mediatek: mt7988a-rfb: align order of SFPs with SDK build

MediaTek SDK uses eth2 as additional LAN port and eth1 as WAN.
Do the same also in OpenWrt to keep settings aligned with the SDK.

Suggested-by: Bc-bocun Chen <bc-bocun.chen@mediatek.com>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>

qoriq: enable 6.6 testing kernel

The qoriq now supports 6.6 kernel as testing.

Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/16079
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>

qoriq: refresh config-6.6

Refresh kernel config for Linux 6.6.

Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/16079
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>

kernel/qoriq: Restore kernel files for v6.1

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: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/16079
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>

kernel/qoriq: Create kernel files for v6.6 (from v6.1)

This is an automatically generated commit.

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

Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/16079
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>

kernel: bump 6.1 to 6.1.107

https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.1.107

All patches automatically rebased.

Build system: bcm53xx octeon

Signed-off-by: Zxl hhyccc <zxlhhy@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16283
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel: bump 6.6 to 6.6.48

Changelog: https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.6.48

Removed upstreamed:
generic/backport-6.6/752-04-v6.6-net-ethernet-mtk_wed-check-update_wo_rx_stats-in-mtk.patch[1]
generic/pending-6.6/601-udp-fix-receiving-fraglist-GSO_packets.patch[2]
mediatek/patches-6.6/951-net-ethernet-mtk_wed-fix-use-after-free-panic-in-mtk.patch[3]
rockchip/patches-6.6/302-mmc-allow-probe-to-defer-if-clock-is-not-ready.patch[4]

Manually rebased:
bcm27xx/patches-6.6/950-0058-Revert-Bluetooth-Always-request-for-user-confirmatio.patch

All other patches automatically rebased.

1. https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.6.48&id=7ede6ef04c9ffc466844a297d3d68bea8a06093d
2. https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.6.48&id=cfa692e7e71b22d35174a796d10b9111bafa0686
3. https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.6.48&id=326a89321f9d5fe399fe6f9ff7c0fc766582a6a0
4. https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.6.48&id=7bbf8f8b8861dd2a6215f1aded425c66e6c376a0

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

Signed-off-by: John Audia <therealgraysky@proton.me>
Link: https://github.com/openwrt/openwrt/pull/16282
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

netfilter: add nf_conntrack_sane to kmod-nf-nathelpers-extra

'SANE' is a protocol for remote access to scanners as implemented by the
'saned' daemon. Like FTP, it uses separate control and data connections.

So let´s enable this in the kernel and add them to 'kmod-nf-nathelper-extra'.

Signed-off-by: Florian Eckert <fe@dev.tdt.de>

netfilter: add nf_conntrack_netbios_ns to kmod-nf-nathelper-extra

NetBIOS name service requests are sent as broadcast messages from an
unprivileged port and responded to with unicast messages to the
same port. This make them hard to firewall properly because connection
tracking doesn't deal with broadcasts.

So let´s enable this in the kernel and add them to 'kmod-nf-nathelper-extra'.

Signed-off-by: Florian Eckert <fe@dev.tdt.de>

ltq-vdsl-vr11-app: perform orderly shutdown on exit

Try to do a clean disconnection via L3 request before the connection is
stopped.

Because this might take up to 6 seconds (the driver does 3 attempts with
a timeout of 2 seconds each), a termination timeout needs to be defined
in the init script.

Signed-off-by: Jan Hoffmann <jan@3e8.eu>

ltq-vdsl-vr11-app: always disconnect on exit

Move the code for disconnection on exit to a separate function, and also
call it in the code paths for SIGINT and the "quit" CLI command.

While at it, make the patch description a bit clearer.

Signed-off-by: Jan Hoffmann <jan@3e8.eu>

ltq-vdsl-vr9-app: always disconnect on exit

Move the code for disconnection on exit to a separate function, and also
call it in the code path for the "quit" CLI command.

While at it, make the patch description a bit clearer.

Signed-off-by: Jan Hoffmann <jan@3e8.eu>

ltq-vdsl-vr11-app: fix error handling during disconnection

Use the correct return value in error message.

Fixes: 6e4c9738bed0 ("ltq-vdsl-vr11-app: add version 4.23.1 for vr11 targets")
Signed-off-by: Jan Hoffmann <jan@3e8.eu>

ltq-vdsl-vr9-app: fix error handling during disconnection

Use the correct return value in error message.

Fixes: 1daaef31b352 ("ltq-vdsl-app: disconnect when service is stopped")
Signed-off-by: Jan Hoffmann <jan@3e8.eu>

kernel: modules: usb: remove deprecated Kconfig option CONFIG_USB_XHCI_HCD_DEBUGGING

The Kconfig option 'CONFIG_USB_XHCI_HCD_DEBUGGING' has been removed with the
following commit upstream in the Linux kernel.

https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?id=b2497509df002e9a09c8550cd0ecd2f77c9640d8

This Kconfig option is therefore no longer valid for the kernel version
6.6 and could be removed.

