x86: generic: config-6.6: refresh config

Add CONFIG_ARCH_CONFIGURES_CPU_MITIGATIONS=y which was introduced
back in April of this year[1]

1. https://github.com/torvalds/linux/commit/ce0abef6a1d540acef85068e0e82bdf1fbeeb0e9

Signed-off-by: John Audia <therealgraysky@proton.me>
Link: https://github.com/openwrt/openwrt/pull/16535
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

ramips: modernize ralink i2c driver

Use more devm and remove _remove function.

Remove owner as the kernel sets it automatically.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16546
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel: ubnt-ledbar: return directly

Removes ret variable.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16508
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel: ubnt-ledbar: use dev_err_probe

Handles EPROBE_DEFER and simplifies the code.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16508
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

ramips: rename interfaces for tplink er605v2

Currently eth1 (which is the first "lan" interface) doesn't work on this device.
During boot the following can be seen in logs:

```
[    2.252804] mt7530-mdio mdio-bus:1f: MT7530 adapts as multi-chip module
[    2.266060] mtk_soc_eth 1e100000.ethernet eth0: mediatek frame engine at 0xbe100000, irq 19
[    2.277889] mtk_soc_eth 1e100000.ethernet eth1: mediatek frame engine at 0xbe100000, irq 19
...
[    2.355157] mt7530-mdio mdio-bus:1f: MT7530 adapts as multi-chip module
[    2.390312] mt7530-mdio mdio-bus:1f: configuring for fixed/rgmii link mode
[    2.398597] mt7530-mdio mdio-bus:1f: Link is Up - 1Gbps/Full - flow control rx/tx
[    2.403872] mt7530-mdio mdio-bus:1f eth1 (uninitialized): PHY [mt7530-0:01] driver [MediaTek MT7530 PHY] (irq=21)
[    2.416988] mtk_soc_eth 1e100000.ethernet eth0: error -17 registering interface eth1
[    2.426973] mt7530-mdio mdio-bus:1f eth2 (uninitialized): PHY [mt7530-0:02] driver [MediaTek MT7530 PHY] (irq=22)
[    2.440996] mt7530-mdio mdio-bus:1f eth3 (uninitialized): PHY [mt7530-0:03] driver [MediaTek MT7530 PHY] (irq=23)
[    2.454405] mt7530-mdio mdio-bus:1f eth4 (uninitialized): PHY [mt7530-0:04] driver [MediaTek MT7530 PHY] (irq=24)
[    2.467198] mtk_soc_eth 1e100000.ethernet eth0: entered promiscuous mode
[    2.474117] DSA: tree 0 setup
...
[    6.820998] mtk_soc_eth 1e100000.ethernet dsa: renamed from eth0
[    6.919904] mtk_soc_eth 1e100000.ethernet wan: renamed from eth1
```

So the problem seems to be the fact that built-in gmacs get default
names (eth0/eth1) and are renamed after switch ports are initialized. This means
that when switch port with name `eth1` is brought up this name is still used by
gmac1 causing switch port's init to fail.

This patch just renames the ports to avoid name collision.
Note: this will break existing configs for this device because it renames all
the ports. This should not be major problem because this device doesn't have a
proper OEM image and is only flashable with serial access, meaning there should
not be many users.

Signed-off-by: Nikolay Martynov <mar.kolya@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/15865
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

wifi-scripts: Fix typo in EHT320 name

All other places name it EHT320.

Fixes: b478b7b1f7b9 ("wifi-scripts: detect and configure EHT")
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel: packages: kmod-pstore: remove deprecated Kconfig options

The following options no longer exist in 6.6:

PSTORE_COMPRESS_DEFAULT
PSTORE_DEFLATE_COMPRESS
CONFIG_PSTORE_DEFLATE_COMPRESS_DEFAULT

https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/fs/pstore/Kconfig?h=v6.6

Signed-off-by: Manuel Fombuena <fombuena@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/16579
Signed-off-by: Robert Marko <robimarko@gmail.com>

mt76: mark mt7925 11BE capable

Build hostapd with 11BE support for
both mt7925e and mt7925u.

Signed-off-by: Marek Puzyniak <marek.puzyniak@holisticon.pl>
Link: https://github.com/openwrt/openwrt/pull/16571
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

mac80211: mark ath12k 80211BE capable

Build hostapd with 80211BE suport.

Signed-off-by: Janusz Dziedzic <janusz.dziedzic@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16571
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

generic: backport upstream r8169 patches

These patches have been added in linux v6.12 release.

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

realtek: add support for Linksys LGS310C

Hardware specification
----------------------

* RTL8380M SoC, 1 MIPS 4KEc core @ 500MHz
* 256MB DRAM
* 32MB NOR Flash
* 8 x 10/100/1000BASE-T ports
* 2 x SFP ports
* Power LED, Fault LED
* Reset button on front panel
* UART (115200 8N1) via populated standard pin header marked JP1

TODO: The SFP ports use a shared SCL GPIO that the driver cannot handle.
The left SFP port (lan9) is defined and fully functional while the laser
on the right SFP port (lan10) is off by default.

UART pinout
-----------

[o]ooo|JP1
 | ||`------ GND
 | |`------- RX
 | `-------- TX
 `---------- Vcc (3V3)

Installation using OEM webinterface
-----------------------------------

1. Make sure you are running OEM firmware in secondary slot
2. Install squashfs-factory.imag to primary slot by upload via http

Installation using serial interface
-----------------------------------

1. Press "a" "c" "p" during message "Enter correct key to stop autoboot"
2. Load image with "upgrade runtime <TFTP IP>:squashfs-sysupgrade.bin" command
3. Switch to primary slot with "setsys bootpartition 0"
4. Store config with "savesys"
5. Boot the image with `boota` command

Dual-boot with stock firmware using writable u-boot-env
-------------------------------------------------------

From stock to OpenWrt / primary image 1 (CLI as admin):
   - > boot system image1
   - > reboot

From OpenWrt to stock / boot image 2: (shell as root)
   - # fw_setsys bootpartition 1
   - # reboot

Debrick using serial interface
------------------------------

1. Press "a" "c" "p" during message "Enter correct key to stop autoboot"
2. Load vendor image with "upgrade runtime <TFTP IP>:LGS310xxxxx.imag"
3. switch to primary partition "setsys bootpartition 0"
4. safe config "savesys"

Further documentation
---------------------
See https://openwrt.org/toh/linksys/lgs352c

It has been developed and tested on device with v1 revision.

Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
Link: https://github.com/openwrt/openwrt/pull/16068
[Add missing 'w' in name of firmware partition]
Signed-off-by: Sander Vanheule <sander@svanheule.net>

wifi-scripts: add better defaults for 6 GHz interop

Indicate stationary AP and set default 6G regulatory power type to indoor

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

base-files: set root password if present inside board.json

Add code to set plain password or put the hash into /etc/shadow.

Signed-off-by: John Crispin <john@phrozen.org>

base-files: add timezone to UCI during config_generate

Signed-off-by: John Crispin <john@phrozen.org>

dropbear: add a uci-defaults script for loading authorized keys

Write the ssh authorized key to /etc/dropbear/ssh_authorized_keys if present
inside boad.json.

Signed-off-by: John Crispin <john@phrozen.org>

wifi-scripts: populate default wifi credentials and country code from board.json

Signed-off-by: John Crispin <john@phrozen.org>

uboot-envtools: add a board.d script to load defaults from the environment

This allows loading default wifi credentials, passwords and ssh keys on
firstboot.

Signed-off-by: John Crispin <john@phrozen.org>

uboot-envtools: add fw_loadenv tool

This tool will load the uboot environment to /var/run/uboot-env/. This allows
more efficient use when accessing multiple variables.

Signed-off-by: John Crispin <john@phrozen.org>

base-files: execute uboot-env script before calling config_generate

This allows /etc/board.d/* scripts to use values from the uboot environment.

Signed-off-by: John Crispin <john@phrozen.org>

base-files: uci-defaults: allow setting the number of MACs a radio can use

Introduce new uci-default functions:
- ucidef_set_wireless_mac_count [count]

Signed-off-by: John Crispin <john@phrozen.org>

base-files: uci-defaults: allow setting default timezone

Introduce new uci-default functions:
- ucidef_set_timezone TZ

Signed-off-by: John Crispin <john@phrozen.org>

base-files: uci-defaults: allow setting default credentials and ssh keys

Introduce new uci-default functions:
     - ucidef_set_root_password_hash [hash]
     - ucidef_set_root_password_plain [plaintext]
     - ucidef_set_ssh_authorized_key [ssh key]

Signed-off-by: John Crispin <john@phrozen.org>

base-files: uci-defaults: allow setting wireless defaults

Introduce new uci-default functions:
 - ucidef_set_wireless band ssid [encryption] [key]
 - ucidef_set_country cc

They are supposed to be used in /etc/board.d/* scripts to define
board-specific defaults for wireless.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Signed-off-by: John Crispin <john@phrozen.org>

hostapd: send a notification via ubus when CSA completed

Signed-off-by: John Crispin <john@phrozen.org>

hostapd: add ifname to generic ubus notify code

Signed-off-by: John Crispin <john@phrozen.org>

hostapd: add ifname and vlan_id to sta-authorized notifications

Signed-off-by: John Crispin <john@phrozen.org>

hostapd: add the ifname to ubus events

Signed-off-by: John Crispin <john@phrozen.org>

umdns: update to latest HEAD

fbaca4b cache: improve update call by doing a full refresh probe
93c9036 dns: reply to A/AAAA questions for additional hostnames

Signed-off-by: John Crispin <john@phrozen.org>

wifi-scripts: detect and configure EHT

Check if EHT/11BE supported, configure in board.json
and config/wireless.

Signed-off-by: Janusz Dziedzic <janusz.dziedzic@gmail.com>

mt76: mark mt7996e 11BE capable

Build hostapd with 11BE support.

Signed-off-by: Janusz Dziedzic <janusz.dziedzic@gmail.com>

wifi-scripts: mac80211.sh: enable EHT320

For EHT320 we should also enable HE160.

Signed-off-by: Janusz Dziedzic <janusz.dziedzic@gmail.com>

wifi-scripts: mac80211.sh: allow EHT* sets

Allow to configure 80211BE.

Signed-off-by: Janusz Dziedzic <janusz.dziedzic@gmail.com>

hostapd: fix build when 80211BE enabled

In file included from hostapd-wpad-basic-mbedtls/hostapd-2024.03.09~695277a5/src/ap/ubus.h:11,
                 from hostapd-wpad-basic-mbedtls/hostapd-2024.03.09~695277a5/src/ap/hostapd.h:21,
                 from main.c:26:
hostapd-2024.03.09~695277a5/src/ap/sta_info.h: In function 'ap_sta_is_mld':
hostapd-2024.03.09~695277a5/src/ap/sta_info.h:425:20: error: invalid use of undefined type 'struct hostapd_data'
  425 |         return hapd->conf->mld_ap && sta && sta->mld_info.mld_sta;
      |                    ^~

Signed-off-by: Janusz Dziedzic <janusz.dziedzic@gmail.com>

hostapd: add CONFIG_DRIVER_11BE_SUPPORT

Add option to enable 802.11BE support.

