kernel: modules: appletalk: add for Netatalk v4

Netatalk v4 reintroduces AppleTalk and this module is required for
layer 3 protocol support.

Module was removed in kernel 4.14 for OpenWrt 18.06 (commmit
14a0131, 22/02/2018). At the time nothing used it as Netatalk v3
did not support AppleTalk.

Not building ipddp feature/module like it was in the past, as
recommended by upstream Netatalk maintainers.

Signed-off-by: Antonio Pastor <antonio.pastor@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16979
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 5eb25dddb17fa3cf4958e91dfc9fc868c9eb03ac)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

stm32: add new stm32 target

New stm32 target introduces support for stm32mp1 based devices.
For now it includes an initial support of the STM32MP135F-DK device.
The specifications bellow only list supported features.

Specifications
--------------

SOC: STM32MP135FAF7
RAM: 512 MiB
Storage: SD Card
Ethernet: 2x 100 Mbps
Wireless: 2.4GHz Cypress CYW43455 (802.11b/g/n)
LEDs: Heartbeat (Blue)
Buttons: 1x Reset, 1x User (USER2)
USB: 4x 2.0 Type-A

Signed-off-by: Thomas Richard <thomas.richard@bootlin.com>
Link: https://github.com/openwrt/openwrt/pull/16716
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 851e7f77e424f24cda8531330f2d82ebc942c0a4)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

kernel: package v4l2-async and v4l2-fwnode drivers

These packages are needed for the stm32/stm32mp1 subtarget.

Signed-off-by: Thomas Richard <thomas.richard@bootlin.com>
Link: https://github.com/openwrt/openwrt/pull/16716
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit e109831e07e398cc50839a30a1cf18c4436983f4)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

kernel: add kmod-media-controller dependency for kmod-usb-audio

The snd-usb-audio drivers uses media controller if media controller is set.
So add kmod-media-controller as dependency.

Signed-off-by: Thomas Richard <thomas.richard@bootlin.com>
Link: https://github.com/openwrt/openwrt/pull/16716
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 109592524663ae8bbda47e4ad51a124153a1c903)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

kernel: package media controller driver

It is needed for the stm32/stm32mp1 subtarget.

Signed-off-by: Thomas Richard <thomas.richard@bootlin.com>
Link: https://github.com/openwrt/openwrt/pull/16716
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit bb3418f21f7b8fe0dcd335375d183dd2a81772bf)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

kernel: package industrialio-buffer-cb driver

It is needed for the stm32/stm32mp1 subtarget.

Signed-off-by: Thomas Richard <thomas.richard@bootlin.com>
Link: https://github.com/openwrt/openwrt/pull/16716
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit b0adf5e0fdc0d4b449bced9191a95bf08124c1ac)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

kernel: package crypto_engine driver

It is needed for the stm32/stm32mp1 subtarget.

Signed-off-by: Thomas Richard <thomas.richard@bootlin.com>
Link: https://github.com/openwrt/openwrt/pull/16716
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 7bcc68ede36dcc66ec6cdb45c3201e0af0f90708)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

kernel: unset CONFIG_DEVFREQ_THERMAL in kmod-thermal package

It is needed for the stm32/stm32mp1 subtarget.

Signed-off-by: Thomas Richard <thomas.richard@bootlin.com>
Link: https://github.com/openwrt/openwrt/pull/16716
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 6b9df468f05fefa21a8cbce2fdb2e59692b4d246)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

kernel: package snd-soc-dmic driver

It is needed for the stm32/stm32mp1 subtarget.

Signed-off-by: Thomas Richard <thomas.richard@bootlin.com>
Link: https://github.com/openwrt/openwrt/pull/16716
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit e606366a2e018e0a3b25e3dd4b3a937bb53c70e1)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

kernel: package industrialio-hw-consumer driver

It is needed for the stm32/stm32mp1 subtarget.

