
Signed-off-by: Hauke Mehrtens <[email protected]>

rockchip: fix eMMC corruption on NanoPC-T6 with A3A444 chips

Some NanoPC-T6 boards with A3A444 eMMC chips experience I/O errors and
corruption when using HS400 mode. Downgrade to HS200 mode to ensure
stable operation.

Fixes: #18844
Signed-off-by: Grzegorz Sterniczuk <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19398
[Add patch header]
Signed-off-by: Hauke Mehrtens <[email protected]>

wireless-regdb: Update to version 2025.07.10

6953f19 wireless-regdb: Update regulatory info for Indonesia (ID) for 2025
2e8214e wireless-regdb: Permit 320 MHz bandwidth in 6 GHz band for GB
a94f685 wireless-regdb: Update regulatory info for Egypt (EG) for 2024
7628ce2 wireless-regdb: Update regulatory rules for Brazil (BR) on 6GHz
4411b39 wireless-regdb: Update regulatory info for Vietnam (VN) for 2025
490f136 wireless-regdb: Update regulatory info for Estonia (EE) for 2024
c56c663 wireless-regdb: update regulatory rules for Paraguay (PY) on 6 GHz for 2025
5a8ced5 wireless-regdb: Update regulatory info for CEPT countries for 6GHz listed by WiFi Alliance
5fd8ee3 wireless-regdb: update regulatory rules for Bosnia and Herzegovina (BA) for 6 GHz
e05260a wireless-regdb: update regulatory database based on preceding changes

Link: https://github.com/openwrt/openwrt/pull/19474
Signed-off-by: Hauke Mehrtens <[email protected]>

mediatek: filogic: increase nand flash speed on Redmi AX6000

This commit increases the SPI bus frequency from 20 to 52 MHz. Reduces boot
time by 2s. Below is a performance comparison.

Before:
root@OpenWrt:~# dd if=/dev/mtd5 of=/dev/null bs=10M count=1 status=progress
10485760 bytes (10 MB, 10 MiB) copied, 2 s, 5.8 MB/

After:
root@OpenWrt:~# dd if=/dev/mtd5 of=/dev/null bs=10M count=1 status=progress
10485760 bytes (10 MB, 10 MiB) copied, 1 s, 9.7 MB/s

Taken from PR #18752 as each device should be tested individually, so I have
created a separate PR for this.

Signed-off-by: Sky Huang <[email protected]>
Signed-off-by: Aleksander Jan Bajkowski <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19439
Signed-off-by: Hauke Mehrtens <[email protected]>

Note inspiration for this script to give credit where it's due

There is some confusion in the Git log of this file.  Specifically,
the commit message on 7b7f1702 incorrectly indicates that there was
"potential fear" about copyright infringement.

Upon review of this situation, there is, in my opinion, no concern of
copyright infringement here.  This is not legal advice; it is my
opinion based of years of work on copyright policy for FOSS.

However, Elliott Mitchell's idea was obviously helpful as inspiration
in writing this script and deserves credit.  Ideas alone, however,
are not to my knowledge copyrightable anywhere in the world.

Signed-off-by: Bradley M. Kuhn <[email protected]>

mediatek: disable MediaTek SoC Information (mtk-socinfo) driver

With the switch to Linux 6.12 this driver was enabled by accident.
However, it doesn't support any of the router SoCs but is meant to be
used only by the smartphone, tablet and chromebook SoCs.
Disable the driver again to silence a kernel error logline during boot.

Fixes: 0a0f5f94ec ("mediatek: mt7623: update config-6.12")
Fixes: 029b7ed9c4 ("mediatek: mt7622: update config-6.12")
Reported-by: https://aparcar.org/openwrt-tests/119/
Signed-off-by: Daniel Golle <[email protected]>

malta: drop 6.6 support

Drop configs and patches for Linux 6.6.

Link: https://github.com/openwrt/openwrt/pull/19414
Signed-off-by: Hauke Mehrtens <[email protected]>

malta: use kernel 6.12 by default

Switch to Linux kernel version 6.12.

Link: https://github.com/openwrt/openwrt/pull/19414
Signed-off-by: Hauke Mehrtens <[email protected]>

mediatek: add support for ELECOM WRC-X3000GS3

ELECOM WRC-X3000GS3 is a 2.4/5 GHz band 11ax (Wi-Fi 6) router, based on
MT7981B.

Specification:

- SoC             : MediaTek MT7981B
- RAM             : DDR3 512 MiB (Winbond W634GU6QB-09)
- Flash           : SPI-NOR 128 MiB (Winbond W25N01GVZEIG)
- WLAN            : 2.4/5 GHz 2T2R (MediaTek MT7981B (SoC))
- Ethernet        : 5x 10/100/1000 Mbps
  - wan (phy)     : MediaTek MT7981B (SoC)
  - lan (switch)  : MediaTek MT7531A
- LEDs/Keys (GPIO): 8x/4x
- UART            : through-hole on PCB (J500)
  - assignment    : 3.3V, TX, RX, NC, GND from tri-angle marking
  - settings      : 115200n8
- Power           : 12 VDC, 1 A (Max. 12.8 W)

Flash instruction using factory.bin image:

1. Boot WRC-X3000GS3 in router mode normally
2. Access to the WebUI ("http://192.168.2.1/") on the device and open
   the firmware update page ("ファームウェア更新")
3. Select the OpenWrt factory.bin image and click apply ("適用") button
4. Wait ~120 seconds to complete flashing

Switching to the stock firmware:

1. Load the elecom.sh script

   . /lib/upgrade/elecom.sh

2. Check the current index of firmware partition

   mstc_rw_bootnum

3. Set the bootnum to opposite value between 1 and 2

   mstc_rw_bootnum <value>

   example:

   - step2 returned "1": mstc_rw_bootnum 2

   - step2 returned "2": mstc_rw_bootnum 1

4. Reboot

Notes:

- ELECOM sells (or sold) multiple models as AX3000 class with different
  hardwares:

  - WRC-X3000GS(N) : Lantiq(Intel) GRX350/GRX550
  - WRC-X3000GS(T)2: Qualcomm IPQ5018
  - WRC-X3000GS3   : MediaTek MT7981B

MAC Addresses:

LAN   : 38:97:A4:xx:xx:40 (Factory, 0x2A(hex)/Ubootenv,"ethaddr"(text))
WAN   : 38:97:A4:xx:xx:43 (Factory, 0x24(hex))
2.4GHz: 38:97:A4:xx:xx:41 (Factory, 0x4 (hex))
5GHz  : 38:97:A4:xx:xx:42 (Factory, 0xA (hex))

Signed-off-by: INAGAKI Hiroshi <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18976
Signed-off-by: Hauke Mehrtens <[email protected]>

mediatek: filogic: enable mstc-boot mtdsplit parser

Enable mstc-boot mtdsplit parser on mediatek/filogic subtarget to handle
dual-boot on ELECOM WRC-X3000GS3.

Signed-off-by: INAGAKI Hiroshi <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18976
Signed-off-by: Hauke Mehrtens <[email protected]>

generic: add mstc-boot mtdsplit parser

Add new mtdsplit parser "mstc-boot" for the devices manufactured by MSTC
(Mitra Star Technology Corp.). This is necessary to handle dual-boot on
those devices.
This parser splits kernel+rootfs or only rootfs(or UBI) based on the
image in the firmware partition or pre-defined partitions in dts, and
"bootnum" value in the "persist" (or "working") partition.

Note: "bootnum" is used for switching active firmware partitions on the
      devices manufactured by MSTC and '1' or '2' are used on most
      devices. But some devices use '0' or '1'. (example: I-O DATA
      WN-DEAX1800GR)

Sequence:

1. obtain "bootnum" value
2. child nodes exsist (regardless of bootnum)
   -> fixed partitions
      (active parts  : without bootnum (ex.: "kernel", "rootfs")
       inactive parts: with bootnum (ex.: "kernel2", "rootfs2"))
3. current partition is active (dt bootnum == mtd bootnum)
   -> image-based partitions

Device Tree:

- common

  - mstc,bootnum  : "bootnum" value for the mtd partition (0/1/2)
  - mstc,persist  : phandle of "persist" partition containing "bootnum"
                    value

- fixed partitions

  - #address-cells: indicate cell count of address of child nodes (1)
  - #size-cells   : indicate cell count of size of child nodes (1)
  - (child nodes) : define the child partitions
    - reg         : define the offset and size
    - label-base  : define the base name of the partition
      - (example) : base:"kernel"->"kernel"(active)/"kernel2"(inactive)

  example:

  partition@3c0000 {
   compatible = "mstc,boot";
   reg = <0x3c0000 0x3240000>;
   label = "firmware1";
   mstc,bootnum = <1>;
   mstc,persist = <&mtd_persist>;
   #address-cells = <1>;
   #size-cells = <1>;

   partition@0 {
   reg = <0x0 0x800000>;
   label-base = "kernel";
   };

   partition@800000 {
   reg = <0x800000 0x2a40000>;
   label-base = "ubi";
  };
  };

- image-based partitions

  (no additional properties)

  example:

  partition@5a0000 {
   compatible = "mstc,boot";
   label = "firmware1";
   reg = <0x5a0000 0x3200000>;
   mstc,bootnum = <1>;
   mstc,persist = <&mtd_persist>;
  };

Signed-off-by: INAGAKI Hiroshi <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18976
Signed-off-by: Hauke Mehrtens <[email protected]>

libxml2: update to 2.14.5

Release Notes:
    https://gitlab.gnome.org/GNOME/libxml2/-/releases/v2.13.7
    https://gitlab.gnome.org/GNOME/libxml2/-/releases/v2.13.8
    https://gitlab.gnome.org/GNOME/libxml2/-/releases/v2.14.3
    https://gitlab.gnome.org/GNOME/libxml2/-/releases/v2.14.4
    https://gitlab.gnome.org/GNOME/libxml2/-/releases/v2.14.5

Fixes:
CVE-2025-32415 CVE-2025-32414 CVE-2025-6170 CVE-2025-49794 CVE-2025-49795 CVE-2025-49796

Build system: x86/64
Build-tested: x86/64-glibc
Run-tested: x86/64-glibc

Signed-off-by: John Audia <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19383
Signed-off-by: Hauke Mehrtens <[email protected]>

ipq40xx: enable 6.12 testing kernel

The 6.12 testing kernel for ipq40xx target is ready now.

Tested on SKSpruce WIA3300-20.

Signed-off-by: Shiji Yang <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18725
Signed-off-by: Hauke Mehrtens <[email protected]>

ipq40xx: migrate wifi configuration device paths for 6.12 kernel

The device tree PCIe host node name has been changed in the new
6.12 kernel[1]. Hence we have to update the wifi device path to
make sure it can work properly.