Signed-off-by: Janusz Dziedzic <janusz.dziedzic@gmail.com>

ucode: update to latest HEAD

b610860 Merge pull request #223 from jow-/nl80211-add-eht-attributes
4e0456b nl80211: add EHT mac/phy capabilities
c938ee9 Merge pull request #218 from jonasjelonek/add-ioctl
56ebeec fs: add ioctl() file method
d0f423b Merge pull request #221 from mikma/fix-getenv
86f1121 lib: test if call to getenv() destroys environ
7af80d9 lib: use copy of environ pointer in getenv()
6ea37c8 ci: bump upload-artifact to v3
26d7292 Merge pull request #214 from jow-/ucode-test-runner
fbabec4 tests: replace test runner shell script with ucode implementation
e391ef5 main: prevent invalid memory access when executing empty stdin
b391fd7 Merge pull request #211 from jow-/hwsim-support
bc13278 nl80211: support the MAC80211_HWSIM netlink protocol family
f5b5a58 nl80211: drop unused structure members from connection context
f6ea6fc nl80211: support conversion from/to struct array attributes
115a84f nl80211: gracefully handle illegal netlink error code

Signed-off-by: John Crispin <john@phrozen.org>

iwinfo: update to latest HEAD

714e419 iwinfo: fix EHT mode reporting for STA interfaces
7eed433 devices: add device id for MediaTek MT7996e

Signed-off-by: John Crispin <john@phrozen.org>

mediatek: change the the default SUBTARGET to filogic

Signed-off-by: John Crispin <john@phrozen.org>

mac80211: do not pass a stopped vif to the driver on get_txpower

Fixes potential crashes in various drivers when checking the status of
interfaces that are down

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

kernel: bump 6.6 to 6.6.53

Changelog: https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.6.53

All patches automatically rebased.

Build system: x86/64
Build-tested: x86/64/AMD Cezanne, flogic/glinet_gl-mt6000, flogic/xiaomi_redmi-router-ax6000-ubootmod
Run-tested: x86/64/AMD Cezanne, flogic/glinet_gl-mt6000, flogic/xiaomi_redmi-router-ax6000-ubootmod

Signed-off-by: John Audia <therealgraysky@proton.me>
Link: https://github.com/openwrt/openwrt/pull/16554
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

lantiq: backport latest upstream patches

Backport patch that fixes memory disclosure in packet padding.
The downstream driver supports statistics, so when a packet
cannot be padded the statistics of dropped packets are incremented.
The other patches do not introduce any functional changes.

Signed-off-by: Aleksander Jan Bajkowski <olek2@wp.pl>
Link: https://github.com/openwrt/openwrt/pull/16563
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

bmips: fix kernel panic caused by missing CBR address

The cbr-reg DTS property uses a wrong name causing the RAC kernel
panic again on BCM6358 BCM6368 boards.

Use the correct cbr-reg name property.

Fixes: 7c9644a7b5 ("bmips: backport upstreamed RAC patches")
Signed-off-by: Daniel González Cabanelas <dgcbueu@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16561
Signed-off-by: Robert Marko <robimarko@gmail.com>

wifi-scripts: wifi-detect.uc: skip null entries on nl80211 phy dump

This can happen on incomplete device probe

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

generic: backport ("net: free_netdev: exit earlier if dummy")

Currently, when removing ath11k or any other driver that uses dummy netdev
kernel will crash with:
[  365.004961] ------------[ cut here ]------------
[  365.004992] kernel BUG at net/core/dev.c:10979!
[  365.008642] Internal error: Oops - BUG: 00000000f2000800 [#1] SMP
[  365.012898] Modules linked in: pppoe ppp_async nft_fib_inet nf_flow_table_inet ath11k_ahb(O) ath11k(O) pppox ppp_generic nft_reject_ipv6 nft_reject_ipv4 nft_reject_inet nft_reject nft_redir nft_quota nft_numgen nft_nat nft_masq nft_log nft_limit nft_hash nft_flow_offload nft_fib_c
[  365.064794] CPU: 3 PID: 3896 Comm: procd Tainted: G           O       6.6.52 #0
[  365.087031] Hardware name: QNAP 301w (DT)
[  365.094058] pstate: 20400005 (nzCv daif +PAN -UAO -TCO -DIT -SSBS BTYPE=--)
[  365.098229] pc : free_netdev+0x164/0x1a0
[  365.104994] lr : free_netdev+0xec/0x1a0
[  365.109159] sp : ffffffc081d33b90
[  365.112718] x29: ffffffc081d33b90 x28: ffffff80039d3c00 x27: ffffff800307f000
[  365.116199] x26: ffffff800538a000 x25: ffffff8005388000 x24: ffffff800538a4c0
[  365.123317] x23: 00000000000023a4 x22: ffffff8005388f68 x21: ffffff8004e37050
[  365.130434] x20: ffffff8004e37000 x19: ffffff8004e36ee8 x18: 0000000000000152
[  365.137552] x17: 63775f357636712e x16: 3030303030646320 x15: ffffffc081436e20
[  365.144670] x14: ffffff80033f261c x13: 00000000000000d9 x12: 0000000000000002
[  365.151789] x11: 0000000000000000 x10: 00000000000008a0 x9 : ffffffc081d33980
[  365.158906] x8 : ffffff80039d4500 x7 : ffffff803fdda6c0 x6 : 0000000000000007
[  365.166025] x5 : 0000000000000000 x4 : 00000000000000c1 x3 : 00000000000001f4
[  365.173143] x2 : ffffff803fdd4b78 x1 : ffffff8004e37050 x0 : 0000000000000005
[  365.180261] Call trace:
[  365.187370]  free_netdev+0x164/0x1a0
[  365.189630]  0xffffffc079b373f0
[  365.193447]  0xffffffc079b374c0
[  365.196311]  platform_shutdown+0x24/0x34
[  365.199438]  device_shutdown+0x160/0x268
[  365.203605]  kernel_restart+0x40/0xc0
[  365.207510]  __do_sys_reboot+0x104/0x220
[  365.211070]  __arm64_sys_reboot+0x24/0x30
[  365.215063]  invoke_syscall.constprop.0+0x5c/0x108
[  365.218971]  do_el0_svc+0x40/0xc8
[  365.223655]  el0_svc+0x30/0xb8
[  365.227041]  el0t_64_sync_handler+0x120/0x12c
[  365.229995]  el0t_64_sync+0x178/0x17c
[  365.234424] Code: f94013f5 a8c37bfd d50323bf d65f03c0 (d4210000)
[  365.238072] ---[ end trace 0000000000000000 ]---

Wireless backports include alloc_netdev_dummy() backport but they dont have
the required free_netdev change first, so backport it.