Signed-off-by: Thomas Richard <thomas.richard@bootlin.com>
Link: https://github.com/openwrt/openwrt/pull/16716
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit ff96bd84ce0bae633007cf0de82d69796cf3508a)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

murata-nvram: add new package

This package contains nvram files for Murata's Wi-Fi/Bluetooth modules.

Signed-off-by: Thomas Richard <thomas.richard@bootlin.com>
Link: https://github.com/openwrt/openwrt/pull/16716
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit be937952fdf8f78fb45ea001258074ddb45e0891)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

murata-firmware: add new package

This package contains firmwares provided by Murata.
See https://community.murata.com/.

Murata firmwares are used by wifi chip on stm32 boards.

Signed-off-by: Thomas Richard <thomas.richard@bootlin.com>
Link: https://github.com/openwrt/openwrt/pull/16716
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 960f5f94d7cefe89abd41347124fd3f0fbb945ae)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

mac80211: enable CONFIG_BRCMFMAC_SDIO for stm32

All stm32 devices use BCM43430 Wi-Fi interface in SDIO mode.

Signed-off-by: Thomas Richard <thomas.richard@bootlin.com>
Link: https://github.com/openwrt/openwrt/pull/16716
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 456e0aad97413816aee41e68b3d3cfdfadc2fd43)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

arm-trusted-firmware-stm32: add ATF for stm32 boards

Add ATF for stm32 boards, with the first being STM32MP135F-DK.

Signed-off-by: Thomas Richard <thomas.richard@bootlin.com>
Link: https://github.com/openwrt/openwrt/pull/16716
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 08dcb73c8d05aa774957422731bfed1b322648d7)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

uboot-stm32: add U-Boot for stm32 boards

Add U-Boot for stm32 boards, with the first being STM32MP135F-DK.

Signed-off-by: Thomas Richard <thomas.richard@bootlin.com>
Link: https://github.com/openwrt/openwrt/pull/16716
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit ac79ad0757764028a138d0138c8d9aa74420f9ed)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

optee-os-stm32: add OP-TEE for stm32 boards

Add OP-TEE for stm32 boards, with the first being STM32MP135F-DK.

Signed-off-by: Thomas Richard <thomas.richard@bootlin.com>
Link: https://github.com/openwrt/openwrt/pull/16716
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit fae25d4001bcbae0bb1ce60b676bd421e97d78fc)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

build: add generic build template for optee-os packages

Add a generic template include/optee-os.mk to build OP-TEE.

Signed-off-by: Thomas Richard <thomas.richard@bootlin.com>
Link: https://github.com/openwrt/openwrt/pull/16716
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit dedb90e8f82a179ad67b6eac3b6f8862b9183077)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

mediatek: change Routerich AX3000 ubi size
All new routers are shipped with ubi size 112MB since early September.
Bootloader update required (ask vendor , see wiki)

These partitions weren't used:
firmware_backup
zrsave
config2

Signed-off-by: Romanov Danila <pervokur@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16686
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit d8a9669093359d60f4d6f183b58a22f24cf01731)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

x86: add all ports to the Gowin 1U default network config

Commit e52c57bb1b30375e0bcc5523db76a672a4a8b4a4 renamed all
network ports to match the faceplate of the Gowin 1U Rack
Mount Server and added the br-lan bridge for the eth* ports.

This commit adds the PoE port to the br-lan bridge and a
br-wan bridge for the two SFP ports so that all ports are
part of the default network configuration.