This script is based on:
target/linux/qualcommax/ipq807x/base-files/etc/hotplug.d/ieee80211/05-wifi-migrate

[1] https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=linux-6.12.y&id=07299ba2e7d98045e6b522f7c5b97f402b15bc82
Signed-off-by: Shiji Yang <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18725
Signed-off-by: Hauke Mehrtens <[email protected]>

ipq40xx: dts: fix incorrect PCIe bus number

On the ipq40xx platform, the bus range is 1 - 255. Therefore, bus
number 0 is clearly incorrect. Change the bus number to 1, just
like other ipq40xx devices. This patch fixes the following dtc
warnings on 6.12 kernel:

qcom-ipq4019.dtsi:476.5-29: Warning (pci_device_bus_num): /soc/pcie@40000000/pcie@0/wifi@0,0:bus-range: PCI bus number 0 out of range, expected (1 - 255)
qcom-ipq4019.dtsi:476.5-29: Warning (pci_device_bus_num): /soc/pcie@40000000/pcie@0/wifi@0,0:bus-range: PCI bus number 0 out of range, expected (1 - 255)

Signed-off-by: Shiji Yang <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18725
Signed-off-by: Hauke Mehrtens <[email protected]>

ipq40xx: dts: correct PCIe device node name

According to the dtc source code, the PCIe device node unitname
needs to follow the following naming rules:

```
reg = fdt32_to_cpu(cells[0]);
dev = (reg & 0xf800) >> 11;
func = (reg & 0x700) >> 8;
snprintf(unitname, sizeof(unitname), "%x,%x", dev, func);
```

These devices' reg cell[0] is equal to 0x10000, hence the correct
node unitname should be "0,0". This patch fixes the following dtc
warnings for 6.12 kernel:

qcom-ipq4019-map-ac2200.dts:211.11-216.4: Warning (pci_device_reg): /soc/pcie@40000000/pcie@0/wifi@1,0: PCI unit address format error, expected "0,0"
qcom-ipq4019-rt-ac42u.dts:315.11-319.4: Warning (pci_device_reg): /soc/pcie@40000000/pcie@0/wifi@1,0: PCI unit address format error, expected "0,0"
qcom-ipq4019-fritzbox-7530.dts:318.10-321.4: Warning (pci_device_reg): /soc/pcie@40000000/pcie@0/dsl@1,0: PCI unit address format error, expected "0,0"
qcom-ipq4019-fritzrepeater-3000.dts:239.11-245.4: Warning (pci_device_reg): /soc/pcie@40000000/pcie@0/wifi@1,0: PCI unit address format error, expected "0,0"
qcom-ipq4019-gl-b2200.dts:340.11-348.4: Warning (pci_device_reg): /soc/pcie@40000000/pcie@0/wifi@1,0: PCI unit address format error, expected "0,0"
qcom-ipq4019-ea8300.dts:97.11-102.4: Warning (pci_device_reg): /soc/pcie@40000000/pcie@0/wifi@1,0: PCI unit address format error, expected "0,0"
qcom-ipq4019-mr8300.dts:83.11-88.4: Warning (pci_device_reg): /soc/pcie@40000000/pcie@0/wifi@1,0: PCI unit address format error, expected "0,0"
qcom-ipq4019-whw03v2.dts:234.11-241.4: Warning (pci_device_reg): /soc/pcie@40000000/pcie@0/wifi@1,0: PCI unit address format error, expected "0,0"
qcom-ipq4029-insect-common.dtsi:282.11-287.4: Warning (pci_device_reg): /soc/pcie@40000000/pcie@0/wifi@1,0: PCI unit address format error, expected "0,0"
qcom-ipq4029-insect-common.dtsi:282.11-287.4: Warning (pci_device_reg): /soc/pcie@40000000/pcie@0/wifi@1,0: PCI unit address format error, expected "0,0"
qcom-ipq4019-lbr20.dts:488.11-495.4: Warning (pci_device_reg): /soc/pcie@40000000/pcie@0/wifi@1,0: PCI unit address format error, expected "0,0"
qcom-ipq4019-orbi.dtsi:317.11-322.4: Warning (pci_device_reg): /soc/pcie@40000000/pcie@0/wifi@1,0: PCI unit address format error, expected "0,0"
qcom-ipq4019-orbi.dtsi:317.11-322.4: Warning (pci_device_reg): /soc/pcie@40000000/pcie@0/wifi@1,0: PCI unit address format error, expected "0,0"
qcom-ipq4019-orbi.dtsi:317.11-322.4: Warning (pci_device_reg): /soc/pcie@40000000/pcie@0/wifi@1,0: PCI unit address format error, expected "0,0"
qcom-ipq4019-orbi.dtsi:317.11-322.4: Warning (pci_device_reg): /soc/pcie@40000000/pcie@0/wifi@1,0: PCI unit address format error, expected "0,0"
qcom-ipq4019-orbi.dtsi:317.11-322.4: Warning (pci_device_reg): /soc/pcie@40000000/pcie@0/wifi@1,0: PCI unit address format error, expected "0,0"
qcom-ipq4019-orbi.dtsi:317.11-322.4: Warning (pci_device_reg): /soc/pcie@40000000/pcie@0/wifi@1,0: PCI unit address format error, expected "0,0"
qcom-ipq4019-a62.dts:218.11-225.4: Warning (pci_device_reg): /soc/pcie@40000000/pcie@0/wifi@1,0: PCI unit address format error, expected "0,0"
qcom-ipq4019-pa2200.dts:198.11-205.4: Warning (pci_device_reg): /soc/pcie@40000000/pcie@0/wifi@1,0: PCI unit address format error, expected "0,0"
qcom-ipq4019-mf18a.dts:478.11-484.4: Warning (pci_device_reg): /soc/pcie@40000000/pcie@0/wifi@1,0: PCI unit address format error, expected "0,0"
qcom-ipq4019-mf289f.dts:429.11-435.4: Warning (pci_device_reg): /soc/pcie@40000000/pcie@0/wifi@1,0: PCI unit address format error, expected "0,0"

Signed-off-by: Shiji Yang <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18725
Signed-off-by: Hauke Mehrtens <[email protected]>

ipq40xx: dts: compatibility changes for 6.12 kernel

- Remove duplicate PCIe bridge. A default bridge node has been
  added upstream.
- Do not refer to the qca807x node. This node label has been
  removed. We can disable the qca807x phy in mdio node.