Fixes: #16531
Fixes: 384d079fd876 ("mac80211: update to version 6.11")
Link: https://github.com/openwrt/openwrt/pull/16549
Signed-off-by: Robert Marko <robimarko@gmail.com>

mt76: add missing eeprom for mt7992 firmware

Avoid the following errors:
[    9.219272] mt7996e 0000:01:00.0: Direct firmware load for mediatek/mt7996/mt7992_eeprom_2i5i.bin failed with error -2
[    9.229975] mt7996e 0000:01:00.0: Falling back to sysfs fallback for: mediatek/mt7996/mt7992_eeprom_2i5i.bin

Fixes: 2f7d22d ("mt76: update to Git HEAD (2024-09-29)")
Signed-off-by: Chukun Pan <amadeus@jmu.edu.cn>

mac80211: reorder patches in subsys/

Close some patch numbering gaps left behind by updates

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

tools/elfutils: prevent invalid include of stack binary

In some versions of Xcode, some C++ header has an include
of the standard C++ header "stack" while "." is in the include paths
which can conflict with the binary "stack" built by elfutils.
This leads to a decode error as the binary is interpreted as text.

Add an arbitrary dependency between stack and the C++ program.

Signed-off-by: Michael Pratt <mcpratt@pm.me>

mediatek: filogic: select MT7996 2+3+3 firmware

Select MT7996 2+3+3 firmware for the Adtran SmartRG SDG-8733A device
and also add it to the image for the BPi-R4 which supports both,
BE19000 (4+5+5) as well as BE14000 (2+3+3) Wi-Fi 7 modules.

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

mt76: remove leftover api update patch

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

mt76: fix typo in last update

Fix broken dependency for mt7992 firmware

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

mt76: update to Git HEAD (2024-09-29)

1231d45dcb5e wifi: mt76: mt7915: check devm_kasprintf() returned value
042082027b3f wifi: mt76: mt7921: Check devm_kasprintf() returned value
4a95c1a81eee wifi: mt76: Avoid multiple -Wflex-array-member-not-at-end warnings
5fdb1841f726 wifi: mt76: mt7925: fix a potential association failure upon resuming
eeb51b527377 wifi: mt76: mt7925: convert comma to semicolon
d4b3b302ceb9 wifi: mt76: mt7615: check devm_kasprintf() returned value
922e515bd8e0 wifi: mt76: mt7925: fix a potential array-index-out-of-bounds issue for clc
bec8cb851d99 wifi: mt76: mt7925: replace chan config with extend txpower config for clc
d10ce9d65dac wifi: mt76: remove mt76_calculate_default_rate()
6685a2f4c217 wifi: mt76: mt7996: remove phy->monitor_vif
a2a9d9826277 wifi: mt76: do not increase mcu skb refcount if retry is not supported
059342ca6a68 wifi: mt76: move mt76_tm_policy to from testmode.h to mt76.h
8bef11443290 wifi: mac80211: inform the low level if drv_stop() is a suspend
4b16d2bfdcbd wifi: mac80211: handle color change per link
4529cddb6718 wifi: mt76: mt7996: extend flexibility of mt7996_mcu_get_eeprom()
4364571e4b2b wifi: mt76: mt7996: add support for more variants
8c86aa304ea1 wifi: mt76: mt7996: set correct background radar capability
addfbd141863 firmware: add latest mt7996 + mt7992 firmware
5ba4981f8b92 wifi: mt76: mt792x: add P2P_DEVICE support
dc32780c7fcd wifi: mt76: mt7921s: fix a potential firmware freeze during startup
7b5d653ea90e wifi: mt76: mt7925: fix off by one in mt7925_load_clc()
1e57d2da94c9 wifi: mt76: mt7615: Convert comma to semicolon
7aca94abac47 wifi: mt76: mt7915: fix slot time for 5/6GHz
e3e6d490ab90 wifi: mt76: mt7915: Fix mesh scan on MT7916 DBDC
ee693260c521 firmware: move mt7996 firmware to the right place
867accf4f4aa wifi: mt76: mt7996: Add eht radiotap tlv
7c4d6e403381 wifi: mt76: Fix EHT NSS radiotap reporting.
71c3949fa8ac wifi: mt76: mt7921: fix a potential scan no APs
61641da04c9e mt76: mt76u_vendor_request: Do not print error messages when -EPROTO
680bc70f161f wifi: mt76: mt7921: fix using incorrect group cipher after disconnection.

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

yafut: fix build on non-linux systems

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

ramips: mtk_eth_soc: don't allocate 4 queues

This driver does not support more than 1 queue.

Fixes: ba24b94e0b8 ("ramips: use more devm in mtk_eth_soc")
Signed-off-by: Rosen Penev <rosenp@gmail.com>

scripts: config.guess: update to 2024-07-27

This script hasn't seen an update in multiple years, update it to the
latest version provided upstream.

Both `config.guess` and `config.sub` are copied from upstream[1] and not
modified.

The full changelog is available within the upstream repository[1].

[1]: https://git.savannah.gnu.org/git/config.git

Link: https://github.com/openwrt/openwrt/pull/16536
Signed-off-by: Robert Marko <robimarko@gmail.com>

kernel: make 'no port node found' output a debug message

There are cases where an unavailable port is not an error, making
this error message a false-positive. The kernel log is flooded with
the messages like:
  OF: graph: no port node found in /soc@0/bus@42000000/i2c@42530000/usb-typec@50

Silence this message by making it a debug message.