Signed-off-by: Til Kaiser <mail@tk154.de>
Link: https://github.com/openwrt/openwrt/pull/16965
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 75af6a0d736a696ca726cb44d813791858c29f0c)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

ramips: sync upstream Ralink clock patches

1. Add sdhc clock for MT7620 and MT76x8 SoCs.
2. Fix clock driver warning for RT2880, RT305x and RT3883.

Link: https://lore.kernel.org/all/20240910044024.120009-1-sergio.paracuellos@gmail.com/
Signed-off-by: Shiji Yang <yangshiji66@qq.com>
Link: https://github.com/openwrt/openwrt/pull/17037
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 7bb99bca3dfd878e6ad950b218c0cb96d36d14f4)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

Revert "mediatek: fix pwn fan settings for sinovoip bpi-r3"

This reverts commit 3467ea905b40f0df87700a15b716b1f5a4efc955.

I merged it by accident.

Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 52b6c9247997e51a97f13bb9e94749bc34e2d52e)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

mediatek: fix pwn fan settings for sinovoip bpi-r3

Popular bpi-r3 pwm fans like this one

  https://www.amazon.com/youyeetoo-Barebone-Fan-BPI-R3-Integrated/dp/B0CCCTY8PS

will not work properly with current openwrt-23.05/24.10 firmware.
Trying different pwm setting

  echo $value > /sys/devices/platform/pwm-fan/hwmon/hwmon1/pwm1

I found:

  pwm1 value     fan rotation speed   cpu temperature     notes
  -----------------------------------------------------------------
    0            maximal              31.5 Celsius        too noisy
   40            optimal              35.2 Celsius        no noise hearable
   95            minimal
   above 95      does not rotate      55.5 Celsius
  -----------------------------------------------------------------

At the moment we have following cooling levels:

  cooling-levels = <255 96 0>;

for cpu-active-high, cpu-active-medium and cpu-active-low modes correspondingly.
Thus only cpu-active-high and cpu-active-low are usable. I think this is wrong.

This patch fixes cpu-active-medium settings for bpi-r3 board.

PS: I know, the patch is not ideal as it can break pwm fan for some users.
    There are some peoples that use handmade cooling solutions, but:
      * discussed cooler is the only 'official' pwm cooler for bpi-r3
        available on the market.
      * most peoples will use passive cooling available on the market or
        the discussed cooler.
      * the pwm-fan dts section was added before the official cooler
        appears on the market.
    Thus it should not be a lot of harm from this fix.

Signed-off-by: Mikhail Kshevetskiy <mikhail.kshevetskiy@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16974
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 3467ea905b40f0df87700a15b716b1f5a4efc955)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

mpc85xx: p2020: add missing symbols

Since kernel_oldconfig doesn't work properly with it, I ran that first
and then moved all the config symbols to config-6.6 and found the
differences.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16847
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 22664498ebdd47aad9b15f62edbcaadd851fd62c)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

mpc85xx: p1020: add missing symbols

Since kernel_oldconfig doesn't work properly with it, I ran that first
and then moved all the config symbols to config-6.6 and found the
differences.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16847
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit bec8edb6d61eb74e8d7578749560cbb5040f307a)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

mpc85xx: p1010: add missing symbols

Since kernel_oldconfig doesn't work properly with it, I ran that first
and then moved all the config symbols to config-6.6 and found the
differences.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16847
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit e9dd6da91661612d2330c2b633f3dd55b93218f8)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

mpc85xx: run make kernel_oldconfig

Adjusts the default config to modern kernels.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16847
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 890d47b60a221328bbf05518eedc35ac5bd922f1)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

odhcp6c: update to Git HEAD (2024-09-25)

b6ae9ffaeb0e odhcp6c: allow to request specific ipv6 prefix

Signed-off-by: He Ping <pinghejk@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17013
Signed-off-by: Petr Štetiar <ynezz@true.cz> [hash, commit]
(cherry picked from commit b0312c10818f36bc6e61a518bb528d2809562796)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

omnia-eeprom: depend only on subtarget

Now that omnia-eeprom is marked nonshared building the cortex-a9 mvebu
subtarget will fail with:
ERROR: unable to select packages:
  omnia-eeprom (no such package):
    required by: world[omnia-eeprom]

This is because omnia-eeprom depends on TARGET_mvebu_cortexa9_DEVICE_cznic_turris-omnia
which will not be satisfied in buildbots since CONFIG_TARGET_ALL_PROFILES
and CONFIG_TARGET_PER_DEVICE_ROOTFS are set in which case
CONFIG_TARGET_mvebu_cortexa9_DEVICE_cznic_turris-omnia is not set.

