ramips: add support for ZyXEL PoE12-3PD

Hardware
========
MediaTek MT7621
256MB DDR3 (Winbond W632GG6MB-11)
128MB NAND (Winbond W29N01HV)
1x Uplink (PoE IN)
3x LAN (PoE Out)
IC+ IP804AR PSE Controller

UART
====

UART is located in the top left corner (unpopulated). VCC is
silkscreened.

GND - TX - RX - VCC

Don't connect VCC. Baudrate is 115200 8N1.

Installation
============

Signed-off-by: David Bauer <mail@david-bauer.net>

ramips: prepare zyxel-nwa-fit to support more models

Pass the string of compatible models to the "zyxel-nwa-fit" image
generation script as an argument.

This is required to support new models using the same script.

Signed-off-by: David Bauer <mail@david-bauer.net>

mpc85xx: fix function name

The function name is misspelled and does not exist. Fix the function
name so the correct function is called.

Signed-off-by: David Bauer <mail@david-bauer.net>

zyxel-bootconfig: fix indentation

Signed-off-by: David Bauer <mail@david-bauer.net>

bcm27xx-gpu-fw: update to v1.20250430

Full changelog: https://github.com/raspberrypi/firmware/compare/1.20250305...1.20250430

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>

ipq806x: config-6.6: add new option

6.6.93 introduced new config option[1]

1. https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.6.93&id=fba6e5d44fb8df6bdadb1e90d2ddca715462248f

Signed-off-by: John Audia <therealgraysky@proton.me>
Link: https://github.com/openwrt/openwrt/pull/19026
Signed-off-by: Nick Hainke <vincent@systemli.org>

ipq40xx: config-6.6: add new option

6.6.93 introduced new config option[1]

1. https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.6.93&id=fba6e5d44fb8df6bdadb1e90d2ddca715462248f

Signed-off-by: John Audia <therealgraysky@proton.me>
Link: https://github.com/openwrt/openwrt/pull/19026
Signed-off-by: Nick Hainke <vincent@systemli.org>

kernel: bump 6.6 to 6.6.93

Changelog: https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.6.93

Removed upstreamed:
backport-6.6/780-51-v6.15-r8169-don-t-scan-PHY-addresses-0.patch[1]

All other patches automatically rebased.

1. https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.6.93&id=961ee132104b09e054308e53d8ab4a2386a21904

Build system: x86/64
Build-tested: flogic/xiaomi_redmi-router-ax6000-ubootmod, ramips/tplink_archer-a6-v3, x86/64
Run-tested: flogic/xiaomi_redmi-router-ax6000-ubootmod, ramips/tplink_archer-a6-v3, x86/64

Signed-off-by: John Audia <therealgraysky@proton.me>
Link: https://github.com/openwrt/openwrt/pull/19026
Signed-off-by: Nick Hainke <vincent@systemli.org>

ramips: mt7621: fix file permissions in source tree

They don't need +x permission.

Fixes: 502916468e28 ("ramips: add support for ASUS 4G-AX56")
Signed-off-by: Zheng Zhang <everything411@qq.com>
Link: https://github.com/openwrt/openwrt/pull/19034
Signed-off-by: Nick Hainke <vincent@systemli.org>

starfive: fix 6.12 kernel patch conflict

The recently committed starfive 6.12 kernel support patchset[1] has
conflict with the 6.12.32 kernel update[2]. Remove upstreamed patch
to fix the issue:

  0031-phy-starfive-jh7110-usb-Fix-usb-2.0-host-detection-f.patch [3]

[1] https://github.com/openwrt/openwrt/pull/18997
[2] https://github.com/openwrt/openwrt/pull/19027
[3] https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.12.32&id=2c09a5cbc0100e88b8872b11f4da01a656c97fbc

Fixes: ebfd69a3e373 ("kernel: bump 6.12 to 6.12.32")
Signed-off-by: Shiji Yang <yangshiji66@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/19039
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>

mediatek: add support for TP-Link Archer AX80v1(US/RU/CA)

This commit adds support for TP-Link Archer AX80v1(US/RU/CA).

Device specification
SoC Type:   MediaTek MT7986AV, Cortex-A53, 64-bit
RAM:        ESMT M15T4G16256 (512MB)
Flash:      ESMT F50L1G41LB (128 MB)
Ethernet:   MediaTek MT7531AE + 2.5GbE MaxLinear GPY211C0VC (SLNW8)
Ethernet:   1x2.5Gbe (WAN/LAN 2.5Gbps), 4xGbE (WAN/LAN 1Gbps, LAN1, LAN2, LAN3)
WLAN 2g:    MediaTek MT7976GN
WLAN 5g:    MediaTek MT7976AN
LEDs:       1 red,1 green,1 blue status LEDs,

Buttons:    4 (Reset,ledswitch,wps,wlan),
USB ports:  1 (USB 3.0)
Power:      12 VDC, 3,3 A
Connector:  Barrel
Bootloader: Main U-Boot - U-Boot 2022.01-rc4. Additionally, both UBI
            slots contain "seconduboot" (also U-Boot 2022.01-rc4)
Serial console (UART)
                            V
+-------+-------+-------+-------+
| +3.3V |  GND  |  TX   |  RX   |
+---+---+-------+-------+-------+
    |
    +--- Don't connect

Installation (UART)
Place OpenWrt initramfs image on tftp server with IP 192.168.1.2

Attach UART, switch on the router and interrupt the boot process by pressing 'Ctrl-C'

Load and run OpenWrt initramfs image:

      tftpboot initramfs-kernel.bin
      bootm

!!Attention!! is very important! After entering OpenWrt, please set / update the environment variables:

fw_setenv bootargs "ubi.mtd=ubi0 console=ttyS0,115200n1 loglevel=8 earlycon=uart8250,mmio32,0x11002000 init=/etc/preinit"
fw_setenv mtdids "spi-nand0=spi-nand0"
fw_setenv mtdparts "spi-nand0:2M(boot),1M(u-boot-env),50M(ubi0),50M(ubi1),8M(userconfig),4M(tp_data),8M(mali_data)"
fw_setenv tp_boot_idx 0

Run 'sysupgrade -n' with the sysupgrade OpenWrt image from console or Luci WebUI.