Signed-off-by: Florian Eckert <fe@dev.tdt.de>

ath79: qihoo,c301: use nvmem

Userspace handling is deprecated.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16244
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

bcm4908: drop 6.1 support

Drop config and files for Linux 6.1.

Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/16057
Signed-off-by: Robert Marko <robimarko@gmail.com>

bcm4908: use kernel 6.6 by default

Switch to Linux kernel version 6.6.

Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/16057
Signed-off-by: Robert Marko <robimarko@gmail.com>

uboot-envtools: ath79: add Belkin F9K1115v2/F9J1108v2 u-boot env support

Add support for Belkin F9K1115v2/F9J1108v2 u-boot env

Signed-off-by: Paweł Owoc <frut3k7@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16033
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

ath79: nec,wg800hp: convert to nvmem

Userspace handling is deprecated.

Unused wan mac added to dts. Requires DSA to implement.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16246
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

ath79: elecom,wab: use nvmem

Userspace handling is deprecated.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16247
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

base-files: fix merge of passwd/shadow/group lines with trailing colons

Empty trailing fields get lost when the lines are split and merged again
at colons, resulting in unparsable entries. Only use the split fields for
matching against the other file, but emit the original line unchanged
to fix the issue.

Fixes: de7ca7dafadf ("base-files: merge /etc/passwd et al at sysupgrade config restore")
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>

ath79: nec,wX1200cr: use nvmem

Userspace handling is deprecated.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16245
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel: set vlan mc emulation for rtl8367d family chips

Set vlan mc emulation for rtl8367d family chips

Co-authored-by: Serge Vasilugin <vasilugin@yandex.ru>
Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/14804
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel: set RGMII mode for rtl8367d family chips

Set RGMII mode for rtl8367d family chips

Co-authored-by: Serge Vasilugin <vasilugin@yandex.ru>
Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/14804
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel: set vlan mc for rtl8367d family chips

Set vlan mc for rtl8367d family chips

Co-authored-by: Serge Vasilugin <vasilugin@yandex.ru>
Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/14804
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel: set and get vlan 4k for rtl8367d family chips

Set and get vlan 4k for rtl8367d family chips

Co-authored-by: Serge Vasilugin <vasilugin@yandex.ru>
Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/14804
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel: set and get mc index for rtl8367d family chips

Set and get mc index for rtl8367d family chips

Co-authored-by: Serge Vasilugin <vasilugin@yandex.ru>
Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/14804
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel: set and get ports speed for rtl8367d family chips

Set and get ports speed for rtl8367d family chips

Co-authored-by: Serge Vasilugin <vasilugin@yandex.ru>
Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/14804
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel: rtl8367b: detect rtl8367d chip family

Detect the RTL8367D chip family and set the appropriate extif

Co-authored-by: Serge Vasilugin <vasilugin@yandex.ru>
Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/14804
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel: netdevices: adjust mlxsw network drivers

* Adds the x86_64 dependency for mlxsw_core

* Removes the redundant mlxsw_core dependency
from mlxsw-minimal and mlxsw-spectrum

* Removes the DCB configuration symbols because
they were moved into the generic configuration

Signed-off-by: Til Kaiser <mail@tk154.de>
Link: https://github.com/openwrt/openwrt/pull/15362
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

generic: 6.6: add missing DCB symbols

Disables Data Center Bridging support for some network drivers by
default to avoid compilation errors when CONFIG_DCB is enabled.

Signed-off-by: Til Kaiser <mail@tk154.de>
Link: https://github.com/openwrt/openwrt/pull/15362
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

linux-firmware: add firmware for all Mellanox Spectrum Switches

This commit adds the currently missing Mellanox
Spectrum-2, Spectrum-3, and Spectrum-4 firmware files.

Signed-off-by: Til Kaiser <mail@tk154.de>
Link: https://github.com/openwrt/openwrt/pull/15362
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel: modules: add x86 dependency for leds-mlxcpld

This commit adds the missing x86 dependency for
the Mellanox mlxcpld LED driver.

Signed-off-by: Til Kaiser <mail@tk154.de>
Link: https://github.com/openwrt/openwrt/pull/15362
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

rockchip: add support for Radxa ROCK 3B

Radxa ROCK 3B is a Pico-ITX form factor SBC[1] using the Rockchip
RK3568(J).