So, lets simply depend on the mvebu/cortex-a9 subtarget.

Fixes: 371e7bef4046 ("omnia-eeprom: Mark it nonshared")
Link: https://github.com/openwrt/openwrt/pull/17007
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 90de3b277b8e1d121b2f7023578979c6b616167f)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

generic: add CONFIG_FB_INTEL

CONFIG_FB_INTEL is now visible on x86 since i915 driver is packaged as kmod
now and it stops compilation, so add it to the generic config.

Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 57daea682e1c6e6da91f92efed717159f9ffffb0)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

dropbear: use config_get_bool enable

The config_get_bool also works with on/off, yes/no, true/false.
Add 'main' section name. This will make it easier to change settings from uci.
Add a link to documentation.

Signed-off-by: Sergey Ponomarev <stokito@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/15579
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 4511fa4b30f73185597990ce563bbf3f96385292)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

kernel: filter out compiler opts from config

These get dynamically set based on compiler version. Not relevant for
targets.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16770
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit cd92cbddf8c473507a9af290c99f2aef8c368784)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

kernel: remove GCC11_NO_ARRAY_BOUNDS

This symbol is no longer present.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16770
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit da8abd4a1e86a37ddcd4fa0e6acc5bf8b1b4c70e)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

kernel: move CONFIG_INTEL_MEI settings to target generic

Move settings CONFIG_INTEL_MEI_GSC_PROXY, CONFIG_INTEL_MEI_HDCP
and CONFIG_INTEL_MEI_PXP to target generic.

Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/16971
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 84665993d6beed8ac2b557f8d3e005a29e3fd09c)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

omnia-eeprom: Mark it nonshared

This tool was build in the phase 2 build, there the
TARGET_mvebu_cortexa9_DEVICE_cznic_turris-omnia dependecy was probably
not meat. Mark it as non shared to build it together with the target
where this option is set.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Link: https://github.com/openwrt/openwrt/pull/16975
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 371e7bef4046fb3992ecfb92b5b741c22fa3ade6)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

kernel: disable drm-i915 module for x86/geode

Disable drm-i915 module for target x86/geode.

Fixes: 77cfe8f ("x86: make i915 as a kmod with required firmware")
Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/16977
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 226f74badd61fb1586dd3531835aeaa3bf6fe467)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

ipq807x: delete unused legacy eMMC sysupgrade code

All ipq807x devices that were using the legacy 'mmc_do_upgrade' eMMC
sysupgrade code were ported to the replacement 'emmc_do_upgrade' code.

Signed-off-by: Rodrigo Balerdi <lanchon@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16505
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 491121288eaa9d20f3db8215a4790e8f0ee1f8ca)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

ipq807x: order devices alphabetically

Alphabetically sort devices in platform.sh

Signed-off-by: Rodrigo Balerdi <lanchon@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16505
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 17f84bba44dbee248b3a79935657a74b21caa280)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

ipq807x: prpl haze & qnap 301w: use the standard eMMC sysupgrade code

Note that the old ad-hoc method did not explicitly align backup data
to 64 KiB boundaries.

Also note that the qnap 301w has a 'rootfs_data' partition in the eMMC
that is being ignored by fstools during boot, presumably due to a bug.
This is why the partition is also ignored in the sysupgrade code and
there is no definition of CI_DATAPART="rootfs_data".

Signed-off-by: Rodrigo Balerdi <lanchon@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16505
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit fe481c9c47d97ebde1bc3c959924f4f9227c6c16)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

ipq807x: nbg7815: use the standard eMMC sysupgrade code

Note that the old ad-hoc method did not explicitly align backup data
to 64 KiB boundaries.