Recovery

Press Reset button and power on the router
Navigate to U-Boot recovery web server (192.168.1.1) and upload the OEM firmware

Stock layout

0x000000000000-0x000000200000 : "boot"
0x000000200000-0x000000300000 : "u-boot-env"
0x000000300000-0x000003500000 : "ubi0"
0x000003500000-0x000006700000 : "ubi1"
0x000006700000-0x000006f00000 : "userconfig"
0x000006f00000-0x000007300000 : "tp_data"
0x000007300000-0x000007B00000 : "mali_data"

ubi0/ubi1 format
U-Boot at boot checks that all volumes are in place:

+-------------------------------+
| Volume Name: uboot   Vol ID: 0|
| Volume Name: kernel  Vol ID: 1|
| Volume Name: rootfs  Vol ID: 2|
+-------------------------------+
MAC addresses
+---------+-------------------+-----------+
| label   | 00:eb:xx:xx:xx:be | label     |
| LAN     | 00:eb:xx:xx:xx:be | label     |
| WAN     | 00:eb:xx:xx:xx:bf | label+1   |
| WLAN 2g | 00:eb:xx:xx:xx:be | label     |
| WLAN 5g | 00:eb:xx:xx:xx:bd | label-1   |
+---------+-------------------+-----------+
label MAC address was found in UBI partition "tp_data", file "default-mac".
OEM wireless eeprom is also there (file "MT7986_EEPROM.bin").

Signed-off-by: Sergey Shlukov <ichizakurain@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17753
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel: add kmod-leds-lp5523

Add kernel module for LP5523/LP55231 LED driver.
The kmod-leds-lp5523 depends on kmod-leds-lp55xx-common.

Signed-off-by: Sergey Shlukov <ichizakurain@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17753
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

build: image: support specifying mksquashfs4 parallel threads

By enabling multi-threading file reading, the squashfs rootfs build
process can be greatly accelerated. We use 4 as the default thread
number. This is the default value for squashfs4 tool 4.7 if we don't
append "-block-readers" or "-small-readers" options.

For more test results:
https://github.com/plougher/squashfs-tools/blob/4.7/Documentation/4.7/README

Signed-off-by: Shiji Yang <yangshiji66@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/19019
Signed-off-by: Nick Hainke <vincent@systemli.org>

tools/squashfs4: update to 4.7.0

This new version introduced parallel file reading, which will
greatly improve the IO performance. OpenWrt only uses GZIP
and XZ compression, hence the LZ4, LZO and ZSTD compression
algorithms were explicitly disabled.

Upstreamed patches:
  001-xz_wrapper-support-multiple-lzma-configuration-optio.patch[1]
  002-xz_wrapper-make-new-OpenWrt-extended-options-non-def.patch[2]

Release Notes:
https://github.com/plougher/squashfs-tools/releases/tag/4.7

[1] https://github.com/plougher/squashfs-tools/commit/dcb976fe4ee40e4bac8ae0dcc836629c625a6fd4
[2] https://github.com/plougher/squashfs-tools/commit/5fb9fdfb8757fc9afb6318a3dcf9dce0a97de352

Signed-off-by: Shiji Yang <yangshiji66@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/19019
Signed-off-by: Nick Hainke <vincent@systemli.org>

mediatek: Create common DTSI for WR3000H and WR3000S

This change moves common elements of the WR3000H and the WR3000S to mt7981b-cudy-wr3000-nand.dtsi.
This will simplify adding of new similar devices, for exapmle WR3000E.

Signed-off-by: Roland Reinl <reinlroland+github@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18619
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel: bump 6.12 to 6.12.32

Changelog: https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.12.32

All patches are automatically refreshed.

Signed-off-by: Shiji Yang <yangshiji66@outlook.com>
Tested-by: John Audia <therealgraysky@proton.me>
Link: https://github.com/openwrt/openwrt/pull/19027
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

starfive: add 6.12 as testing kernel

Runtime-tested on:
 - Visionfive v2-1.3b (JH7110)
 - Visionfive v1 (JH7100)

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>

starfive: jh7110: update bootscript to use addresses in bootloader

Rather than hardcoding the kernel/fdt addresses in the boot.scr script,
use the addresses provided by the bootloader.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>

starfive: 6.12: refresh patches and drop upstreamed ones

 - refresh, rebase and reorder patches
 - JH7110 media drivers have been dropped for now
 - JH7110 E24 and mailbox drivers were added
 - JH7100 DMA- and errata-patches have been dropped as they were
   upstreamed

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>

kernel/starfive: Restore kernel files for v6.6

This is an automatically generated commit which aids following Kernel patch
history, as git will see the move and copy as a rename thus defeating the
purpose.

For the original discussion see:
https://lists.openwrt.org/pipermail/openwrt-devel/2023-October/041673.html

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>

kernel/starfive: Create kernel files for v6.12 (from v6.6)

This is an automatically generated commit.

When doing `git bisect`, consider `git bisect --skip`.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>

mac80211: ath9k: avoid a double reset

The original OF code effectively does a reset at ahb.c but then again in
hw.c. For AR9330, it's already done in the driver and with the others,
there are patches in here that do the same. hw.c looks like the proper
place to handle this.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/19031
Signed-off-by: Robert Marko <robimarko@gmail.com>

mac80211: ath9k: fix refclock typo

Fixes ath9k on ar934x platforms.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/19031
Signed-off-by: Robert Marko <robimarko@gmail.com>

archs38: drop target

archs38 has been on life support for the last couple of releases,eventually
leading to marking it as source-only in 2023.

It has been basically only touched to do a kernel bump so that we can make
the new OpenWrt release.

Link: https://github.com/openwrt/openwrt/pull/19001
Signed-off-by: Robert Marko <robimarko@gmail.com>

armsr: enable 6.12 as a testing kernel version

Allow the armsr target to be built with a 6.12 kernel
when CONFIG_TESTING_KERNEL is set.

Signed-off-by: Mathew McBride <matt@traverse.com.au>
Link: https://github.com/openwrt/openwrt/pull/18849
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

config: remove kernel version restriction for ARM_PMUV3

The kernel CONFIG_ARM_PMUV3 option was not being presented
for 6.12 kernels with armsr/armv7.

As the minimum kernel version is now 6.6, there is no need
to limit this to specific kernel versions anymore.

This caused a kernel compile stoppage when attempting to
compile the 'perf' tool for armsr/armv7.

Signed-off-by: Mathew McBride <matt@traverse.com.au>
Link: https://github.com/openwrt/openwrt/pull/18849
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

armsr: 6.12: add new KConfigs and refresh

This includes both new additions to 6.12
and an attempted "refresh" of the config
to remove duplicates from target/generic/config-6.12.

(The OpenWrt kernel_makeoldconfig does not
work well with the armv8 subtarget for reasons
I am yet to determine, so that file has been
pruned manually)

Most new options are in armv8, where the
KConfig relates to a platform that will likely
be armsr compatible (like the i.MX91/93/95),
it has been enabled.