Signed-off-by: Shiji Yang <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18725
Signed-off-by: Hauke Mehrtens <[email protected]>

ipq40xx: add PCIe bridge node reference label for ipq4019.dtsi

Add a bridge node label so that we can insert PCIe peripheral nodes.

Signed-off-by: Shiji Yang <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18725
Signed-off-by: Hauke Mehrtens <[email protected]>

ipq40xx: tag_oob/ipqess: update DSA APIs

In the new 6.12 LTS kernel, master/slave keywords were replaced
by conduit/user.

Link: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=linux-6.12.y&id=6ca80638b90cec66547011ee1ef79e534589989a
Signed-off-by: Shiji Yang <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18725
Signed-off-by: Hauke Mehrtens <[email protected]>

ipq40xx: refresh 6.12 kernel config files

Manually selected symbols:

- Enable ARCH_QCOM_RESERVE_SMEM
  Reserve SMEM at the beginning of RAM

- Enable QCOM_TZMEM_MODE_GENERIC
  TrustZone interface memory allocator mode

- Disable QCOM_QSEECOM
  Qualcomm QSEECOM interface driver

- Disable IPQ_NSSCC_QCA8K
  QCA8K(QCA8386 or QCA8084) NSS Clock Controller

- Disable MSM_IOMMU
  MSM IOMMU Support

- Disable INTERCONNECT_QCOM
  Qualcomm Network-on-Chip interconnect drivers

All other symbols are automatically refreshed by
`make kernel_oldconfig CONFIG_TARGET=target` and
`make kernel_oldconfig CONFIG_TARGET=subtarget`.

Signed-off-by: Shiji Yang <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18725
Signed-off-by: Hauke Mehrtens <[email protected]>

ipq40xx: refresh 6.12 kernel patches

Remove upstreamed:
004-v6.7-firmware-qcom_scm-disable-SDI-if-required.patch [1]
709-ARM-dts-qcom-ipq4019-add-QCA8075-PHY-Package-nodes.patch [2]

Manually rebased:
422-firmware-qcom-scm-fix-SCM-cold-boot-address.patch
701-net-dsa-add-out-of-band-tagging-protocol.patch
850-soc-add-qualcomm-syscon.patch
900-PCI-qcom-add-hack-compatible-for-ipq4019-Lantiq-DSL.patch
910-Revert-firmware-qcom_scm-Clear-download-bit-during-r.patch

[1] https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=linux-6.12.y&id=ff4aa3bc98258a240b9bbab53fd8d2fb8184c485
[2] https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=linux-6.12.y&id=2338f4315f16b937e924ff679b91bb8c0ab53f25
Signed-off-by: Shiji Yang <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18725
Signed-off-by: Hauke Mehrtens <[email protected]>

ipq40xx: restore kernel 6.6 arch files

Copy arch directory from 6.12 kernel to restore the default
6.6 device tree files.

Signed-off-by: Shiji Yang <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18725
Signed-off-by: Hauke Mehrtens <[email protected]>

ipq40xx: restore kernel 6.6 config files and patches

Copy patches and kernel configs from 6.12 kernel to restore the
default 6.6 kernel support files.

Signed-off-by: Shiji Yang <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18725
Signed-off-by: Hauke Mehrtens <[email protected]>

ipq40xx: move arch files to 6.12 directory

This is a preparation for 6.12 kernel support. It can help us
track the device tree files history by using the Git tool.

Signed-off-by: Shiji Yang <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18725
Signed-off-by: Hauke Mehrtens <[email protected]>

ipq40xx: rename patchset and kernel configs to 6.12

This is a preparation for 6.12 kernel support. It can help us
track the patches and Kconfig history by using the Git tool.

Signed-off-by: Shiji Yang <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18725
Signed-off-by: Hauke Mehrtens <[email protected]>

ipq40xx: refresh kernel symbol configs

This is a preparation for introducing the 6.12 kernel support.
All configs are automatically refreshed. In theory, they will
generate the same .config files in the kernel build directory
as before.

Signed-off-by: Shiji Yang <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18725
Signed-off-by: Hauke Mehrtens <[email protected]>

ipq40xx: qca8k-ipq4019: add missing header platform_device.h

The prototype of `devm_platform_ioremap_resource_byname()` was defined
in `linux/platform_device.h`.

Signed-off-by: Shiji Yang <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18725
Signed-off-by: Hauke Mehrtens <[email protected]>

ipq40xx: ipqess: remove explicit phylink_generic_validate() references

The phylink_generic_validate() can be automatically called if
phylink_validate method was not given in the switch operations.

Link: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=linux-6.6.y&id=5938bce4b6e2146d5194badd7ecf0acc6bf877fc
Signed-off-by: Shiji Yang <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18725
Signed-off-by: Hauke Mehrtens <[email protected]>

ipq40xx: gcc-ipq4019: mark qcom_find_freq_close() function as static

This function is only used by the current source file gcc-ipq4019.c.