Hardware
--------
- Rockchip RK3568(J) SoC
- Quad A55 CPU
- Mali-G52 GPU
- 1 TOPS @ INT8 NPU
- 2GB/4GB/8GB LPDDR4 RAM
- eMMC connector
- Micro SD Card slot
- NVMe SSD through the M.2 M Key (2-lane PCIe 3.0)
- SPI Flash for bootloader
- 2x Gigabit ethernet port (one supports PoE with add-on PoE HAT)
- 1x M.2 E Key socket with SDIO, UART and USB interfaces
- 1x M.2 B Key socket with PCIe, SATA, and USB interfaces
- 1x SIM card socket
- 1x USB 3.0 Type-A HOST port
- 1x USB 3.0 Type-A OTG port
- 2x USB 2.0 Type-A HOST ports
- 40 Pin GPIO header

[1] https://radxa.com/products/rock3/3b

Installation
------------
Uncompress the OpenWrt sysupgrade and write it to a micro SD card or
internal eMMC using dd.

Signed-off-by: FUKAUMI Naoki <naoki@radxa.com>
Link: https://github.com/openwrt/openwrt/pull/16185
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

rockchip: add support for Radxa ZERO 3E and 3W

Radxa ZERO 3E and 3W are light, compact and tiny SBC[1][2] using the
Rockchip RK3566.

Hardware
--------
- Rockchip RK3566 SoC
- Quad A55 CPU
- Mali-G52-2EE GPU
- 1 TOPS @ INT8 NPU
- 1GB/2GB/4GB/8G LPDDR4 RAM
- Optional 8GB/16GB/32GB/64GB eMMC (3E)
- Micro SD Card slot
- 1x Gigabit ethernet port (supports PoE with add-on PoE HAT) (3E)
- WiFi6/BT5.4 (3W) (not supported yet on OpenWrt)
- 1x USB 3.0 Type-C HOST port
- 1x USB 2.0 Type-C OTG port
- Optional 40 Pin GPIO header

[1] https://radxa.com/products/zeros/zero3e
[2] https://radxa.com/products/zeros/zero3w

Installation
------------
Uncompress the OpenWrt sysupgrade and write it to a micro SD card or
internal eMMC using dd.

Signed-off-by: FUKAUMI Naoki <naoki@radxa.com>
Link: https://github.com/openwrt/openwrt/pull/16185
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

rockchip: add support for Radxa ROCK 3C

Radxa ROCK 3C is a high-performance, low-cost SBC[1] using the
Rockchip RK3566.

Hardware
--------
- Rockchip RK3566 SoC
- Quad A55 CPU
- Mali-G52-2EE GPU
- 1 TOPS @ INT8 NPU
- 1GB/2GB/4GB LPDDR4 RAM
- eMMC connector
- Micro SD Card slot
- NVMe SSD through the M.2 M Key connector(2230) or M.2 Extension
  board(2232/2260/2280)
- SATA through the Radxa Penta SATA HAT
- 1x Gigabit ethernet port(supports PoE with add-on PoE HAT)
- WiFi6/BT5.4 (not supported yet on OpenWrt)
- 1x USB 3.0 Type-A HOST port
- 2x USB 2.0 Type-A HOST ports
- 1x USB 2.0 Type-A OTG port
- 40 Pin GPIO header

[1] https://radxa.com/products/rock3/3c

Installation
------------
Uncompress the OpenWrt sysupgrade and write it to a micro SD card or
internal eMMC using dd.

Signed-off-by: FUKAUMI Naoki <naoki@radxa.com>
Link: https://github.com/openwrt/openwrt/pull/16185
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

uboot-rockchip: renumber patches

renumber patches for upcoming new device support

Signed-off-by: FUKAUMI Naoki <naoki@radxa.com>
Link: https://github.com/openwrt/openwrt/pull/16185
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

tools: libtool: do not symlink files in bootstrap

Another instance of files in build_dir symlinking to staging_dir. While
the symlinks do not currently cause any bugs in the libtool package,
such symlinks were found to make the build more fragile, as writing to
the symlink may accidentally modify the shared file in staging_dir. Pass
--copy to bootstrap to disable the symlinking.

Link: https://github.com/openwrt/openwrt/pull/15825
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>

include: autotools: do not symlink files in autoreconf

In Gluon's Github Actions CI, we were occasionally seeing bizarre build
errors that looked like a config.sub file had been corrupted, or changed
while it was being executed.

The cause turned out to be an interaction of the symlinks created by
autoreconf (pointing from individual tools' build dirs into
`staging_dir/host/share/automake-1.16`) and OpenWrt's host-build.mk,
which replaced config.guess and config.sub *after* autoreconf. The
result was that the replacement of these files ended up following the
symlinks and writing the files in `staging_dir/host/share/automake-1.16`
instead of a package's build dir. This could cause other packages' builds
to fail if they were currently executing the scripts while they were
being written.

To fix this, disable autoreconf's symlinking feature, so that modifying
these files in a package's build directory can't accidentally affect the
staged versions.