Signed-off-by: Tianling Shen <cnsztl@immortalwrt.org>
Link: https://github.com/openwrt/openwrt/pull/16524
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

CI: labeler: Add new targets

Some new targets were added, but we did not add them to the labeler.
Add them now.

Link: https://github.com/openwrt/openwrt/pull/16454
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

CI: labeler: remove bcm63xx

Some packages are now used by other targets, move them over

Fixes: 5cd8e037b506 ("bcm63xx: drop target")
Link: https://github.com/openwrt/openwrt/pull/16454
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

tools: fakeroot: update to 1.36

The type 'id_t' check has been moved to configure.ac in upstream commit:
819063eb4d27 ("test for id_t with autoconf instead of blind typedef, fixes FTBFS on FreeBSD")

changelog:
https://salsa.debian.org/clint/fakeroot/-/blob/e4780b1aad6fae0be1868a76cc7ffa6149afa912/debian/changelog

Signed-off-by: Shiji Yang <yangshiji66@qq.com>
Link: https://github.com/openwrt/openwrt/pull/16530
Signed-off-by: Nick Hainke <vincent@systemli.org>

base-files: emmc: invalidate kernel image while rootfs is being written

Some setups have alternate boot actions in case the main OS fails to
boot. These can include a secondary copy of the OS, a recovery OS, a
fallback to TFTP boot, etc.

This commit invalidates the kernel image while rootfs is being written
which, if a sysupgrade is interrupted, will trigger an alternate boot
action in devices that support it. This results in safer sysupgrades.

Signed-off-by: Rodrigo Balerdi <lanchon@gmail.com>

mac80211: remove outdated patches

The only supported kernel in the main branch now is version 6.6. It's
time to clean up these useless codes and patches.

Signed-off-by: Shiji Yang <yangshiji66@qq.com>

mac80211: fix build error on kernel 6.6

The virtio_find_vqs() ABI has been changed since the 6.11 kernel.
Switch back to using the old ABI to fix the build error:

../mac80211-regular/backports-6.11/drivers/net/wireless/virtual/mac80211_hwsim.c: In function 'init_vqs':
../mac80211-regular/backports-6.11/drivers/net/wireless/virtual/mac80211_hwsim.c:6632:31: error: array type has incomplete element type 'struct virtqueue_info'
 6632 |         struct virtqueue_info vqs_info[HWSIM_NUM_VQS] = {
      |                               ^~~~~~~~
../mac80211-regular/backports-6.11/drivers/net/wireless/virtual/mac80211_hwsim.c:6637:16: error: too few arguments to function 'virtio_find_vqs'
 6637 |         return virtio_find_vqs(vdev, HWSIM_NUM_VQS,
      |                ^~~~~~~~~~~~~~~
In file included from ../mac80211-regular/backports-6.11/drivers/net/wireless/virtual/mac80211_hwsim.c:38:
./include/linux/virtio_config.h:224:5: note: declared here
  224 | int virtio_find_vqs(struct virtio_device *vdev, unsigned nvqs,
      |     ^~~~~~~~~~~~~~~
../mac80211-regular/backports-6.11/drivers/net/wireless/virtual/mac80211_hwsim.c:6632:31: error: unused variable 'vqs_info' [-Werror=unused-variable]
 6632 |         struct virtqueue_info vqs_info[HWSIM_NUM_VQS] = {
      |                               ^~~~~~~~
../mac80211-regular/backports-6.11/drivers/net/wireless/virtual/mac80211_hwsim.c:6639:1: error: control reaches end of non-void function [-Werror=return-type]
 6639 | }
      | ^

Signed-off-by: Shiji Yang <yangshiji66@qq.com>

ath79: Refresh kernel patches

Make the patches apply cleanly again.

Fixes: 774badd8a840 ("kernel: fix crashes after linearizing fraglist GSO skbs")
Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/16517
Signed-off-by: Robert Marko <robimarko@gmail.com>

mediatek: remove "Ralink" from the target name

None of the supported chips have been sold under the Ralink brand.

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

tools: e2fsprogs: bypass libuuid check

The check is the wrong way around. It uses AC_CHECK_LIBS and then
pkg-config, probably because the latter is a second class citizen.

Just bypass it. pkg-config gets used anyway.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16480
Link: https://github.com/openwrt/openwrt/pull/16480
Link: https://github.com/openwrt/openwrt/pull/16480
Signed-off-by: Robert Marko <robimarko@gmail.com>

tools: util-linux: fix static libuuid pkgconf

Fixes uuid.pc usage on hosts with an older glibc.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16480
Link: https://github.com/openwrt/openwrt/pull/16480
Link: https://github.com/openwrt/openwrt/pull/16480
Signed-off-by: Robert Marko <robimarko@gmail.com>

kernel: fix crashes after linearizing fraglist GSO skbs

This fixes some corner cases triggered by enabling fraglist GRO, where some
protocols may accidentally or intentionally linearize fraglist skbs.
Previously, these skbs became unusable and segmenting them led to crashes.
With this patch, they are properly handled by passing them to skb_segment
instead of skb_segment_list.

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

mac80211: update to version 6.11

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

kernel: backport MHI patch required by an upcoming mac80211 update

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

openssl: update download mirrors

New releases of openssl are only published on GitHub, and official
downloads are also redirected to GitHub. So remove the old download
mirrors (file 404), and replace the current address with https.

Link: https://openssl-library.org/source/
Signed-off-by: Chukun Pan <amadeus@jmu.edu.cn>
Link: https://github.com/openwrt/openwrt/pull/16470
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

ramips: use more devm in mtk_eth_soc

These got added to the kernel after this driver was written.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16415
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

ath79: gpio-latch-mikrotik: remove fwnode

Not needed. Can just use dev.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16350
Signed-off-by: Robert Marko <robimarko@gmail.com>

ath79: gpio-latch-mikrotik: remove _remove

Not needed with devm.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16350
Signed-off-by: Robert Marko <robimarko@gmail.com>

ath79: gpio-latch-mikrotik: use dev_err_probe

It automatically adds the error code at the end of the message. It also
deals with -EPROBE_DEFER automatically (doesn't output). Simpler code.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16350
Signed-off-by: Robert Marko <robimarko@gmail.com>

uboot-kirkwood: fix compilation with GCC14

Missing header.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16467
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

tools: e2fsprogs: drop e2fsck patch

After digging in the history of the e2fsck exit code patch, it seems like
there is no reason for us to carry this patch so lets drop it.