Signed-off-by: Rodrigo Balerdi <lanchon@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16505
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 5583d2318966d5d5dcdbdec2e34b8815d3f6419a)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

elfutils: Backport some patches to fix errors

On aarch64 musl gcc 14.x compiler, trying compiling elfutils 0.192 with
lto option enabled will cause null-dereference error.
Example error message:

...
elf_compress.c: In function 'elf_compress':
elf_compress.c:675:26: error: potential null pointer dereference [-Werror=null-dereference]
  675 |           shdr->sh_flags |= SHF_COMPRESSED;
      |                          ^
elf_compress_gnu.c: In function 'elf_compress_gnu':
elf_compress_gnu.c:127:25: error: potential null pointer dereference [-Werror=null-dereference]
  127 |           shdr->sh_size = new_size;
      |                         ^                      ^
...

This is a false postive warning but will abort compilation if gcc has
`-Werror` flag. This commit add a patch for this, see the bugzilla
report below.

This commit backports a series of patches to fix some errors.

Add patch:
- 007-add-libeu-symbols-to-libelf.patch
- 008-fix-autoconf-ENABLE_IMA_VERIFICATION.patch
- 009-fix-null-dereference-with-lto.patch

Link: https://sourceware.org/bugzilla/show_bug.cgi?id=32311
Signed-off-by: Ryan Keane <the.ra2.ifv@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16886
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit afffcd09e5f15be53f327a80ee87d391312eb805)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

elfutils: Add mirrors.kernel.org as mirrror

Add mirrors.kernel.org as mirror, listed on sourceware mirror sites
page.

Link: https://sourceware.org/mirrors.html
Signed-off-by: Ryan Keane <the.ra2.ifv@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16886
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 63caa2b1685ef5d2b4259784085f64cb9e5574a8)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

elfutils: Update to 0.192

Add a patch still under review to fix some errors.

Refresh patch:
- 003-libintl-compatibility.patch
- 005-build_only_libs.patch
- 006-Fix-build-on-aarch64-musl.patch

Add patch:
- 102-fix-potential-deref-of-null-error.patch

Release notes are in the link below.

Link: https://inbox.sourceware.org/elfutils-devel/CAJDtP-T3+gXqHWp3T0mejWWbPr0_1tHetEXwfB67-o+zz7ShiA@mail.gmail.com/T/#u
Signed-off-by: Ryan Keane <the.ra2.ifv@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16886
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit f5132df69db50a99e556e6fbd165002aac0e3e38)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

realtek: align kernel config with upstream

Since the start of the Realtek target OpenWrt works with RTL83XX as the
target architecture. Upstream is using MACH_REALTEK_RTL instead. To
simplify further development align that.

Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
Link: https://github.com/openwrt/openwrt/pull/16963
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 65964c42f8744ed8ca09448a155b99e3a99be283)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

libreadlne: create symlinks for .so

Currently, libreadline only installs

```
 /usr/lib/libhistory.so.8 -> libhistory.so.8.2
 /usr/lib/libhistory.so.8.2
 /usr/lib/libreadline.so.8 -> libreadline.so.8.2
 /usr/lib/libreadline.so.8.2
```

But there is no `libreadline.so` or `libhistory.so` available.

So this happens:

```
root@OpenWRT:~# cat a.c
int main() {
}
root@OpenWRT:~# gcc a.c -lreadline
/usr/bin/ld: cannot find -lreadline: No such file or directory
collect2: error: ld returned 1 exit status
```

Unless, of course, one uses `-l:libreadline.so.8`... But that
doesn't help with binaries that try to dynamically open

`libreadline.so`. I have one of those here (the STklos Scheme
compiler -- I didn't make a PR for it because it's far from
being ready, but one issue is that it does use dlopen to use
readline...)