Signed-off-by: Mathew McBride <matt@traverse.com.au>
Link: https://github.com/openwrt/openwrt/pull/18849
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

armsr: patches: refresh printk console patch for 6.12

This carries the text console behaviour fix as described in
[1] over to the 6.12 kernel.

Signed-off-by: Mathew McBride <matt@traverse.com.au>
[1] https://github.com/openwrt/openwrt/pull/17012#issuecomment-2591751115
Link: https://github.com/openwrt/openwrt/pull/18849
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel/armsr: Restore kernel files for v6.6

This is an automatically generated commit which aids following Kernel patch
history, as git will see the move and copy as a rename thus defeating the
purpose.

For the original discussion see:
https://lists.openwrt.org/pipermail/openwrt-devel/2023-October/041673.html

Signed-off-by: Mathew McBride <matt@traverse.com.au>
Link: https://github.com/openwrt/openwrt/pull/18849
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel/armsr: Create kernel files for v6.12 (from v6.6)

This is an automatically generated commit.

When doing `git bisect`, consider `git bisect --skip`.

Signed-off-by: Mathew McBride <matt@traverse.com.au>
Link: https://github.com/openwrt/openwrt/pull/18849
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

tools/7z: update to 24.09

There are no patches that need to be refreshed.

Signed-off-by: Shiji Yang <yangshiji66@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/19020
Signed-off-by: Nick Hainke <vincent@systemli.org>

wireguard-tools: update to v1.0.20250521

Update to latest version.

Release notes: https://lists.zx2c4.com/pipermail/wireguard/2025-May/008789.html

% wg -v
wireguard-tools v1.0.20250521 - https://git.zx2c4.com/wireguard-tools/

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

Signed-off-by: John Audia <therealgraysky@proton.me>
Link: https://github.com/openwrt/openwrt/pull/19023
Signed-off-by: Nick Hainke <vincent@systemli.org>

tools/ccache: update to 4.11.3

Update ccache to the latest version. All patches are automatically
refreshed. Changelogs:

https://github.com/ccache/ccache/blob/v4.11.3/doc/NEWS.adoc

Signed-off-by: Shiji Yang <yangshiji66@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/19025
Signed-off-by: Nick Hainke <vincent@systemli.org>

kernel: mtdsplit_uimage: return 0 if not fatal

Introduced with Linux 6.7, in commit:
5c2f7727d437 ("mtd: mtdpart: check for subpartitions parsing result"),
when a parser returns an error, this will be passed up, and
consequently, all parent mtd partitions get torn down.

Adjust the mtdsplit_uimage driver to only return an error if there is a
critical problem in reading from the mtd device or allocating memory.
Otherwise return 0 to indicate that no partitions were found.
Also add logging to indicate what went wrong.

E.g. on Realtek devices that are booted for the first time through
initramfs with OpenWrt never installed before boot log will show

[    0.932985] Creating 8 MTD partitions on "spi0.0":
[    0.938412] 0x000000000000-0x000000080000 : "u-boot"
[    0.990151] 0x000000080000-0x0000000c0000 : "u-boot-env"
[    0.999907] 0x0000000c0000-0x000000100000 : "board-name"
[    1.019971] 0x000000100000-0x000000e80000 : "firmware"
[    1.051582] mtdsplit_uimage: no uImage found in "firmware"
[    1.069365] 0x000000e80000-0x000001000000 : "kernel2"
[    1.078959] 0x000001000000-0x000001040000 : "sysinfo"
[    1.099747] 0x000001040000-0x000001c40000 : "rootfs2"
[    1.119865] 0x000001c40000-0x000002000000 : "jffs2"

Similar issue was fixed before with commit ade045084bd3f8696
("kernel: mtdsplit_minor: return 0 if not fatal")

Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
Link: https://github.com/openwrt/openwrt/pull/19016
Signed-off-by: Robert Marko <robimarko@gmail.com>

uboot-mediatek: update patch for Cudy TR3000-v1 and SNR SNR-CPE-AX2

Update the reference to gpio props.

Fixes: 2a32d215ba47 ("uboot-mediatek: bump to v2025.04")
Signed-off-by: Tianling Shen <cnsztl@immortalwrt.org>
Link: https://github.com/openwrt/openwrt/pull/18988
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

mediatek: filogic: fix for new GL.iNet GL-MT2500/GL-MT2500A hardware revision

GL.iNet shipped a hardware change of the WAN PHY going from the MaxLinear
GPY211C to the Airoha EN8811H.

Signed-off-by: Matthew Bilker <me@mbilker.us>
Link: https://github.com/openwrt/openwrt/pull/18799
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

gemini: Move back to using statically compiled PHYs

Trying to use modules for the PHY:s does not work:

The ethernet driver does not like to probe if it can't find
the PHY.

The ethernet driver really likes to be compiled-in. It will
not probe otherwise. Some hardware issue.

Revert things to how they always worked until I maybe solve
this.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>

mwlwifi: add pending patch to fix compilation with kernel 6.12

Building against recent kernel versions (noticed with 6.12) and -Werror
can fail because a macro MAX(a,b) is already defined in minmax.h or
kernel.h before 5.10.

In file included from ../mwlwifi-2025.02.06~db97edf2/hif/fwcmd.h:23,
                 from ../mwlwifi-2025.02.06~db97edf2/core.c:25:
../mwlwifi-2025.02.06~db97edf2/hif/hostcmd.h:1124:9: error: "MAX" redefined [-Werror]
 1124 | #define MAX(a, b) (((a) > (b)) ? (a) : (b))
      |         ^~~
In file included from usr/include/mac80211-backport/linux/minmax.h:4,
                 from ./include/linux/kernel.h:28,
                 from usr/include/mac80211-backport/linux/kernel.h:3,
                 from ./include/linux/skbuff.h:13,
                 from usr/include/mac80211-backport/linux/skbuff.h:3,
                 from ./include/linux/if_ether.h:19,
                 from usr/include/mac80211-backport/linux/if_ether.h:3,
                 from ./include/linux/etherdevice.h:20,
                 from usr/include/mac80211-backport/linux/etherdevice.h:3,
                 from ../mwlwifi-2025.02.06~db97edf2/core.c:18:
./include/linux/minmax.h:330:9: note: this is the location of the previous definition
  330 | #define MAX(a,b) __cmp(max,a,b)
      |         ^~~

Add a pending upstream patch which replaces the MAX(a,b) macro to avoid
conflicts and allow compilation with 6.12

Signed-off-by: Stefan Kalscheuer <stefan@stklcode.de>
Link: https://github.com/openwrt/openwrt/pull/18980
Signed-off-by: Robert Marko <robimarko@gmail.com>

odhcp6c: Conditionally request S46 OROs

Request Softwire46 (S46) [RFC 7598] options when the map and/or ds-lite
packages are installed. This is required as the behaviour of odhcp6c has
changed to not include these OROs by default.