Signed-off-by: Shiji Yang <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18725
Signed-off-by: Hauke Mehrtens <[email protected]>

ipq40xx: dts: fix dtc warnings for MikroTik hAP ac3 LTE6

Correct the mtd partition reg property size to fix the following
dtc warnings:

qcom-ipq4019-hap-ac3-lte6-kit.dts:202.5-25: Warning (reg_format): /soc/spi@78b5000/flash@0/partitions/partition@0:reg: property has invalid length (8 bytes) (#address-cells == 2, #size-cells == 1)
qcom-ipq4019-hap-ac3-lte6-kit.dts:209.5-29: Warning (reg_format): /soc/spi@78b5000/flash@0/partitions/partition@80000:reg: property has invalid length (8 bytes) (#address-cells == 2, #size-cells == 1)
qcom-ipq4019-hap-ac3-lte6-kit.dts:226.5-31: Warning (reg_format): /soc/spi@78b5000/flash@0/partitions/partition@110000:reg: property has invalid length (8 bytes) (#address-cells == 2, #size-cells == 1)

Signed-off-by: Shiji Yang <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18725
Signed-off-by: Hauke Mehrtens <[email protected]>

ipq40xx: dts: fix led-controller node name for Netgear LBR20

The node name should include the reg property value. Fix warning:

qcom-ipq4019-lbr20.dts:415.17-449.4: Warning (i2c_bus_reg): /soc/i2c@78b7000/led-controller: I2C bus unit address format error, expected "27"

Signed-off-by: Shiji Yang <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18725
Signed-off-by: Hauke Mehrtens <[email protected]>

bcm27xx: package modules for official 7" touchscreen TFT display

Add driver modules for the official RaspberryPi 7" touchscreen TFT display.

Signed-off-by: Daniel Golle <[email protected]>

kernel: modules: video: package driver for TC358762

Create module package for Toshiba TC358762 DSI/DPI bridge driver.

Signed-off-by: Daniel Golle <[email protected]>

kernel: modules: video: individually package simple-panel.ko

The DRM modules for simple (ie. raw, non-eDP) panels can be used also
by platforms others than Freescale i.MX.
Create an individual package for it.

Signed-off-by: Daniel Golle <[email protected]>

targets: add USES_PM auto-feature

Not all targets support power management, some older or more simple
targets don't have CONFIG_PM set. Allow kernel module packages to
depend on USES_PM to only be available on targets which got
CONFIG_PM=y in their kernel config.

Signed-off-by: Daniel Golle <[email protected]>

uboot-mediatek: restore Ethernet PHY drivers

Partially revert commit 332645a610 ("uboot-mediatek: sync with
mtk-openwrt/u-boot 20250711") by dropping unused PHY drivers which come
with large embedded firmware (and are hence inacceptable for upstream)
and restoring our previous version of the Airoha EN8811H driver which is
known to work and load firmware from the eMMC boot1 hardware partition
instead of embedding it in the driver.

In future we may switch to the upstream Airoha EN8811H driver which
allows loading the firmware either from a filesystem or overloading the
en8811h_read_fw() for board which may eg. store that firmware directly
inside a eMMC hardware partition or UBI volume, like the BPi-R3 mini
does.

See also:
https://patchwork.ozlabs.org/project/uboot/patch/20250615131243[email protected]/

Fixes: 332645a610 ("uboot-mediatek: sync with mtk-openwrt/u-boot 20250711")
Reported-by: Adrian Bente
Signed-off-by: Daniel Golle <[email protected]>

cli: fix editor remove call validation

Skip type validation, fix passing values to remove

Signed-off-by: Felix Fietkau <[email protected]>

tools/llvm: update to 20.1.8

- release notes:
  - https://discourse.llvm.org/t/18-1-8-released/79725
  - https://discourse.llvm.org/t/llvm-19-1-0-released/81285
  - https://discourse.llvm.org/t/llvm-19-1-1-released/82321
  - https://discourse.llvm.org/t/llvm-19-1-2-released/82550
  - https://discourse.llvm.org/t/llvm-19-1-3-relased/82829
  - https://discourse.llvm.org/t/llvm-19-1-4-released/83230
  - https://discourse.llvm.org/t/llvm-19-1-5-released/83455
  - https://discourse.llvm.org/t/llvm-19-1-6-released/83723
  - https://discourse.llvm.org/t/llvm-19-1-7-released/84062
  - https://discourse.llvm.org/t/llvm-20-1-0-released/85122
  - https://discourse.llvm.org/t/llvm-20-1-1-released/85337
  - https://discourse.llvm.org/t/llvm-20-1-2-released/85654
  - https://discourse.llvm.org/t/llvm-20-1-3-released/85874
  - https://discourse.llvm.org/t/llvm-20-1-4-released/86147
  - https://discourse.llvm.org/t/llvm-20-1-5-released/86371
  - https://discourse.llvm.org/t/llvm-20-1-6-released/86598
  - https://discourse.llvm.org/t/llvm-20-1-7-released/86855
  - https://discourse.llvm.org/t/llvm-20-1-8-released/87259
- remove "100-gcc15.patch" (not needed anymore)
- refresh configuration options

Signed-off-by: Konstantin Demin <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19417
Signed-off-by: Hauke Mehrtens <[email protected]>

policycoreutils: bump PKG_RELEASE

improves 6cac528f7227 "policycoreutils: adjust with glibc and libcrypt-compat"

Signed-off-by: Konstantin Demin <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19404
Signed-off-by: Hauke Mehrtens <[email protected]>

rpcd: bump PKG_RELEASE

improves 3a188488197d "rpcd: adjust with glibc and libcrypt-compat"

Signed-off-by: Konstantin Demin <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19404
Signed-off-by: Hauke Mehrtens <[email protected]>

libpcap: bump PKG_RELEASE

improves 34c2e7701a9a "libpcap: adjust with glibc and libcrypt-compat"

Signed-off-by: Konstantin Demin <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19404
Signed-off-by: Hauke Mehrtens <[email protected]>

uhttpd: bump PKG_RELEASE

improves ec62e87c0c01 "uhttpd: adjust with glibc and libcrypt-compat"

Signed-off-by: Konstantin Demin <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19404
Signed-off-by: Hauke Mehrtens <[email protected]>

ppp: bump PKG_RELEASE

improves ea6bcebc417c "ppp: adjust with glibc and libcrypt-compat"

Signed-off-by: Konstantin Demin <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19404
Signed-off-by: Hauke Mehrtens <[email protected]>

dropbear: bump PKG_RELEASE

improves eecc53cbe5c2 "dropbear: adjust with glibc and libcrypt-compat"

Signed-off-by: Konstantin Demin <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19404
Signed-off-by: Hauke Mehrtens <[email protected]>

busybox: bump PKG_RELEASE

improves f0e7868b7e0c "busybox: adjust with glibc and libcrypt-compat"

Signed-off-by: Konstantin Demin <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19404
Signed-off-by: Hauke Mehrtens <[email protected]>

toolchain: bump PKG_RELEASE

improves c571ac4bdf25 "toolchain: adjust with recent glibc"

Signed-off-by: Konstantin Demin <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19404
Signed-off-by: Hauke Mehrtens <[email protected]>

armsr: drop 6.6 support

Drop configs and patches for Linux 6.6.