Link: https://github.com/openwrt/openwrt/pull/16500
Signed-off-by: Robert Marko <robimarko@gmail.com>

tools: e2fsprogs: pass --with-crond-dir=no instead of patching

Currently, e2fsprogs is being patched to avoid detecting host crond, but
instead of doing that we can simply pass --with-crond-dir=no as an
argument and drop the patch.

Link: https://github.com/openwrt/openwrt/pull/16500
Signed-off-by: Robert Marko <robimarko@gmail.com>

tools: e2fsprogs: drop the e4defrag patch

Considering that this patch to not build e4defrag was added 12 years ago
because:
Compilation fails on older systems due to missing syscalls

I think its safe to say that enough time is passed so that whatever those
older systems were are either updated or not used anymore so lets drop
the patch.

Link: https://github.com/openwrt/openwrt/pull/16500
Signed-off-by: Robert Marko <robimarko@gmail.com>

realtek: d-link: add support for dgs-1210-28p-f

General hardware info:
----------------------

D-Link DGS-1210-28P rev. F1 is a switch with 24 ethernet ports and 4
combo ports, all ports Gbit capable. It is based on a RTL8382 SoC
@500MHz, DRAM 128MB and 32MB flash. 24 ethernet ports are 802.3af/at PoE
capable with a total PoE power budget of 193W.

Power over Ethernet:
--------------------

The PSE hardware consists of three BCM59121 PSE chips, serving 8 ports
each. They are controlled by a Nuvoton MCU.  In order to enable PoE, the
realtek-poe package is required. It is installed by default, but
currently it requires the manual editing of /etc/config/poe. Keep in
mind that the port number assignment does not match on this switch,
alway 8 ports are in reversed order: 8-1, 16-9 and 24-17.

LEDs and Buttons:
-----------------

On stock firmware, the mode button is supposed to switch the LED
indicators of all port LEDs between Link Activity and PoE status. The
currently selected mode is visualized using the respective LEDs. PoE Max
indicates that the maximum PoE budget has been reached.  Since there is
currently no support for this behavior, these LEDs and the mode button
can be used independently.

Serial connection:
------------------
The UART for the SoC (115200 8N1) is available via unpopulated standard
0.1" pin header marked J6. Pin1 is marked with arrow and square.

Pin 1: Vcc 3.3V
Pin 2: Tx
Pin 3: Rx
Pin 4: Gnd

OEM installation from Web Interface:
------------------------------------

    1. Make sure you are booting using OEM in image 2 slot. If not,
       switch to
        image2 using the menus
        System > Firmware Information > Boot from image2
        Tools > reboot
    2. Upload image in vendor firmware via Tools > Backup / Upgrade
        Firmware > image1
    3. Toggle startup image via System > Firmware Information > Boot
       from
        image1
    4. Tools > reboot

Other installation methods not tested, but since the device shares the
board with the DGS-1210-28, the following should work:

Boot initramfs image from U-Boot:
---------------------------------

    1. Press Escape key during `Hit Esc key to stop autoboot` prompt
    2. Press CTRL+C keys to get into real U-Boot prompt
    3. Init network with `rtk network on` command
    4. Load image with `tftpboot 0x8f000000
        openwrt-rtl838x-generic-d-link_dgs-1210-28p-f-initramfs-kernel.bin`
        command
    5. Boot the image with `bootm` command

Signed-off-by: Luiz Angelo Daros de Luca <luizluca@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/15938
Signed-off-by: Sander Vanheule <sander@svanheule.net>

tools/flex: respect STAGING_DIR_HOST

flex currently leaks the path of m4 as found on the build host.
While it is possible to override this using the M4 environment
variable (which we always did for autotools based builds) when
using CMake or Ninja the M4 variable is not set.
One easy fix is to make flex take STAGING_DIR_HOST into account
and expect m4 there if that variable is set in the environment.

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

tfa-layerscape: fix fiptool's build

Platform specified fiptool files was moved before lf-6.6.23-2.0.0 bump.
But PLAT_FIPTOOL_HELPER_MK still pointed to old location.
This cause problems with ls-ddr-phy build.

This patch fix PLAT_FIPTOOL_HELPER_MK path.

Fixes: 0ec659bd2b7e ("tfa-layerscape: Bump to lf-6.6.23-2.0.0")
Signed-off-by: Pawel Dembicki <paweldembicki@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16472
Signed-off-by: Robert Marko <robimarko@gmail.com>

mac80211: ath12k: add PCI_SUPPORT dependency

Always depend on PCI and don't build on platforms
without PCI.

Signed-off-by: Janusz Dziedzic <janusz.dziedzic@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16475
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.
* adapt the labeler configuration to the new format

Follow-up to #16251 this was reverted with b870c16534c05ddc94149c6ff56976d8de8a353f
Link: https://github.com/openwrt/openwrt/pull/16451
Signed-off-by: Goetz Goerisch <ggoerisch@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16451
Signed-off-by: Robert Marko <robimarko@gmail.com>

kernel: r8126: add CONFLICT to rss variant

The rss variant should conflict with the default.

Signed-off-by: Chukun Pan <amadeus@jmu.edu.cn>
Link: https://github.com/openwrt/openwrt/pull/16460
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>

kernel: r8125: add CONFLICT to rss variant

The rss variant should conflict with the default.