See openwrt/odhcp6c#89

Signed-off-by: Richard Patterson <richard@helix.net.nz>
Signed-off-by: Shengyu Qu <wiagn233@outlook.com>

odhcp6c: update to Git HEAD (2025-02-06)

ffbb2d559af4 odhcp6c: Do not request S46 OROs by default
8aa8b706727a Use GNUInstallDirs macros

Signed-off-by: Shengyu Qu <wiagn233@outlook.com>

realtek: extend simple thermal driver to support rtl839x

* Extends SoC thermal sensor on rtl839x
* Tested on HP JG928A

Signed-off-by: Stephen Howell <howels@allthatwemight.be>
Link: https://github.com/openwrt/openwrt/pull/18825
Signed-off-by: Robert Marko <robimarko@gmail.com>

ramips: mt7621: use regulators for USB GPIO

Regulators as implemented by the XHCI driver only accept one GPIO.
However, we can abuse the fact that the XHCI driver accepts two
regulators, one for 5V and the other for 3.3V, for USB 2 and 3 GPIOs.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16967
Signed-off-by: Robert Marko <robimarko@gmail.com>

bmips: huawei,hg556a-b: configure LAN LEDs

Now that the external switch is working properly we can configure the LAN LEDs.

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>

ixp4xx: Switch to kernel v6.12

This can be done with so little effort, all but two patches are now
upstream. No need to keep support for v6.6, everything just works
the same or better with v6.12.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>

qualcommax: ipq50xx: fix uniphy soft reset issue

The resets in the GCC of the uniphy found in the IPQ5018 SoC are
incorrect which broke the ability to shift between 1G and 2.5G link
speeds. So let's correct the resets based on below two downstream
commits.

In a seperate and prequisite PR to the QCA-SSDK repo, logic has been
implemented to select the right reset based on the link setup so fixed
link scenarios don't break.

Signed-off-by: George Moussalem <george.moussalem@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/18774
Signed-off-by: Robert Marko <robimarko@gmail.com>

qca-ssdk: update to Git HEAD (2025-05-30)

446db12b1fd3 MP: fix uniphy reset in phy to phy link scenario

Signed-off-by: George Moussalem <george.moussalem@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/18774
Signed-off-by: Robert Marko <robimarko@gmail.com>

qualcommax: ipq50xx: prepare for UNIPHY reset fix

Ahead of the actual fix in both the GCC and QCA-SSDK, add the required
AHB reset so it can be picked up by updated QCA-SSDK. This is needed
as the SSDK needs to use different resets depending on the link
architecture. If it's a fixed link, AHB needs to be reset. In a phy to
phy link setup (such as QCA8081), SYS, RX, and TX need to be reset using
one reset with a bitmask in the GCC (GCC_UNIPHY_SOFT_RESET).

Signed-off-by: George Moussalem <george.moussalem@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/18774
Signed-off-by: Robert Marko <robimarko@gmail.com>

qualcommax: ipq50xx: remove named clock references for GE PHY

With completely carving out GE PHY out of the QCA-SSDK, the named clock
references to the GE PHY RX and TX clocks are no longer needed.
So, let's revert to using the DT indices as per the upstream GCC driver.

Signed-off-by: George Moussalem <george.moussalem@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/18774
Signed-off-by: Robert Marko <robimarko@gmail.com>

qca-ssdk: update to Git HEAD (2025-05-30)

3423fdacac63 ssdk_clk: Remove MP_PHY clocks resets and init logic

Signed-off-by: George Moussalem <george.moussalem@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/18774
Signed-off-by: Robert Marko <robimarko@gmail.com>

qualcommax: ipq50xx: update patches for IPQ5018 GE PHY

Use latest patches sent upstream for review for IPQ5018 GE PHY support:
- Move enablement of the LDO controller to the mdio-ipq4019 driver away
  from the CMN PLL driver
- Remove the different patches to add CDT, MSE, AZ, and DAC support they
  are all contained in the upstreamed driver.

Accordingly, also set the right property in the DTS for Linksys SPNMX56
to set the right DAC values to accommodate for the short cable length.

Signed-off-by: George Moussalem <george.moussalem@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/18774
Signed-off-by: Robert Marko <robimarko@gmail.com>

gemini: switch default NTFS driver package to kmod-fs-ntfs3

kmod-fs-ntfs is not available on the 6.12 kernel.

Signed-off-by: Shiji Yang <yangshiji66@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/18954
Signed-off-by: Robert Marko <robimarko@gmail.com>

qualcommbe: drop kernel v6.6 support

As qualcommbe is now supported by the v6.12 kernel, there is no point
in v6.6 as well. Drop v6.6 support.

Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18982
Signed-off-by: Robert Marko <robimarko@gmail.com>

qualcommbe: use kernel 6.12 by default

The qualcommbe target was introduced after openwrt-24.10. The v6.12
kernel is now available, and is likely to be used by the next openwrt
release.While the v6.6 kernel served as an interim development vehicle,
it is no longer useful for the qualcommbe target

The v6.12 patches contain more recent submissions of pending ipq95xx
drivers. I expect that it will be much easier to update v6.12 patches
with new submissions. For ease of maintenance, it makes sense to use
a single kernel for qualcommbe.

For these reasons, enable v6.12 by default.

Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18982
Signed-off-by: Robert Marko <robimarko@gmail.com>

bmips: huawei,hg556a-b: enable external switch

The external switch of the Huawei HG556a is a BCM5325E connected by MDIO.

All the DSA brcm legacy FCS tag and b53 patches have been submitted upstream
and will be backported when accepted.

There are still some sporadic FDB errors, but at least the switch is working
and stable on my device:
bcm53xx fffe4800.ethernet-mii:1e: port 0 failed to add 72:31:59:xx:xx:xx vid 1 to fdb: -28
bcm53xx fffe4800.ethernet-mii:1e: port 0 failed to add 5c:4c:a9:xx:xx:xx vid 0 to fdb: -28
bcm53xx fffe4800.ethernet-mii:1e: port 0 failed to add 5c:4c:a9:xx:xx:xx vid 1 to fdb: -28
bcm53xx fffe4800.ethernet-mii:1e: port 0 failed to delete 72:31:59:xx:xx:xx vid 1 from fdb: -2

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>

pistachio: drop 6.6 support

Drop configs and patches for Linux 6.6.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>

pistachio: use kernel 6.12 by default

Switch to Linux kernel version 6.12.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>

realtek: support EnGenius EWS2910P v3

Add support for V3 of the Engenius EWS2910P PoE switch. Like its v1
brother, This is an RTL8380 based switch with two SFP slots, and PoE
802.3af one every RJ-45 port.