Link: https://github.com/openwrt/openwrt/pull/19413
Signed-off-by: Hauke Mehrtens <[email protected]>

armsr: use kernel 6.12 by default

Switch to Linux kernel version 6.12.

Link: https://github.com/openwrt/openwrt/pull/19413
Signed-off-by: Hauke Mehrtens <[email protected]>

mvebu: Use common files folder

The files-6.6 and files-6.12 folder contain the same content, merge them
into the files folder for both kernel versions.

Link: https://github.com/openwrt/openwrt/pull/19423
Signed-off-by: Hauke Mehrtens <[email protected]>

kirkwood: add kmod-ata-ahci as default to ix4-200d

The ix4-200d is a 4 bay NAS supported since OpenWrt 23.05. Drive trays
controller. Due to a kernel-related issue with the SATA controller,
trays #1 and #2 were not recognized. This issue has been fixed in
kernel 6.10, so with the adoption of kernel 6.12 in the Kirkwood
target trays #1 and #2 can now be made accessible. This requires the
kmod-ata-ahci package, which this commit adds as default to the
makefile.

Signed-off-by: Sander van Deijck <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19435
Signed-off-by: Hauke Mehrtens <[email protected]>

mbedtls: Apply configuration in Configure instead of Prepare

This ensures mbedtls_config.h is correctly updated and the package rebuilt
accordingly when configuration options are changed after the source is first
prepared.

Signed-off-by: Karsten Sperling <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19358
Signed-off-by: Hauke Mehrtens <[email protected]>

kernel: modules: i2c: package i2c-mux-pinctrl

Create kernel module package for i2c-mux-pinctrl.

Signed-off-by: Daniel Golle <[email protected]>

targets: add PINCTRL_SUPPORT auto-feature

Not all targets have CONFIG_PINCTRL=y set in their kernel config.
Let's introduce a feature for that so kernel module packages which
select or depend on CONFIG_PINCTRL=y may depend on that, so we can
try to prevent leaking CONFIG_PINCTRL=y also into targets which do
not require it.

Signed-off-by: Daniel Golle <[email protected]>

uboot-mediatek: refresh patches

Refresh patch which has been made before the update to 2025.07 and
doesn't apply cleanly any more.

Signed-off-by: Daniel Golle <[email protected]>

uboot-mediatek: backport filogic clk driver fix

This patch can fix the crash when initializing the mmc device.

Signed-off-by: Shiji Yang <[email protected]>

kernel: bump 6.12 to 6.12.39

Changelog: https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.12.39

All patches auto-refreshed.

Signed-off-by: Leo Barsky <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19448
Signed-off-by: Hauke Mehrtens <[email protected]>

kernel: bump 6.6 to 6.6.99

Changelog: https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.6.99

All patches auto-refreshed.

Signed-off-by: Leo Barsky <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19448
Signed-off-by: Hauke Mehrtens <[email protected]>

arm-trusted-firmware-mediatek: update to new release 20250711

Update mediatek arm-trusted-firmware to the latest version.

Remove upstream patch:
  0004-mediatek-snfi-fix-return-code-when-reading.patch

Signed-off-by: Shiji Yang <[email protected]>

uboot-mediatek: sync with mtk-openwrt/u-boot 20250711

- Backport upstream Winbond W25N04KV Flash support.
- Backport upstream GigaDevice series Flash support.
- Backport pending Airoha AN8855 switch TPID value fix.
- Backport Mediatek UART baudrate accuracy compensation support.
- Pull mtk patchset from MTK SDK mtksoc-20250711 branch:
    Remove mt7622_rfb changes. The MTK SDK already dropped them.
    Replace Airoha ethernet PHY driver with new version.
    Split downstream snfi changes into independent patches.
    Add new Marvell CUX3410 PHY driver.
    Add new MediaTek built-in 2.5Gbps PHY driver.

Signed-off-by: Shiji Yang <[email protected]>

uboot-mediatek: drop build without OpenSSL fix

The build error has been fixed by the upstream commit:
03e598263e38 ("tools: fix build without LIBCRYPTO support")

Signed-off-by: Shiji Yang <[email protected]>

uboot-mediatek: support getting legacy uImage size

Most ramips target devices use the legacy uImage format. This
patch extends the imsz/imszb commands to support detecting the
image size of legacy uImage.

Signed-off-by: Shiji Yang <[email protected]>

mediatek: stage mt7622/mt7629 rfb u-boot images to bin folder

These u-boot images are no longer automatically copied to the
bindir since commit:
ed5000431940 ("uboot-mediatek: add support for Linksys E8450")

Signed-off-by: Shiji Yang <[email protected]>

uboot-mediatek: fix mt7622 rfb1 build

Add back the default LK image header to generate u-boot-mtk.bin. This
is the default behavior before u-boot v2022.01. If "u-boot-mtk.bin"
doesn't boot, please try "u-boot.bin" and report it.