With the symlink, it works:

```
root@OpenWRT:~# ln -s /usr/lib/libreadline.so.8 /usr/lib/libreadline.so
root@OpenWRT:~#
root@OpenWRT:~# gcc a.c -lreadline
root@OpenWRT:~#
```

Another example: when trying to package rlwrap, the build failed
complaining it could not find readline (using `-lreadline`).
It would then be necessary to change rlwrap's `configure.ac`
(and also in all packages that use readline), but it seems
simpler to add the symlinks...

This PR changes the Makefile so it will include the links.

Signed-off-by: Jeronimo Pellegrini <j_p@aleph0.info>
Link: https://github.com/openwrt/openwrt/pull/16445
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit 0000ba6ab8062f965d8fed240b76d36f9a493b55)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

x86: switch config to a tickless kernel

This commit fixes and closes #16313.

Switch the x86 kernel's timer to tickless operation which is
more power efficient since it is not woken up by periodic timer
interrupts when idle.  Also add several other options for CPU
idle governors particularly the upstream default for tickless
kernels, CONFIG_CPU_IDLE_GOV_MENU. Without this commit, my AMD
Ryzen 7 5800U can only achieve a minimum core frequency of 1,384
MHz which is over 3x higher than the processor's minimum
frequency of 400 MHz which is accessible with this modification.
In addition to the lower clock rate, I have seen a concomitant
reduction in both idle temps and at-the-wall power consumption.

Summary:
* Idle CPU freqs dropped from 1,384 MHz to 400 Mhz.
* Idle power consumption dropped from 7 W avg to 5 W.
* Idle temps have dropped from 50C on avg to 43C.

There are other well known reasons to switch to a tickless
timer including: reduced interrupt overhead, better use of CPU
resources, and reduced latency to name a few.

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/16317
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit 31111680f6fc5f17201b02b8ce656b955be56d7a)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

ppp: remove more unnecessary kernel checks

The ppp package can support all features since Linux 4.7.0 kernel.
Therefore, most kernel version checks can pass unconditionally on
OpenWrt v18.06 and later version. This patch can reduce the size
of ppp package by approximately 2.5 KB.

Signed-off-by: Shiji Yang <yangshiji66@qq.com>
Link: https://github.com/openwrt/openwrt/pull/16695
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit 3dbe730080e70e96990a5e89cbe2035adcca8d4b)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

base-files: Mount debugfs and pstore with nosuid,nodev,noexec

These permissions are not needed. Systemd also mounts these file systems
without these permissions on other Linux distributions.

Dropping these permissions should make the system more secure.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Link: https://github.com/openwrt/openwrt/pull/16960
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit b88d51898d126d2f918cb476d4158e9fcd62492c)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

base-files: Create /run and /run/lock folder

Create the folder /run and /run/lock using symlinks. Other Linux
distributions also have these folders and some applications might already
depend on them. Just create symlinks pointing to the older folder.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Link: https://github.com/openwrt/openwrt/pull/16961
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit 3b710375dd97af1e0416904c4174d11c51e3f09c)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

x86: make i915 as a kmod with required firmware

i915 driver requires to load correct firmware to work on latest x86
GPU, it is more reasonable to make it as a kernel module, so that
initramfs is not required, and it can also save some space from the
kernel image comparing being a built-in driver

Signed-off-by: Joe Zheng <joe.zheng@intel.com>
Link: https://github.com/openwrt/openwrt/pull/16276
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 77cfe8fd15d3d0f77ee16660d06a174f41b53444)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

linux-firmware: rename packages for i915 firmware

Change the package name from intel-igpu-firmware-* to i915-firmware-*,
the prefix "intel-igpu" is misleading, i915 firmware is not only for
iGPU but also for dGPU now.

Remove the redundant "intel" as i915 is already well known.