Unlike its older brother, the max budget is 55W instead of 61.6 W.
Investigation into the communication protocol with the PoE controller
is ongoing, though it appears the vendor firmware configures the PSE
with a per-port budget of 30.0W.

Specifications:
---------------
* SoC:       Realtek RTL8380M
* Flash:     32 MiB SPI flash Macronix MX25L25635E
* RAM:       256 MiB (As reported by bootloader)
* Ethernet:  8x 10/100/1000 Mbps with PoE
             2x SFP slots
* Buttons:   1 "Reset" button on front panel
             1 "LED mode: button on front panel
             1 "On/Off" Toggle switch on the back
* Power:     48V-54V DC barrel jack
* UART:      1 serial header (JP1) with populated 2.54mm pitch header
             Labeled GRTV for ground, rx, tx, and 3.3V respectively
* PoE:       1 STM ST32... microcontroller (U15)
             1 RTL8238B PSE controller
Works:
------
  - (8) RJ-45 ethernet ports
  - Switch functions
  - LEDs and buttons

Not yet enabled:
----------------
  - Power-over-Ethernet (requires realtek-poe support for RTL8232B)

Install via web interface:
-------------------------

The factory firmware will accept and flash the initramfs image. It is
recommended to flash to "Partition 0". Flashing to "Partition 1" is
not supported at this point.

The factory web GUI will show the following warning:

 " Warning: The firmware version is v0.00.00-c0.0.00
     The firmware image you are uploading is older than the current
     firmware of the switch. The device will reset back to default
     settings. Are you sure you want to proceed?"

This is expected when flashing OpenWrt. After the initramfs image
boots, flash the -sysupgrade using either the commandline or LuCI.

Install via serial console/tftp:
--------------------------------

See commit 2cfaab4549f5 ("realtek: add support for EnGenius EWS2910P").

Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/15217
Signed-off-by: Robert Marko <robimarko@gmail.com>

realtek: engenius_ews2910p: support multiple hardware versions

When the Engenius EWS-2910P was added, only v1 was known. Move the
common parts to a dtsi, and split up the support to acccount for the
hardware version.

On v3, for example, the root partition uses a different uImage magic.

Add a "engenius,ews2910p-v1" compatible, while leaving the legacy
"engenius,ews2910p" to also mean v1.

Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/15217
Signed-off-by: Robert Marko <robimarko@gmail.com>

uboot-mediatek: bump to v2025.04

The following upstreamed / superseded patches were dropped:

 060-01-clk-mediatek-mt7629-fix-parent-clock-of-some-top-clo.patch
 060-02-arm-dts-mt7629-fix-sgmii-clock-selection-for-etherne.patch
 060-03-net-mediatek-use-correct-register-field-for-SGMII-sp.patch
 060-04-net-mediatek-correct-register-name-of-ethsys-syscfg1.patch
 060-05-net-mediatek-fix-sgmii-selection-for-mt7622.patch
 060-06-net-mediatek-fix-gmac2-usability-for-mt7629.patch
 060-07-net-mediatek-add-support-for-10GBASE-R.patch
 060-08-net-mediatek-make-sgmii-usxgmii-optional.patch
 060-09-net-mediatek-don-t-enable-GDMA-cpu-bridge-unconditio.patch
 060-10-net-mediatek-fix-usability-with-wget-command.patch
 061-01-net-mediatek-split-ethernet-switch-code-from-mtk_eth.patch
 061-02-net-mediatek-add-support-for-MediaTek-MT7987-SoC.patch
 061-03-net-mediatek-add-support-for-Airoha-AN8855-ethernet-.patch
 070-01-board-mediatek-mt7622-remove-board_late_init.patch
 070-02-clk-mediatek-fix-uninitialized-fields-issue-in-INFRA.patch
 070-03-configs-mt7629-move-image-load-address-to-0x42000000.patch
 070-04-configs-mt7988-move-image-load-address-to-0x44000000.patch
 070-05-spi-mtk_spim-add-support-to-use-DT-live-tree.patch
 070-06-spi-mtk_spim-check-slave-device-mode-in-spi-mem-s-su.patch
 070-07-arm-dts-mediatek-add-quad-mode-capabilities-for-SPI-.patch
 070-08-pwm-mediatek-add-pwm3-support-for-mt7981.patch
 070-09-pci-mediatek-add-support-for-multiple-ports-in-media.patch
 070-10-arm-dts-mediatek-add-pcie-support-for-mt7988.patch
 070-11-arm-dts-medaitek-fix-internal-switch-link-speed-of-m.patch
 070-12-arm-dts-mediatek-add-support-for-all-three-GMACs-for.patch
 070-13-arm-dts-medaitek-add-flash-interface-driving-setting.patch
 070-14-arm-dts-mediatek-update-mt7981-mmc-node.patch
 070-15-MAINTAINERS-update-file-list-for-MediaTek-ARM-platfo.patch
 071-01-pinctrl-mediatek-update-mt7981-pinctrl-driver-based-.patch
 100-18-board-mt7629-add-support-for-booting-from-SPI-NAND.patch
 290-mt7981-add-USB-nodes.patch

Refreshed all the patches needing it.

Run-tested: GatoNetworks GDSP, Arcadyan Mozart (ynezz),
            Zbt WG3526 (dangowrt)

Signed-off-by: Enrico Mioso <mrkiko.rs@gmail.com>
Signed-off-by: Petr Štetiar <ynezz@true.cz> [patch refresh, Mozart testing]
Signed-off-by: Daniel Golle <daniel@makrotopia.org> [rebased, tested on MT7621]

sysupgrade: create 10_disable_services w/ fixed modified time

Every time "sysupgrade -b -" runs it would generate a new
(synthetic) "/etc/uci-defaults/10_disable_services" file with
the current time as the modified time.  This unfortunately
creates a non-deterministic tarball, so if you run a cron job
to save your state, you don't have a trivial way of seeing if
it changed or not without unpacking the archive, deleting this
file, and comparing the entire directory tree to the previous
backup.