Link: https://github.com/openwrt/openwrt/pull/15825
Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>

build: fix CMake generator for non-Ninja builds

OpenWRT by default uses the Ninja generator, but some packages disable
Ninja and use the default Unix Makefiles generator. This generator can
be overridden in the user environment with `CMAKE_GENERATOR`. This patch
explicitly sets the correct generator when `PKG_USE_NINJA:=0`.

In particular, the `mt76` package uses the Makefiles generator.

Signed-off-by: Ivan Romanov <drizt72@zoho.eu>
Link: https://github.com/openwrt/openwrt/pull/16263
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

mediatek: phase out uImage.FIT partition parser

All boards using the deprecated uImage.FIT partition parser have
been migrated to the new fitblk driver. Drop the now no longer
needed partition parser.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>

mediatek: filogic: convert Zyxel EX5601-T0 to fitblk

Use fitblk driver instead of deprecated partition parser.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>

mediatek: filogic: convert Netcore N60 to fitblk

Use fitblk driver instead of deprecated partition parser.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>

mediatek: filogic: convert Xiaomi WR30U to fitblk

Use fitblk driver instead of deprecated partition parser.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>

mediatek: filogic: convert Xiaomi AX3000T to fitblk

Use fitblk driver instead of deprecated partition parser.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>

mediatek: filogic: convert Qihoo 360T7 to fitblk

Use fitblk driver instead of deprecated partition parser.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>

mediatek: filogic: convert JCG Q30 PRO to fitblk

Use fitblk driver instead of deprecated partition parser.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>

mediatek: filogic: convert H3C Magic NX30 Pro to fitblk

Use fitblk driver instead of deprecated partition parser.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>

mediatek: filogic: convert MT7981 RFB to fitblk

Use fitblk driver instead of deprecated uImage.FIT partition parser.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>

uboot-mediatek: detect rootdisk on MT7981

Set /chosen/rootdisk according to boot_mode register, similar to
MT7986.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>

generic: hack: fix patch description

That patch title of 911-kobject_add_broadcast_uevent.patch has been
wrongly copied from 910-kobject_uevent.patch.
Change the description from "lib: add uevent_next_seqnum()" to
"lib: add broadcast_uevent()", so that the git history doesn't look
all weird when importing both patches to a git tree.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>

hostapd: fix anqp_3gpp_cell_net list delimiter

This patch fixes the list delimiter between 3GPP networks
passed to hostapd.

> list iw_anqp_3gpp_cell_net '262,001'
> list iw_anqp_3gpp_cell_net '262,002'

When passing a list of "iw_anqp_3gpp_cell_net" parameters via UCI,
hostapd would crash at startup:
> daemon.err hostapd: Line 73: Invalid anqp_3gpp_cell_net: 262,001:262,002

Using a semicolon as a delimiter, hostapd will start as expected.

Signed-off-by: Sarah Maedel <git@tbspace.de>

Revert "CI: update actions/labeler to v5"

This reverts commit a70555fb4f9b008494007e69dd520651b3da608d.

It is breaking GH PR-s currently, so revert until fixed.

Signed-off-by: Robert Marko <robimarko@gmail.com>

CI: update actions/labeler to v5

* Version 5 of this action updated the runtime to Node.js 20. All scripts are now run with Node.js 20 instead of Node.js 16 and are affected by any breaking changes between Node.js 16 and 20.

Signed-off-by: Goetz Goerisch <ggoerisch@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16251
Signed-off-by: Robert Marko <robimarko@gmail.com>

uqmi: update to latest HEAD

28b48a1 uim: add support for ICC communication channel
f582e00 qmi: fix dynamic array macro
d381f80 data: add support for ICC channel

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

mt76: update to Git HEAD (2024-08-25)

dbc9b0df7c01 linux-firmware: update firmware for MT7996
4e8167b94175 wifi: mt76: connac: fix checksum offload fields of connac3 RXD
084eaaf4792e wifi: mt76: mt7915: fix sta poll list corruption on hw restart
aea16bfe8d50 firmware: update MT7981 firmware to version 20240507201102
65bbd4c394a9 firmware: update MT7986 firmware to version 20240507160203
1f4ac8faa529 wifi: mt76: connac: move mt7615_mcu_del_wtbl_all to connac
b76f16ea6521 wifi: mt76: mt7915: improve hardware restart reliability
e6afe9218a27 wifi: mt76: mt7915: fix unused variable error
d6611d015efd firmware: update mt7916/mt7981/mt7986 firmware to version 20240823
ddeb304aae6d wifi: mt76: mt7921: fix wrong UNII-4 freq range check for the channel usage
904ef52a8d04 wifi: mt76: mt7996: fix uninitialized TLV data