Signed-off-by: Chukun Pan <amadeus@jmu.edu.cn>
Link: https://github.com/openwrt/openwrt/pull/16460
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>

mac80211: introduce EHT rate support in AQL airtime

Backport required by mt76 for decent throughput with EHT rates

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

ipq807x: sax1v1k: fix sysupgrade not touching rootfs_data

Before this commit, sysupgrade saved the settings backup data to
the area following rootfs even though the device has an actual
rootfs_data partition. The backup data was completely ignored on
the following boot, but since rootfs_data was not being cleared
during sysupgrade, the issue was not noticed earlier.

Note that this commit changes the filesystem of rootfs_data from
ext4 to F2FS (if OEM partitioning is used).

Signed-off-by: Rodrigo Balerdi <lanchon@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16463
Signed-off-by: Robert Marko <robimarko@gmail.com>

rockchip: add ArmSoM Sige7 support

This board is also as known as Bananapi BPi-M7.

Hardware
--------
RockChip RK3588 ARM64 (8 cores)
8/16/32GB LPDDR4/LPDDR4x RAM
2x 2500 Base-T (PCIe, rtl8125b)
2 LEDs (RED / GREEN)
16GB/32GB/64GB/128GB eMMC on-board
Micro-SD Slot
USB 2.0 Port
USB 3.0 Port
M.2 M-Key
40-Pin Header
USB PD 2.0 9/12/15V Power

Installation
------------
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/16462
Signed-off-by: Nick Hainke <vincent@systemli.org>

uboot-rockchip: add ArmSoM Sige7 support

Add support for the ArmSoM Sige7 board.

Signed-off-by: Tianling Shen <cnsztl@immortalwrt.org>
Link: https://github.com/openwrt/openwrt/pull/16462
Signed-off-by: Nick Hainke <vincent@systemli.org>

rockchip: reorder nanopc t6 patches

Reorder nanopc t6 patches to 055 to avoid conflict with nanopi r6c patches.

Fixes: 55c46cbabd76 ("rockchip: add FriendlyElec NanoPi R6C")
Signed-off-by: Tianling Shen <cnsztl@immortalwrt.org>
Link: https://github.com/openwrt/openwrt/pull/16461
Signed-off-by: Nick Hainke <vincent@systemli.org>

generic: phy: aquantia: fix MDI pair property parsing

of_property_read_u32 returns -EINVAL when property does not exist,
according to the documentation -ENOENT is not a valid return code.

So, instead of checking for -ENOENT check for -EINVAL as otherwise the
blamed commit breaks AQR probe since it will return -EINVAL during probe.

Fixes: cb2a11f49c98 ("generic: phy: aquantia: add pending patch to force MDI pair order")
Link: https://github.com/openwrt/openwrt/pull/16466
Signed-off-by: Robert Marko <robimarko@gmail.com>

gpio-button-hotplug: use dev_err_probe

Avoids having to handle EPROBE_DEFER manually.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16456
Signed-off-by: Robert Marko <robimarko@gmail.com>

x86: generic: disable rust support

Add missing ksym CONFIG_RUST

Building r27563 without this commit will fail due to this ksym
being undefined in the generic config.

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

Signed-off-by: John Audia <therealgraysky@proton.me>
Link: https://github.com/openwrt/openwrt/pull/16464
Signed-off-by: Robert Marko <robimarko@gmail.com>

linux-firmware: Update to version 20240909

This updates the following firmware files:
amd64-microcode/lib/firmware/amd-ucode/microcode_amd_fam17h.bin
amd64-microcode/lib/firmware/amd-ucode/microcode_amd_fam19h.bin
amdgpu-firmware/lib/firmware/amdgpu/*
ar3k-firmware/lib/firmware/qca/hpnv21.bin
ar3k-firmware/lib/firmware/qca/hpnv21g.bin
ar3k-firmware/lib/firmware/qca/nvm_usb_00130201.bin
ar3k-firmware/lib/firmware/qca/nvm_usb_00130201_010a.bin
ar3k-firmware/lib/firmware/qca/nvm_usb_00130201_010b.bin
ar3k-firmware/lib/firmware/qca/nvm_usb_00130201_gf.bin
ar3k-firmware/lib/firmware/qca/nvm_usb_00130201_gf_010a.bin
ar3k-firmware/lib/firmware/qca/nvm_usb_00130201_gf_010b.bin
ar3k-firmware/lib/firmware/qca/rampatch_usb_00130201.bin
ath11k-firmware-wcn6855/lib/firmware/ath11k/WCN6855/hw2.0/amss.bin
ath11k-firmware-wcn6855/lib/firmware/ath11k/WCN6855/hw2.0/board-2.bin
ath11k-firmware-wcn6855/lib/firmware/ath11k/WCN6855/hw2.0/m3.bin
ath11k-firmware-wcn6855/lib/firmware/ath11k/WCN6855/hw2.1/amss.bin
ath11k-firmware-wcn6855/lib/firmware/ath11k/WCN6855/hw2.1/board-2.bin
ath11k-firmware-wcn6855/lib/firmware/ath11k/WCN6855/hw2.1/m3.bin
ath12k-firmware-wcn7850/lib/firmware/ath12k/WCN7850/hw2.0/board-2.bin
ibt-firmware/lib/firmware/intel/ibt-0040-0041.sfi
ibt-firmware/lib/firmware/intel/ibt-0040-4150.sfi
ibt-firmware/lib/firmware/intel/ibt-0041-0041.sfi
ibt-firmware/lib/firmware/intel/ibt-0180-0041.sfi
ibt-firmware/lib/firmware/intel/ibt-0180-1050.sfi
ibt-firmware/lib/firmware/intel/ibt-0180-4150.sfi
ibt-firmware/lib/firmware/intel/ibt-0291-0291.sfi
ibt-firmware/lib/firmware/intel/ibt-1040-0041.sfi
ibt-firmware/lib/firmware/intel/ibt-1040-1050.sfi
ibt-firmware/lib/firmware/intel/ibt-1040-4150.sfi
ibt-firmware/lib/firmware/intel/ibt-17-16-1.sfi
ibt-firmware/lib/firmware/intel/ibt-17-2.sfi
ibt-firmware/lib/firmware/intel/ibt-18-16-1.sfi
ibt-firmware/lib/firmware/intel/ibt-18-2.sfi
intel-igpu-firmware-guc/lib/firmware/i915/adlp_guc_70.bin
intel-igpu-firmware-guc/lib/firmware/i915/dg1_guc_70.bin
intel-igpu-firmware-guc/lib/firmware/i915/dg2_guc_70.bin
intel-igpu-firmware-guc/lib/firmware/i915/mtl_guc_70.bin
intel-igpu-firmware-guc/lib/firmware/i915/tgl_guc_70.bin
intel-igpu-firmware-huc/lib/firmware/i915/dg2_huc_gsc.bin
iwlwifi-firmware-ax200/lib/firmware/iwlwifi-cc-a0-77.ucode
iwlwifi-firmware-ax201/lib/firmware/iwlwifi-QuZ-a0-hr-b0-77.ucode
iwlwifi-firmware-ax210/lib/firmware/iwlwifi-ty-a0-gf-a0.pnvm
iwlwifi-firmware-be200/lib/firmware/iwlwifi-gl-c0-fm-c0.pnvm
mt7921bt-firmware/lib/firmware/mediatek/BT_RAM_CODE_MT7961_1_2_hdr.bin
mt7922bt-firmware/lib/firmware/mediatek/BT_RAM_CODE_MT7922_1_1_hdr.bin
r8169-firmware/lib/firmware/rtl_nic/rtl8126a-3.fw