Fixes: #16145
Fixes: 0ad062a21b ("base-files: sysupgrade: include uci-defaults script disabling services #2")
Signed-off-by: Philip Prindeville <philipp@redfish-solutions.com>

kernel: bump 6.12 to 6.12.31

Changelog: https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.12.31

Remove upstreamed patches:
  backport-6.12/780-27-v6.15-r8169-don-t-scan-PHY-addresses-0.patch [1]
  backport-6.12/780-33-v6.15-r8169-disable-RTL8126-ZRX-DC-timeout.patch [2]
  bcm27xx/patches-6.12/950-0315-media-i2c-imx219-Correct-the-minimum-vblanking-value.patch [3]
  bcm27xx/patches-6.12/950-0857-drm-v3d-Add-clock-handling.patch [4]
  bcm27xx/patches-6.12/950-0874-PCI-brcmstb-Expand-inbound-window-size-up-to-64GB.patch [5]
  bcm27xx/patches-6.12/950-0877-PCI-brcmstb-Adding-a-softdep-to-MIP-MSI-X-driver.patch [6]
  bcm27xx/patches-6.12/950-0960-media-imx335-Set-vblank-immediately.patch [7]

Manually rebased patches:
  d1/patches-6.12/0009-ASoC-sunxi-sun4i-codec-add-basic-support-for-D1-audi.patch

[1] https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.12.31&id=ba59747562c49974cbace989d76b94a8331da442
[2] https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.12.31&id=2780aa8394415df0a69e3b908d6dd8c79e1d1bcc
[3] https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.12.31&id=9a981079097bee6a0583877798de0b240717bdde
[4] https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.12.31&id=bbd6dc1fb6c56267ad1d58810d92287fcd5b0058
[5] https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.12.31&id=3ffaa2e999380477774e76680ff4cef247451168
[6] https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.12.31&id=12153e3948c596737853c3ec4ff7e4e3f4a9f9a6
[7] https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.12.31&id=8d7e13c31c52690655883dff604238b0760a3644

Signed-off-by: Shiji Yang <yangshiji66@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/18953
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

cli: return the correct data type for enum attributes

Return entries from the value array instead of the provided string

Signed-off-by: Felix Fietkau <nbd@nbd.name>

mac80211: allow scanning operating on DFS channels

Only for multi-radio configurations where the scan would not affect the radio
running DFS detection

Signed-off-by: Felix Fietkau <nbd@nbd.name>

hostapd: tweak ACS retry behavior

When running ACS on multi-radio devices, ACS on one band can block another.
Increase the number of retries and prevent bouncing interfaces between AP
and STA mode during attempts.

Signed-off-by: Felix Fietkau <nbd@nbd.name>

cli: fix dealing with enum types with object value lists

Convert to array before calling filter()

Signed-off-by: Felix Fietkau <nbd@nbd.name>

unetmsg: set TCP timeout for outgoing connections

Avoid running into long timeouts

Signed-off-by: Felix Fietkau <nbd@nbd.name>

procd: update to Git HEAD (2025-05-31)

59d22013f9dd jail: make udebug socket optional

Fixes: https://github.com/openwrt/openwrt/issues/18967
Signed-off-by: Felix Fietkau <nbd@nbd.name>

generic: backport brcm legacy dsa tag fix

Backport a small fix for brcm legacy dsa tags that has been accepted for
linux v6.16.

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>

generic: backport bcm63xx SPI reset fix

Fixes SPI and HSSPI coexistance on bmips bcm6362 and bcm63268.

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>

realtek: enhance & harmonize dsa/phy max port patches

DSA silently drops internal phy access to ports >= 32 in dsa_user_phy_read()
and dsa_user_phy_write(). The code shows:

static int dsa_user_phy_read(struct mii_bus *bus, int addr, int reg)
{
struct dsa_switch *ds = bus->priv;

if (ds->phys_mii_mask & (1 << addr))
return ds->ops->phy_read(ds, addr, reg);

return 0xffff;
}

With ds->phys_mii_mask being a 32 bit variable the reason is clear. So do
not only increase the max values but also adapt the needed bitmasks in
the dsa and phy code. This fixes the dsa_user_ports() and dsa_cpu_ports()
too.

While we are here combine the old separated patches because dsa, mdio and
phy are tigthly coupled.

Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
Link: https://github.com/openwrt/openwrt/pull/18846
Signed-off-by: Robert Marko <robimarko@gmail.com>

qualcommbe: v6.12: enable 2.5G and 10G phylink modes in pcs-qcom-ipq9574

The PCS driver in the 6.12 patchset is the v5 submission (see link
below). It solves a number of issues and crashes with teh pcs driver
from the 6.6 patchset. However, this new driver is missing support for
"10gbase-r", "10g-qxgmii", and 1000/2500base-x modes.

Port these modes to the 6.12 patchset. "2500base-x" in particular seems
to be needed to establish a 2.5G link on phy-mode="usxgmii";

Link: https://lore.kernel.org/lkml/20250207-ipq_pcs_6-14_rc1-v5-0-be2ebec32921@quicinc.com/
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18796
Signed-off-by: Robert Marko <robimarko@gmail.com>

qualcommbe: enable v6.12 testing kernel

Now that patches for v6.12 are available, add the config, and enable
6.12 as a testing kernel.

Compared to v6.6 the patches to add "10gbase-r", "2500base-x",
"1000base-x", and "10g-qxgmii" modes have not been ported yet. They do
not apply cleanly to the new PCS driver. As the currently supported
hardware does not use those modes, it is not possible to test them at
this time.

Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18796
Signed-off-by: Robert Marko <robimarko@gmail.com>

qualcommbe: add v6.12 config from v6.6

Copy the v6.6 config to v6.6. It will be updated in a subsequent
commit in order to show the changes.

Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18796
Signed-off-by: Robert Marko <robimarko@gmail.com>

qualcommbe: v6.12: add NSSCC clock and reset fixes

Port reset and clock fixes from target/linux/qualcommbe/patches-6.6.
Modifications of v6.6 patches are noted in each commit body.

Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18796
Signed-off-by: Robert Marko <robimarko@gmail.com>

qualcommbe: v6.12: enable ethernet for RDP433

Now that ethernet is supported on IPQ9574, add the port nodes and
other requirements to enable ethernet on RDP433.

Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18796
Signed-off-by: Robert Marko <robimarko@gmail.com>

qualcommbe: v6.12: add PPE driver (part 2)

Add the second part of the PPE driver. This includes the EDMA and
network device support. This part does not appear to have been
officially submitted for upstream review. The series is taken from
target/linux/qualcommbe/patches-6.6, and had to be heavily modified
in order to compile of v6.12. Changes to patches are noted in the
respective patch body.