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

ustream-ssl: update to Git HEAD (2024-07-28)

99bd3d2b167c ustream-openssl: fix compilation with OPENSSL_NO_DEPRECATED

Signed-off-by: Petr Štetiar <ynezz@true.cz>
Link: https://github.com/openwrt/openwrt/pull/16020
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

mediatek: filogic: add support for Cudy AP3000 Outdoor

Hardware highlights:
  - MediaTek MT7981 WiSoC
  - 256MB DDR3 RAM
  - 64MB SPI-NAND
  - MediaTek MT7981 2x2 DBDC 802.11ax 2T2R (2.4 / 5)
  - 1x LED in two colors
  - 2x Button
  - 1x 1GbE
  - Two detachable antenas, two internal pcb antenas
  - PoE powered (standalone)

MAC:
LAN MAC: label mac
2.4G MAC: label mac -1
5G MAC: label mac

How to install:
1. Apply Cudy Intermediate OpenWrt image for AP3000 Outdoor V1
2. Install OpenWrt sysupgrade image

Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16156
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

treewide: rename ZyXEL to Zyxel

The company Zyxel rebranded some years ago.
Currently the casing is according to the old branding even
for newer devices which already use the new branding.

This commit aligns the casing of Zyxel everywhere.

Signed-off-by: Goetz Goerisch <ggoerisch@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/15652
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

rockchip: fix model name for Radxa ROCK 3A

This patch fixes model name in dts as below:

Radxa ROCK3 model A -> Radxa ROCK 3A

Signed-off-by: FUKAUMI Naoki <naoki@radxa.com>
Link: https://github.com/openwrt/openwrt/pull/16232
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

uboot-rockchip: fix model name for Radxa ROCK 3A, 5A, and 5B

This patch fixes model name in dts as below:

Radxa ROCK3 model A  -> Radxa ROCK 3A
Radxa ROCK 5 model A -> Radxa ROCK 5A
Radxa ROCK 5 model B -> Radxa ROCK 5B

Signed-off-by: FUKAUMI Naoki <naoki@radxa.com>
Link: https://github.com/openwrt/openwrt/pull/16232
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

armsr: use console=tty1 to make console more readily available

Like x86, armsr is frequently virtualized, and is used for development
and debugging. Kernel messages should be more readily apparent by
default. This can be achieved by adding console=tty1 to the kernel
command line, enabling the console on a (possibly virtual) display and
keyboard, in addition to a serial port.

This also enables failsafe on tty1. Failsafe mode operates on consoles
known by the kernel, without regard to /etc/inittab.

armsr's /etc/inittab is also updated to specify tty1 instead of tty0.
tty1 is technically more correct: tty1 is the first virtual console,
where tty0 reflects the current active virtual console (which is likely
to be tty1).

This configuration matches x86, which is another target commonly used
for virtualization, development, and debugging in the same way. x86's
kernel command line had specified console=tty0, although console=tty1 is
more correct for the reasons given above. This also brings x86's kernel
command line console= into agreement with its /etc/inittab, which
already used tty1.

Signed-off-by: Mark Mentovai <mark@mentovai.com>
Link: https://github.com/openwrt/openwrt/pull/16213
Signed-off-by: Robert Marko <robimarko@gmail.com>

ramips: mt7621_nand: don't set owner

Found with coccinelle:

No need to set .owner here. The core will do it.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16217
Signed-off-by: Robert Marko <robimarko@gmail.com>

trace-cmd: update to 3.3

Use local tarballs instead of upstream generated ones. Smaller.

Fix version to be compatible with apk.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16219
Signed-off-by: Robert Marko <robimarko@gmail.com>

dnsmasq: Add EDNS0 Upstream support

Forward client mac address and subnet on dns queries. Pi-hole and Adguard use this feature to send the originators ip address/subnet so it can be logged and not just the nat address of the router. This feature has been added since version 2.56 of dnsmasq and would be nice to expose this feature in openwrt.

Signed-off-by: Carsten Schuette <schuettecarsten@googlemail.com>
Link: https://github.com/openwrt/openwrt/pull/15965
Signed-off-by: Robert Marko <robimarko@gmail.com>

bmips: add support for TP-Link TD-W8968 V3

TP-Link TD-W8968 v3 is an 300Mbps Wireless N USB ADSL2+ Modem Router based on
Broadcom BCM6318 SoC.

Hardware:
   CPU:          Broadcom BCM6318, 333 MHz, 1 core
   Flash:        8MB
   RAM:          64 MB
   Ethernet:     4x 10/100 Mbps
   Wireless:     802.11b/g/n, BCM43217
   LEDs/Buttons: 10x / 3x
   USB:          1x 2.0

Flash instructions:
* Assign static IP 192.168.1.100 to PC
* Unplug the power source
* Press the RESET button at the router, don't release it yet!
* Plug the power source. Wait for some seconds
* Release the RESET button
* Browse to http://192.168.1.1
* Upload the openwrt-bmips-bcm6318-tp-link_td-w8968-v3-squashfs-cfe.bin file
* Wait some minutes until the firmware upgrade finish.