More accurate file classification to handle following files correctly:

  adlp_dmc.bin
  mtl_huc.bin
  mtl_huc_gsc.bin
  mtl_gsc_1.bin

The pattern in regex is "([[:alnum:]]+)_([[:alnum:]]+)(_[\w-.]+)?\.bin",
where $1 is the platform, $2 is the firmware type (dmc, guc, huc, etc.),
and the optional $3 which is revision or other suffix.

Glob first to narrow down the target file set, and then split with "_"
to extract the firmware type (remove the ".bin" in case there is no $3)

Add package "i915-firmware" as a meta package to install all the i915
firmwares, it is a balance between simplicity and optimization.

* Installing all the available firmwares as a whole, can support all the
  platforms, not only the current one but also the future ones. The
  price to pay is the increased size.
* If we want to minimize the storage, we can customize to install the
  necessary ones only, even for the target platform only (e.g. ADL) and
  skip the others. The price to pay is the time to tune.

What I am going to do is:

* Let drm-i915 driver depend on i915-firmware-dmc, which is small and
  can cover most of the old platforms
* Let the user select i915-firmware to install all the i915 firmwares as
  a whole to cover the latest or future platforms

Signed-off-by: Joe Zheng <joe.zheng@intel.com>
Link: https://github.com/openwrt/openwrt/pull/16276
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit ca00bafd7e05b3f77e38acac6add223ce46df585)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

utils: Add the omnia-eeprom utility

Add a new utility, `omnia-eeprom`, which can be used to print / set
EEPROM fields on Turris Omnia.

One example when this utility might be useful is if the board
experiences random crashes due to newer versions of the DDR training
algorithm in newer U-Boot. The user can change the DDR speed from 1600K
to 1333H to solve these issues, with

  ```
  omnia-eeprom set ddr_speed 1333H
  ```

Signed-off-by: Marek Behún <kabel@kernel.org>
Link: https://github.com/openwrt/openwrt/pull/16264
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 749a43325b98afe4430fa7e2f23ac3ee44791ad6)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

mediatek: filogic: add support for GatoNetworks GDSP

The GatoNetworks GDSP is a re-branded version of the R5000 5G Industrial
router from Yinghua Technologies.
The re-branded device comes with OpenWrt preinstalled, and an OpenWrt-based
U-Boot bootloader version. While the flash layout has been kept compatible
with the OpenWrt version found on the stock device (see [5]), the image format
changed, making a bootloader upgrade necessary.

Specifications:
SoC: Mediatek MT7981BA
RAM: 256MB
Flash: SPI-NOR 32 MiB (Winbond W25Q256)
WLAN: MT7976CN DBDC AX Wi-Fi
Switch: MT7531AE (4x LAN Gigabit ports, 1x WAN Gigabit port)
5G: Quectel RM520N modem
Watchdog: an external WDT connected to GPIO 6 is present and always running;
          the built-in Mediatek watchdog is also present and effective, but
          not used at the moment.

This porting has been tested only with 1x 5G modems installed (the device
supports up to two).

Installation:
Installation is possible via sysupgrade both in the stock device and
re-branded version. However, in the former case, updating the bootloader is
required.

OpenWrt-based U-Boot Bootloader installation
--------------------------------------------
The firmware flashed in the re-branded device at manifacturing time will
flash an OpenWrt-based U-Boot bootloader with some extra recovery features
(see [1]) at first boot.
To update the bootloader, you need to install the mtd-rw module and
insmod it:

insmod mtd-rw i_want_a_brick=1

Then update relevant flash partitions:

mtd erase u-boot-env
mtd erase BL2
mtd erase FIP
mtd write openwrt-mediatek-filogic-gatonetworks_gdsp-preloader.bin BL2
mtd write openwrt-mediatek-filogic-gatonetworks_gdsp-bl31-uboot.fip FIP

And reboot, making sure all previous commands ran succesfully.
If something goes wrong, you can recover your device via the mtk_uartboot
tool.
In my testing, it was possible to start the process even without (un)-plugging
the device, may be handy for remote recovery.