Also add the PPE and EDMA nodes in this series.

Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18796
Signed-off-by: Robert Marko <robimarko@gmail.com>

qualcommbe: v6.12: add pending PPE driver (part 1)

Add v3 submission of the qualcomm PPE driver. As of this writing, it
is the latest version. This lacks the EDMA driver and network device
support. That will be added in part 2.

Link: https://lore.kernel.org/lkml/20250209-qcom_ipq_ppe-v3-0-453ea18d3271@quicinc.com
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18796
Signed-off-by: Robert Marko <robimarko@gmail.com>

qualcommbe: v6.12: add ethernet PCS driver for IPQ9574

Add the v5 of the PCS patch. This is the latest submission as of this
writing. THe last four patches are not part of the submission. They
make the series work with v6.12 kernel, resolve a circular dependency
with the clocks, and add the DTS node. Include them as bundle.

Link: https://lore.kernel.org/lkml/20250207-ipq_pcs_6-14_rc1-v5-0-be2ebec32921@quicinc.com/
Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18796
Signed-off-by: Robert Marko <robimarko@gmail.com>

qualcommbe: v6.12: add QCA8084 ethernet PHY driver

This driver is cherry-picked from target/linux/qualcommbe/patches-6.6.
While Qualcomm did submit past patches for QCA8084, the code has since
ben split from at803x. The existing OpenWRT version of the patch is
the cleanest version I could find. Add it here.

Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18796
Signed-off-by: Robert Marko <robimarko@gmail.com>

qualcommbe: v6.12: add various DTS updates for IPQ9574 and RDP433

These devicetree updates can work with the existing upstreamcode. They
do not require code changes. Add them before any code change.

Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18796
Signed-off-by: Robert Marko <robimarko@gmail.com>

qualcommbe: v6.12: add pending patches

Add dts fixes from linux-next. Two patches from the NSSCC series are
still in -next did not yet land in mainline, as well as misc other DTS
changes. Add them here.

Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18796
Signed-off-by: Robert Marko <robimarko@gmail.com>

qualcommbe: v6.12: add upstreamed patches

Add relevant patches from upstream, up to v5.16-rc6. The gaps in the
patch numbersing are either patches that were picked into the stable
kernel (6.12.y), or that are already backported in
target/linux/generic. The gaps makes it easy for me to pick these
patches from my working kernel git branch.

Signed-off-by: Alexandru Gagniuc <mr.nuke.me@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18796
Signed-off-by: Robert Marko <robimarko@gmail.com>

base-files: use cidr for config_generate

Generate network configuration replacing netmask with CIDR.

Depends on:
https://github.com/openwrt/openwrt/pull/13765

Using CIDR provides the following advantages:
* Consolidate notation for IPv4 and IPv6 addresses.
* Consolidate notation for IP addresses and routing targets.
* Simplify network configuration and troubleshooting.
* Follow the transition from net-tools to iproute2.

Resulting configuration example:
```
config interface 'loopback'
option device 'lo'
option proto 'static'
list ipaddr '127.0.0.1/8'

config interface 'lan'
option device 'br-lan'
option proto 'static'
list ipaddr '192.168.1.1/24'
```

Signed-off-by: Vladislav Grigoryev <vg.aetera@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/13780
Signed-off-by: Robert Marko <robimarko@gmail.com>

ath79: add calibtation variant for TP-Link Archer C6 v2

Now that we have a board file, add calibration variant for TP-Link
Archer C6 v2 and add ipq-wifi package for it.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>

ipq-wifi: Add entry for TP-Link Archer C6 v2

Add IPQ Wifi entry for Ath79 TP-Link Archer C6 v2.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>

ipq-wifi: update to Git HEAD (2025-05-30)

576ade747259 CI: add action that prints info about submited BDF-s
1d688e8654a5 qca9888 : add BDF for Yuncore WB5G08
ddfd9d2f3a8e ipq60xx: rename BDF for EAP623-Outdoor HD v1
bab36c1d947d ath10k: drop upstramed board files
09e5afab85b2 CI: add QCA9988
fde3d2a7ce59 qca9888: add TP-Link Archer C6 v2 (EU) BDF

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>

sunxi: enable crypto offload engine

Enable the sun8i_ce and sun8i_ss crypto engines that are included on
various Allwinner SoCs, along with their respective hash/prng/trng
options.

sun8i_ce: H2+, H3, H5, H6, R40 and A64
sun8i_ss: A80, A83T

BananaPi P2 Zero (H2+):

[    0.373860] sun8i-ce 1c15000.crypto: Set mod clock to 50000000 (50 Mhz) from 24000000 (24 Mhz)
[    0.374326] sun8i-ce 1c15000.crypto: will run requests pump with realtime priority
[    0.374516] sun8i-ce 1c15000.crypto: will run requests pump with realtime priority
[    0.374657] sun8i-ce 1c15000.crypto: will run requests pump with realtime priority
[    0.374788] sun8i-ce 1c15000.crypto: will run requests pump with realtime priority
[    0.374924] sun8i-ce 1c15000.crypto: Register cbc(aes)
[    0.374962] sun8i-ce 1c15000.crypto: Register ecb(aes)
[    0.374976] sun8i-ce 1c15000.crypto: Register cbc(des3_ede)
[    0.374990] sun8i-ce 1c15000.crypto: Register ecb(des3_ede)
[    0.375005] sun8i-ce 1c15000.crypto: Register md5
[    0.375019] sun8i-ce 1c15000.crypto: Register sha1
[    0.375032] sun8i-ce 1c15000.crypto: Register sha224
[    0.375046] sun8i-ce 1c15000.crypto: Register sha256
[    0.375061] sun8i-ce 1c15000.crypto: Register sha384
[    0.375075] sun8i-ce 1c15000.crypto: Register sha512
[    0.375109] sun8i-ce 1c15000.crypto: Register stdrng
[    0.375150] sun8i-ce 1c15000.crypto: TRNG not supported
[    0.375157] sun8i-ce 1c15000.crypto: CryptoEngine Die ID 1

BananaPi M3 (A83t):