Signed-off-by: Mathesh Velayudan <123v.mathesh@gmail.com>

generic: 6.6: backport upstream Realtek PHY patch

- Fix setting of PHY LEDs Mode B bit on RTL8211F.
- Rename pending Realtek PHY patches.

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

mt76: Change AutoLoad macro call for mt7925e and mt7925u

There was a typo done for mt7925e and mt7925u in the KernelPackage
definitions, which caused the system to load the wrong kernel modules.

Signed-off-by: Daniel Pawlik <pawlik.dan@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16236
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

rockchip: fix compatible string for Radxa ROCK 5A

revert unwanted change for compatible string.

Signed-off-by: FUKAUMI Naoki <naoki@radxa.com>
Link: https://github.com/openwrt/openwrt/pull/16221
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

rockchip: add support for nanopc t6

SoC: Rockchip RK3588
CPU: Quad-core ARM Cortex-A76(up to 2.4GHz) and quad-core Cortex-A55 CPU (up to 1.8GHz)
GPU: Mali-G610 MP4, compatible with OpenGLES 1.1, 2.0, and 3.2, OpenCL up to 2.2 and Vulkan1.2
VPU: 8K@60fps H.265 and VP9 decoder, 8K@30fps H.264 decoder, 4K@60fps AV1 decoder, 8K@30fps H.264 and H.265 encoder
NPU: 6TOPs, supports INT4/INT8/INT16/FP16
RAM: 64-bit 4GB/8GB/16GB LPDDR4X at 2133MHz
Flash: 32GB/64GB/256GB eMMC, at HS400 mode
microSD: support up to SDR104 mode
Ethernet: 2x PCIe 2.5G Ethernet

Signed-off-by: xiaobo tian <peterwillcn@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16158
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

mt76: Add firmware files into mt7925-firmware package

The firmware binaries were missing in kmod-mt7925-firmware package.

Signed-off-by: Daniel Pawlik <pawlik.dan@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16239
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

linux-firmware: add Intel CPU-integrated GPU (iGPU) firmware

On latest Intel x86 CPUs, DMC firmware is required for the iGPU to reach
its lowest power states. If the driver cannot load it, it will print a
warning and unnecessarily make the iGPU draw a bit more power when idle.

GUC firmware (various "offload" mechanisms that deal with scheduling GPU
workloads) and HUC firmware (required for accelerated media codec
operations for HEVC/H.265) are probably more niche, but could also
provde useful for some - for example, when building an
Intel/OpenWrt-based security camera.

Signed-off-by: Johannes Truschnigg <johannes@truschnigg.info>
Link: https://github.com/openwrt/openwrt/pull/16069
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

ath79: update Sophos AP15 to indicate that it uses an QCA9557 SoC

Device support for Sophos AP15 is based on Sophos AP55(C) and AP100(C).
Those other Sophos access points uss a QCA9558 SoC (some of them with
one of the three chains on the built-in SoC's wifi disabled) while the
AP15 uses a QCA9557 SoC (which only has two chains enabled in the
package or silicon).

This is mostly cosmetic since QCA9558 and QCA9557 are virtually
identical and all differences are automatically detected and/or managed
by the ART calibration.

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Link: https://github.com/openwrt/openwrt/pull/16187
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

ath79: remove 5GHz wifi bits from Sophos AP15

Sophos AP15 only uses the SoC's built-in wifi. There's no external PCIe
chipset.

Signed-off-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Link: https://github.com/openwrt/openwrt/pull/16187
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

lantiq: vgv7510kw22: use NVMEM for mac addresses

Userspace handling is deprecated.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16230
Signed-off-by: Robert Marko <robimarko@gmail.com>

mpc85xx: Add QCA8327 LED nodes to tl-wdr4900-v1 dts

This commit introduces led nodes in tl-wdr4900-v1 dts.
It allows to configure switch leds from userspace.

Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16226
Signed-off-by: Robert Marko <robimarko@gmail.com>

mpc85xx: add usb trigger to tl-wdr4900-v1 leds

TP-Link tl-wdr4900-v1 has two usb leds. Modern kernels can handle usb
led trigger. Let's enable it.

Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16226
Signed-off-by: Robert Marko <robimarko@gmail.com>

mpc85xx: refresh tl-wdr4900-v1 led configuration

This commit change led description in dts file from old method to new
approach accepted by upstream.

USB power gpio is handled by gpio-export now.

Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16226
Signed-off-by: Robert Marko <robimarko@gmail.com>

tmon: make version apk compatible

No need to override version.

Signed-off-by: Rosen Penev <rosenp@gmail.com>

generic: 5.15, 6.1: enable CNP support

Ever since CONFIG_ARM64_PAN was enabled Common Not Private (CNP) is now
visible and kernel builds will stop as they are not set in kernel config
for 5.15 and 6.1.

So, lets enable Common Not Private (CNP) which is ARMv8.2 feature and will
be NOP of CPU-s that dont support it.

Fixes: a2662309aae1 ("kernel: Enable CONFIG_ARM64_PAN to restrict kernel access to user space memory")
Link: https://github.com/openwrt/openwrt/pull/16211
Signed-off-by: Robert Marko <robimarko@gmail.com>

generic: 6.6: disable EPAN support

Enhanced Privileged Access Never (EPAN) is an ARMv8.7 feature, and since
we dont have any SoC-s implementing it lets disable it.

Link: https://github.com/openwrt/openwrt/pull/16211
Signed-off-by: Robert Marko <robimarko@gmail.com>

mediatek/mt7622: fix config symbol for 2.5GE MTK driver

The commit backporting new MTK patches did not update the
Kernel config for mediatek/mt7622 causing the build to
fail.

Simply use the new config symbol name for the driver to
fix the issue.

Fixes: 1069514978 ("mediatek: backport pending Ethernet PHY driver patches")
Signed-off-by: Rany Hany <rany_hany@riseup.net>
Link: https://github.com/openwrt/openwrt/pull/16225
Signed-off-by: Robert Marko <robimarko@gmail.com>

mediatek: backport pending Ethernet PHY driver patches

Use pending patchset for 2.5GE PHY driver, unifying LED handling
accross all MediaTek Ethernet PHYs.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>

kernel: add missing config symbols

Kconfig symbols CONFIG_ARM64_CNP and CONFIG_ARM64_EPAN got exposed
by enabling CONFIG_ARM64_PAN. Enable them as well, as just like for
PAN, also EPAN and CNP will be detected at runtime at no cost.

Fixes: a2662309aa ("kernel: Enable CONFIG_ARM64_PAN to restrict kernel access to user space memory")
Signed-off-by: Daniel Golle <daniel@makrotopia.org>

mpc85xx: tl-wdr4900: move platform code to dts

No benefit in doing so in platform file. Code for dts has already been
written. Might as well use it.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16125
Signed-off-by: Robert Marko <robimarko@gmail.com>

mpc85xx: use NVMEM for wifi

Userspace handling is deprecated. No need for any of this.

Calibration size was adjusted to 440, which is the standard value for
ath9k radios.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16125
Signed-off-by: Robert Marko <robimarko@gmail.com>

mpc85xx: tl-wdr4900: use NVMEM for WAN

Now that the ports are defined in DTS, this is trivial to do.

Avoids userspace handling.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16125
Signed-off-by: Robert Marko <robimarko@gmail.com>

mt76: update to Git HEAD (2024-08-21)

5c9fbdd64313 wifi: mt7915: fix wcid allocation leak
d547c25cabab wifi: mt76: mt7915: fix last argument to mt7915_mcu_add_sta
caed4843c5cd wifi: mt76: fix station muar index
3cd1c1740280 wifi: mt76: mt7915: hold dev->mt76.mutex while disabling tx worker
a5e630ef458c wifi: mt76: mt7916: fix regression in .sta_state migration

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

x86: enable console keyboard

The kernel support necessary to use a console keyboard was not built on
x86, affecting real and virtual machines alike. The console keyboard
would function properly in GRUB, but would not work at all once Linux
booted. It appeared that the console was intended to work because
console video appeared on the display, including prompts to enter
failsafe or select the debug log level from the keyboard, and the prompt
to "Press Enter to activate this console", but there was no way to
provide input to it. All keystrokes were ignored.

This enables several kernel configuration options to enable HID and USB
HID support (CONFIG_HID, CONFIG_HID_SUPPORT, CONFIG_HID_GENERIC, and
CONFIG_USB_HID), making the keyboard functional. For alignment with
armsr, CONFIG_USB_HIDDEV is also added, although not strictly necessary
for keyboard support. Note that this change also causes
CONFIG_HID_HYPERV_MOUSE to be enabled for x86/64 and x86/generic: it was
already set in these subarchitectures' kernel configurations, but was
ineffective due to CONFIG_HID being absent.

The omission of keyboard support on x86 may not have been widely noticed
because USB HID is not used on production OpenWrt x86 machines such as
pc-engines,apu2 which only have a serial console, or with the default
x86 configuration used by scripts/qemustart, which uses -nographic and
does not configure a virtual physical console but instead uses a serial
console.