Signed-off-by: Shiji Yang <[email protected]>

uboot-mediatek: update to v2025.07

Update to the latest version.

Suppressed patch:
  100-04-env-add-support-for-generic-MTD-device.patch[1]

[1] https://github.com/u-boot/u-boot/commit/03fb08d4aef8c342b583e148d1b5c4d289c5572f

Signed-off-by: Shiji Yang <[email protected]>

kernel: modules: input: add modules for PS/2 mouse support

Package kernel modules for PS/2 mouse support, mostly to allow
using touchpads and trackpoints built-into laptops (many of those
are connected using classic i8052-compatible PS/2 I/O).

Signed-off-by: Daniel Golle <[email protected]>

kernel: add missing dependency on kmod-drm to kmod-i2c-hid

Buildbot revealed that on the x86/geode target, on which CONFIG_DRM=y isn't set,
kmod-i2c-hid misses the dependency on kmod-drm. Fix that to fix the build for
x86/geode.

Fixes: c83e275c9b ("kernel: package modules for I2C HID devices")
Signed-off-by: Daniel Golle <[email protected]>

projectsmirrors: drop ftp.kddilabs.jp

ftp.kddilabs.jp has been shutdown and
now returns 404.

Signed-off-by: Yanase Yuki <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19407
Signed-off-by: Hauke Mehrtens <[email protected]>

apm821xx: enable kernel 6.12

It builds now.

Tested on Cisco Meraki MX60W.

Signed-off-by: Rosen Penev <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18798
Signed-off-by: Hauke Mehrtens <[email protected]>

apm821xx: refresh 6.12 kernel config

Some symbols changed.

Signed-off-by: Rosen Penev <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18798
Signed-off-by: Hauke Mehrtens <[email protected]>

apm821xx: refresh 6.12 patches

pci driver went to a different directory and thus manually edited.

All else done with kernel_oldconfig.

Signed-off-by: Rosen Penev <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18798
Signed-off-by: Hauke Mehrtens <[email protected]>

apm821xx: restore 6.6 kernel config/patches

Copy kernel config and patches from 6.12 to restore the default 6.6
kernel support files.

Signed-off-by: Rosen Penev <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18798
Signed-off-by: Hauke Mehrtens <[email protected]>

apm821xx: rename kernel config/patches to 6.12

This is a preparation for 6.12 kernel support. It can help us
track the files history by using the Git tool.

Signed-off-by: Rosen Penev <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/18798
Signed-off-by: Hauke Mehrtens <[email protected]>

realtek: avoid interrupt storm on mass packet receive

RTL83xx devices have two types of receive interrupts for each of its
8 rings. One for packet received and another for ring overflow. When
the switch is flooded with incoming packets the receive handler will
disable the packet receive notification but still keeps the overflow
notification enabled. While the receive path "slowly" processes the
received packets each new packet triggers the overflow IRQ again. The
device becomes unresponsive and eventually produces messages like:

[18441.709764] rcu: Stack dump where RCU GP kthread last ran:
[18441.727892] Sending NMI from CPU 1 to CPUs 0:
[18441.742300] NMI backtrace for cpu 0 skipped: idling at 0x8080e994
[18415.251700] rcu: INFO: rcu_sched detected stalls on CPUs/tasks:
[18415.271350] rcu:     0-...!: (0 ticks this GP) idle=d740/0/0x0 ...
[18415.303046] rcu:     (detected by 1, t=6004 jiffies, g=230925, ...
[18415.326095] Sending NMI from CPU 1 to CPUs 0:
[18415.340540] NMI backtrace for cpu 0

Fix this issue by always disabling receive and overflow interrupts at
the same time.

Test with hping3 --udp -p 5021 -d 1400 --flood 192.168.2.72

Before (3sec run):
[183260.324846] rtl838x-eth 1b00a300.ethernet eth0: RX buffer overrun: status 0x101, mask: 0x7ffeff
[183260.340524] rtl838x-eth 1b00a300.ethernet eth0: RX buffer overrun: status 0x1, mask: 0x7ffeff
[183260.345799] net_ratelimit: 489997 callbacks suppressed

After (3 sec run):
[  373.981479] rtl838x-eth 1b00a300.ethernet eth0: rx ring overrun: status 0x101, mask: 0x7fffff
[  374.031118] rtl838x-eth 1b00a300.ethernet eth0: rx ring overrun: status 0x101, mask: 0x7fffff
[  377.919996] net_ratelimit: 34 callbacks suppressed

Signed-off-by: Markus Stockhausen <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19365
Signed-off-by: Hauke Mehrtens <[email protected]>

realtek: simplify RTL8218B/RTL8214Fx detection

The current implementation has several issues:

- it uses the hacky phy_port* macros
- it uses SoC dependent raw pages
- it disables/enables SoC dependent polling

Get rid of these dependencies and access the mdio bus the normal way.

Signed-off-by: Markus Stockhausen <[email protected]>
Link: https://github.com/openwrt/openwrt/pull/19372
Signed-off-by: Hauke Mehrtens <[email protected]>

uboot-mediatek: add support for GL.iNet GL-MT2500

Add an OpenWrt-based U-Boot build for this device, allowing for more
flexibility and customization.

Expected behaviour
------------------
When plugging the device, keeping the reset button pressed will enter
TFTP recovery mode: the board will send requests for the initramfs file
(openwrt-mediatek-filogic-glinet_gl-mt2500-initramfs-kernel.bin) from
IP 192.168.1.1 to 192.168.1.10 TFTP server.
The bootloader will populate the environment with the unit serial number
as reported by the "sn" value in eMMC, the "sn_bak" value, the country
code and ddns. WAN and LAN MAC addresses are reported in the environment
as well.

Limitations
-----------
No web interface is available, and only the LAN port can be used for
system recovery.