Link: https://github.com/openwrt/openwrt/pull/16458
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

mac80211: ath12k: prevent ltssm crash

Fix ltssm crashes on BPI-Rx boards.
Seems read32/write32 using wrong address which
is not a problem on x86/64 PCI controllers.
But have issues on BPI-Rx boards.

Signed-off-by: Janusz Dziedzic <janusz.dziedzic@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/15945
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

linux-firmware: qualcomm: add WCN7850 firmware

Add option to add WCN7850 ath12k firmware.

Signed-off-by: Janusz Dziedzic <janusz.dziedzic@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/15945
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

mac80211: add ath12k PCI support

Add upstream ath12k 802.11BE driver support.
Allow to build and use it from openwrt.

Tested on BPI-R4/BPI-R64 with:
 - Qualcomm Technologies, Inc WCN785x Wi-Fi 7(802.11be) 320MHz 2x2 [FastConnect 7800] (rev 01)

Signed-off-by: Janusz Dziedzic <janusz.dziedzic@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/15945
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

mac80211: ath: add qcom-qmi-helpers as package

Allow to build ath11k and other drivers using
qcom-qmi-helpers together (eg. ath12k).

Signed-off-by: Janusz Dziedzic <janusz.dziedzic@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/15945
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

ath79: disable ath79 USB phy drivers by default

We already have an kernel module package "kmod-phy-ath79-usb" to
drive the ath79 USB. It will be selected by the usb base package
"kmod-usb2" and "kmod-usb-ohci".

Signed-off-by: Shiji Yang <yangshiji66@qq.com>
Link: https://github.com/openwrt/openwrt/pull/16380
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel: usb: improve support for ath79 target

1. Remove outdated symbol CONFIG_USB_OHCI_ATH79.
The ath79 OHCI USB was already supported by the generic driver
kmod-usb-ohci. And this kernel symbol has been dropped since
upstream commit:
53d473fc1e38 ("usb: host: Remove the deprecated ATH79 USB host config options")

2. Add ath79 USB phy package to the OHCI dependencies.
Both EHCI and OHCI require it on the ath79 target.

Signed-off-by: Shiji Yang <yangshiji66@qq.com>
Link: https://github.com/openwrt/openwrt/pull/16380
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

uboot-mediatek: support GL.iNet GL-X3000 and GL-XE3000

Add u-boot support based on the kernel dts introduced in d1016446 and
the GL-MT6000 u-boot support in fe10f974.

The pcie-mediatek-gen3 kernel driver doesn't like hotplug, so to work in
PCIe mode, the 5G modem on this device needs to be switched on by u-boot
before starting the kernel. Include an init_modem step in the boot_system
action to set the relevant gpios. (The factory bootloader does the same,
using Mediatek SDK-specific gpio_power_clr and gpio_pull_up.)

Ideally the modem would be started using gpio-hog in the device tree, but
this will need to wait until mediatek gpio-hog support is fixed upstream:

  https://lore.kernel.org/u-boot/6ef2583e85eea60560d7776377d662779e7c44e5.1722419839.git.chris@arachsys.com/

The bootloader can be replaced using the built-in web interface of the
factory bootloader. Hold the reset button for five seconds while powering
on the device and it will boot into a recovery http server.

http://192.168.1.1/uboot.html and http://192.168.1.1/bl2.html can then
be used to upload openwrt-mediatek-filogic-glinet_gl-x3000-bl31-uboot.fip
and openwrt-mediatek-filogic-glinet_gl-x3000-preloader.bin respectively.

Alternatively, from a root shell on the running system, unlock the boot
partition with

  echo 0 >/sys/block/mmcblk0boot0/force_ro

then write openwrt-mediatek-filogic-glinet_gl-x3000-bl31-uboot.fip to
/dev/mmcblk0p4 and openwrt-mediatek-filogic-glinet_gl-x3000-preloader.bin
to /dev/mmcblk0boot0.

Signed-off-by: Chris Webb <chris@arachsys.com>
Link: https://github.com/openwrt/openwrt/pull/15645
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

comgt: directip: add delegate option support

Ipv6 delegate option is not respected by proto directip
this add support for it.

Signed-off-by: Chen Minqiang <ptpt52@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/15508
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>