[    0.429615] sun8i-ss 1c15000.crypto: will run requests pump with realtime priority
[    0.429866] sun8i-ss 1c15000.crypto: will run requests pump with realtime priority
[    0.430042] sun8i-ss 1c15000.crypto: DEBUG: Register cbc(aes)
[    0.430095] sun8i-ss 1c15000.crypto: DEBUG: Register ecb(aes)
[    0.430112] sun8i-ss 1c15000.crypto: DEBUG: Register cbc(des3_ede)
[    0.430128] sun8i-ss 1c15000.crypto: DEBUG: Register ecb(des3_ede)
[    0.430152] sun8i-ss 1c15000.crypto: Register md5
[    0.430171] sun8i-ss 1c15000.crypto: Register sha1
[    0.430187] sun8i-ss 1c15000.crypto: Register sha224
[    0.430203] sun8i-ss 1c15000.crypto: Register sha256
[    0.430220] sun8i-ss 1c15000.crypto: Register hmac(sha1)
[    0.430276] sun8i-ss 1c15000.crypto: Security System Die ID 7
[    0.493087] sun8i-ss 1c15000.crypto: Fallback driver could no be loaded

OrangePi Zero 2 (H616):

[    0.176833] sun8i-ce 1904000.crypto: will run requests pump with realtime priority
[    0.177054] sun8i-ce 1904000.crypto: will run requests pump with realtime priority
[    0.177167] sun8i-ce 1904000.crypto: will run requests pump with realtime priority
[    0.177258] sun8i-ce 1904000.crypto: will run requests pump with realtime priority
[    0.177361] sun8i-ce 1904000.crypto: Register cbc(aes)
[    0.177408] sun8i-ce 1904000.crypto: Register ecb(aes)
[    0.177420] sun8i-ce 1904000.crypto: Register cbc(des3_ede)
[    0.177432] sun8i-ce 1904000.crypto: Register ecb(des3_ede)
[    0.177444] sun8i-ce 1904000.crypto: Register md5
[    0.177456] sun8i-ce 1904000.crypto: Register sha1
[    0.177468] sun8i-ce 1904000.crypto: Register sha224
[    0.177479] sun8i-ce 1904000.crypto: Register sha256
[    0.177503] sun8i-ce 1904000.crypto: Register sha384
[    0.177515] sun8i-ce 1904000.crypto: Register sha512
[    0.177527] sun8i-ce 1904000.crypto: Register stdrng
[    0.177648] sun8i-ce 1904000.crypto: CryptoEngine Die ID 0

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>

sunxi: cortexa53: enable rtc drivers

RTC has been disabled so far in the cortexa53 subtarget so far, enable
it. It's also required for the crypto trng/prng drivers on these SoC
variants (to gather IOSC noise).

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>

sunxi: drop 6.6 support

Drop configs and patches for Linux 6.6.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>

sunxi: use kernel 6.12 by default

Switch to Linux kernel version 6.12.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>

pistachio: add 6.12 testing kernel

Allow selecting 6.12 as testing kernel on pistachio.

Runtime-tested:
 - Creator Ci40 Marduk 02KZ/5+

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>

pistachio: 6.12: refresh patches and kernel config

Refresh patches and configs as required.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>

kernel/pistachio: Restore kernel files for v6.6

This is an automatically generated commit which aids following Kernel patch
history, as git will see the move and copy as a rename thus defeating the
purpose.

For the original discussion see:
https://lists.openwrt.org/pipermail/openwrt-devel/2023-October/041673.html

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>

kernel/pistachio: Create kernel files for v6.12 (from v6.6)

This is an automatically generated commit.

When doing `git bisect`, consider `git bisect --skip`.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>

qmi: increase SIM power-cycle timeouts

Some modems and SIM cards take a bit longer to initialize after UIM has been
powered off. Waiting too little time can cause the qmi protocol to end up
in a loop repeatedly power-cycling the SIM card.

Avoid that by
 a) increasing the time we unconditionally sleep after --uim-power-on
 b) increasing the time we allow uqmi to wait for response for --uim-get-sim-state

Signed-off-by: Antti Seppälä <a.seppala@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18772
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

base-files: treat 'ipaddr is inside range' as warning again

The call to 'ipcalc' is used in 'dnsmasq' init script to create the
configuration. If the 'ipaddr' is in the configured range then 'ipcalc' exited
with an error whereby the START/STOP variables are unavailable.

This behaviour has changed during 'ipcalc' refactoring and now leads to a
problem when starting 'dnsmasq' if the 'ipaddr' is inside this range. To
restore the old behaviour, only a warning is issued as before and the
required variables for the 'dnsmasq' are still set.

Fixes: 854739b32c7f (base-files: ipcalc.sh: Rewrite in pure shell)
Signed-off-by: Florian Eckert <fe@dev.tdt.de>
Link: https://github.com/openwrt/openwrt/pull/18641
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

dnsmasq: fix start if dhcp-range is not correct

If the uci 'dhcp' configuration for the dhcp leases is incorrect then
the call to 'ipclac' fails. However, the problem is that the dnsmasq
configuration option 'dhcp-range' is still written for this uci section
even though the information generated by ipcalc is incorrect or not set.

Due to the incorrectly generated configuration for dnsmasq, the service
cannot start.

To prevent an incorrect configuration from being written to the configuration,
a check is now made beforehand to ensure that the required variables are
present and valid. If the configuration is incorrect, a message is emitted
to the log that this configuration section is incorrect and this uci
configuration section is omitted.

Signed-off-by: Florian Eckert <fe@dev.tdt.de>
Link: https://github.com/openwrt/openwrt/pull/18641
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

wifi-scripts: add missing 'network' property to wifi-iface schema

The ucode-based wifi interface validation is based on `hostapd.conf`
specific options, which means it's missing the OpenWrt-specific
'network' property.

This causes schema validation warnings like:
```
daemon.notice netifd: radio1 (1340): wifi-scripts: network is not present in the schema
```

The description is taken from the OpenWrt wiki:
https://openwrt.org/docs/guide-user/network/wifi/basic#common_options1

Signed-off-by: Sean Khan <datapronix@protonmail.com>
Link: https://github.com/openwrt/openwrt/pull/18946
Signed-off-by: Robert Marko <robimarko@gmail.com>

generic: power: backport v6.15 patch to fix build with GCC15

Backports an upstream fix ([1]) for a build regression of
'drivers/power/supply' with GCC15 -Wunterminated-string-initialization.

[1] https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=77f5bb150132bbbcd6bc37ffdc80c9e140e373a4

Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18950
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

generic: net: add pending wireguard patch to fix build with GCC15

This patch adds __nonstring annotation to address GCC15 with
-Wunterminated-string-initialization.
It is accepted in wireguard-linux in [1] and pending for v6.16
in [2].

[1] https://git.zx2c4.com/wireguard-linux/commit/?h=devel&id=71e5da46e78c1cd24e2feed251a2845327447ad8
[2] https://patch.msgid.link/20250521212707.1767879-3-Jason@zx2c4.com

Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18950
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>