Notes
-----
This port has been tested with the Maxlinear version only, and with a
board that does not exhibit eMMC communication problems. Even though eMMC
frequency has been lowered, some testing is probably needed and always
very welcome.
A -factory image has been introduced, and is only needed when using the
"Load Firmware via TFTP then write to eMMC." boot menu function.
The device has not been converted to use uImage.FIT including the rootfs
to keep compatibility with stock bootloader and firmware.

Installation
------------
From a Linux root shell:
0. Transfer the needed files to the board, placing them in /tmp.
1. Make sure your U-Boot environment is erased: all of my units came with
unpopulated environment, but I am not sure this is always the case.
Issue the command
cat /dev/zero >/dev/mmcblk0p2
(you will get a "no space left" error, which is reasonable and expected).
2. Unlock the eMMC boot area where BL2 is located:
echo 0 >/sys/block/mmcblk0boot0/force_ro
3. Write the new BL2 code:
cat openwrt-mediatek-filogic-glinet_gl-mt2500-emmc-preloader.bin >/dev/mmcblk0boot0
4. Write new BL31+U-Boot image:
cat openwrt-mediatek-filogic-glinet_gl-mt2500-emmc-bl31-uboot.fip >/dev/mmcblk0p4
5. Reboot.

Signed-off-by: Enrico Mioso <[email protected]>

uboot-mediatek: GatoNetworks GDSP support fixes and updates

This commit updates the OpenWrt-based U-Boot for this device, allowing
to boot a FIT firmware image stored in an SD card.
The environment has been updated accordingly to allow for managing images
on the two different supported firmware boot media configurations
(NOR or SDMMC).
Erasing the U-Boot environment when upgrading is recommended, so that
the correct DTS overlay is applied.

New behaviour
-------------

The default environment "bootconf" variable is set to
config-1#mt7981b-gatonetworks-gdsp-gps
rendering the GPS module accessible.
You may set it to something like
config-1#mt7981b-gatonetworks-gdsp-sd
to be able to use the MMC controller.
To boot from a firmware image stored in an SD card, you may set the
bootmedia environment variable:
# fw_setenv bootmedia sd
and reboot.
Should the bootmedia variable be absent or it's content not recognized
as valid (e.g., does not contain "nor" or "sd"), NOR boot will be
attempted.

Signed-off-by: Enrico Mioso <[email protected]>

mediatek: filogic: add SD card support to GatoNetworks GDSP

The device is equipped with a GPS module, reporting data via /dev/ttyS1.
A TF card reader is also present. Only one of those components can be
used at once, since they share some PINs.
This commit adds two devicetree overlays to allow for the user to select
the desired configuration. Another overlay configuration to allow booting
from SD card is provided.

Signed-off-by: Enrico Mioso <[email protected]>

lantiq: enable 6.12 testing kernel

The 6.12 testing kernel for lantiq target is ready now.

Signed-off-by: Shiji Yang <[email protected]>

lantiq: refresh 6.12 kernel config files

All kernel symbols are automatically refreshed by
`make kernel_oldconfig CONFIG_TARGET=target` and
`make kernel_oldconfig CONFIG_TARGET=subtarget`.

Signed-off-by: Shiji Yang <[email protected]>

lantiq: add the missing function pointer ltq_pci_plat_dev_init

Partial revert upstream commit 075dd2eef025 ("MIPS: lantiq: Remove
unused function pointer variables") to fix the build error:

mips-openwrt-linux-musl-ld: arch/mips/lantiq/xway/ath5k_eep.o: in function `of_ath5k_eeprom_probe':
arch/mips/lantiq/xway/ath5k_eep.c:(.init.text+0x31c): undefined reference to `ltq_pci_plat_dev_init'
mips-openwrt-linux-musl-ld: arch/mips/lantiq/xway/ath5k_eep.c:(.init.text+0x338): undefined reference to `ltq_pci_plat_dev_init'

Signed-off-by: Shiji Yang <[email protected]>
lantiq: fix vmmc driver build warnings

lantiq: fix i2c driver compatibility issue on 6.12 kernel

* Drop obsolete flag I2C_CLASS_SPD. [1]
* Change .remove callback return type to NULL.
* Use strscpy instead of strlcpy.

This patch fixes the following build errors:

drivers/i2c/busses/i2c-lantiq.c:599:41: error: 'I2C_CLASS_SPD' undeclared (first use in this function); did you mean 'I2C_CLASS_HWMON'?
  599 |         adap->class = I2C_CLASS_HWMON | I2C_CLASS_SPD;
      |                                         ^~~~~~~~~~~~~
      |                                         I2C_CLASS_HWMON
drivers/i2c/busses/i2c-lantiq.c:600:9: error: implicit declaration of function 'strlcpy'; did you mean 'strncpy'? [-Wimplicit-function-declaration]
  600 |         strlcpy(adap->name, DRV_NAME "-adapter", sizeof(adap->name));
      |         ^~~~~~~
      |         strncpy
drivers/i2c/busses/i2c-lantiq.c:729:19: error: initialization of 'void (*)(struct platform_device *)' from incompatible pointer type 'int (*)(struct platform_device *)' [-Wincompatible-pointer-types]
  729 |         .remove = ltq_i2c_remove,
      |

[1] https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=linux-6.12.y&id=e61bcf42d290e73025bab38e0e55a5586c2d8ad5
Signed-off-by: Shiji Yang <[email protected]>

lantiq: lantiq-flash: platform driver .remove_new conversion

The .remove has been converted to .remove_new in the 6.12 kernel.
We have to change remove function return type from int to void
to match the upstream changes.

Link: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=linux-6.12.y&id=5882bf98089f3a9ecbee5a57ae0aec6f54797855
Signed-off-by: Shiji Yang <[email protected]>