This configuration change results in, for x86_64, +40kB in kernel.bin
and just over +40kB in gzip-compressed "combined" images. This should
not be a problem for the non-storage-constrained x86 target.

Until 2a86425de107, CONFIG_HID, CONFIG_USB_HID, and CONFIG_USB_HIDDEV
were set in the target-level kernel configuration, and
CONFIG_HID_GENERIC was set at the subtarget level. These are
reintroduced strictly at the subtarget level by request. This applies to
the 64, generic, and legacy subtargets, omitting geode.

Fixes: https://github.com/openwrt/openwrt/issues/16157
Signed-off-by: Mark Mentovai <mark@mentovai.com>
Link: https://github.com/openwrt/openwrt/pull/16208
Signed-off-by: Robert Marko <robimarko@gmail.com>

kernel: Enable CONFIG_ARM64_PAN to restrict kernel access to user space memory

Enable the CONFIG_ARM64_PAN kernel security option, which leverages the
ARMv8.1 Privileged Access Never (PAN) extension to prevent the kernel
from directly accessing user space memory.

Instead, copy_to_user and similar functions must be used for data
transfer between kernel and user space. This feature is automatically
disabled at runtime on CPUs without PAN support, making it a no-op in
those cases.

Link: https://github.com/openwrt/openwrt/pull/16189
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel: Activate CONFIG_LIST_HARDENED

Activate the kernel option CONFIG_LIST_HARDENED for all targets.
This adds some inline checks to list_add() and list_del() operations
in the kernel. Before kernel 6.6 these checks were only available with
CONFIG_DEBUG_LIST option, but now a light version is available which
should only add very few extra instructions to such operations.

The performance penalty is very low from my point of view. It should
make it much harder to use bugs in Linux kernel list handling when
exploiting the Linux kernel.

Link: https://github.com/openwrt/openwrt/pull/16189
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

rockchip: add Radxa ROCK 5B support

Hardware
--------
RockChip RK3588 ARM64 (8 cores)
4/8/16/32GB LPDDR4X RAM
2500 Base-T
RGB LED
eMMC Connector
SPI-NOR 16MB
Micro-SD Slot
2x USB 2.0 Port
2x USB 3.0 Port
Headphone Jack
M.2 E-Key
M.2 M-Key
USB PD 5/9/12/15/20V Power

Install
--------
Uncompress the OpenWrt sysupgrade and write it to a micro SD card or
internal eMMC using dd.

Signed-off-by: Tianling Shen <cnsztl@immortalwrt.org>
Link: https://github.com/openwrt/openwrt/pull/16149
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

uboot-rockchip: add ROCK 5B support

Add support for the Radxa ROCK 5B board.

Signed-off-by: Tianling Shen <cnsztl@immortalwrt.org>
Link: https://github.com/openwrt/openwrt/pull/16149
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

rockchip: add Radxa ROCK 5A support

Hardware
--------
RockChip RK3588 ARM64 (8 cores)
4/8/16/32GB LPDDR4X RAM
1000 Base-T
Status LED
eMMC/SPI Connector
Micro-SD Slot
2x USB 3.0 Port
2x USB 2.0 Port
Headphone Jack
M.2 E-Key
USB PD/QC 5/9/12/15/20V Power

Install
--------
Uncompress the OpenWrt sysupgrade and write it to a micro SD card or
internal eMMC using dd.

Signed-off-by: Tianling Shen <cnsztl@immortalwrt.org>
Link: https://github.com/openwrt/openwrt/pull/16149
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

uboot-rockchip: add ROCK 5A support

Add support for the Radxa ROCK 5A board.

Signed-off-by: Tianling Shen <cnsztl@immortalwrt.org>
Link: https://github.com/openwrt/openwrt/pull/16149
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

rockchip: extend BOARDNAME

Now we have rk3588 support :)

Signed-off-by: Tianling Shen <cnsztl@immortalwrt.org>
Link: https://github.com/openwrt/openwrt/pull/16149
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

rockchip: update Kconfig for rk3588

Enable OTP/USBDP drivers for rk3588.

Signed-off-by: Tianling Shen <cnsztl@immortalwrt.org>
Link: https://github.com/openwrt/openwrt/pull/16149
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

rockchip: backport dts updates for rk3588

Backport upstreamed dts updates for rk3588.

Signed-off-by: Tianling Shen <cnsztl@immortalwrt.org>
Link: https://github.com/openwrt/openwrt/pull/16149
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

rockchip: backport driver updates for rk3588

Backport upstreamed clk/mfd/phy/usb updates for rk3588.

Signed-off-by: Tianling Shen <cnsztl@immortalwrt.org>
Link: https://github.com/openwrt/openwrt/pull/16149
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>