Installation from stock device and firmware
-------------------------------------------
To install vanilla OpenWrt in the stock device (R5000 5G Industrial router
from Yinghua Technologies) running the stock vendor firmware, you will need
to update your bootloader as described in previous section. Remember to use
-F (force upgrade) and -n (not keeping settings).

U-Boot Recovery
---------------
This procedure has been tested only with the OpenWrt-based U-boot bootloader.
Assign your system static IP address 192.168.1.1 and start a TFTP server. The
device will look for an initramfs image named
openwrt-mediatek-filogic-gatonetworks_gdsp-initramfs-kernel.bin
(so you may use openwrt/bin/targets/mediatek/filogic as root dir for your
TFTP server).
Power on the device while keeping the reset button pressed, until you see
a TFTP request from 192.168.1.10. Your environment will be restored to it's
default state.

MAC addresses assignment
------------------------
MAC addresses are assigned slightly differently than in stock firmware. In
particular, the 5 GHz Wi-Fi uses 2.4 GHZ MAC + 1, rather than reusing it with
LA bit set as done in stock firmware. This MAC address is allocated to the
device, so it can be used.
The 2.4 GHz Wi-Fi MAC address is the label MAC. LAN MAC is used to set the
special U-Boot environment ethaddr variable.

device MAC address U-Boot env variable factory partition offset
2.4 GHz Wi-Fi :84 wifi_mac 0x4
5.8 GHz Wi-Fi :85 not present not present
WAN :86 wan_mac 0x24
LAN :87 lan_mac 0x2A

Notes
-----
[1]: the OpenWrt-based U-Boot bootloader you will find installed in the
re-branded device is configured to request for the initramfs image via
TFTP for $gdsp_tftp_tries times before trying normal boot from NOR flash.
Setting this U-Boot environment variable to 0x0 will disable the feature,
which is not implemented in this patch.
[2]: the exposed UART port is connected to ttyS1; the ttyS0 console port is
not exposed.
[3]: the provided bootloader environment has no provision for operating on
BL2 and the FIP partitions. This is an intentional choice to make it
(slightly) more difficult to brick the device.
[4]: it seems GPIO 6 is used both for the "SYS" LED and external WDT.
[5] BL2 expects to find FIP payload at a fixed offset, so some constraints
apply.

Signed-off-by: Enrico Mioso <mrkiko.rs@gmail.com>
(cherry picked from commit b43194e041b17fbb574fb3721dafedcc30a20ab5)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

uboot-envtools: add support for GatoNetworks GDSP

This configuration should work with both stock and OpenWrt-based U-Boot.

Signed-off-by: Enrico Mioso <mrkiko.rs@gmail.com>
(cherry picked from commit 785ebf2baff4adadf818d71a99fdb2d2300590ba)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

uboot-mediatek: add support for GatoNetworks GDSP

The GatoNetworks GDSP is a re-branded version of the R5000 5G Industrial
router from Yinghua Technologies.

Advantages over stock bootloader:
1. supports serving the external GPIO WDT, allowing for easier work in U-Boot
   shell
2. supports cool features like netconsole, easy recovery, scripting and so on
3. allows using FIT image and image integrity validation

and ultimately gives you much more flexibility to implement your tweaks.

Known issues
------------
To make it easier to operate the device, console I/O multiplexing support has
been enabled in U-Boot configuration. Setting I/O related U-Boot environment
variables to something like "serial,nc" will have the desired effect. Still,
setting these variables to such a value in the persistent environment will
lead to a crash and make it impossible to boot the system or recover it. I
decided to leave it on anyway since I think it can be very practical in
development.

Signed-off-by: Enrico Mioso <mrkiko.rs@gmail.com>
(cherry picked from commit 7f2c7b3238ee26180d44528ae98fd2f54e874511)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

scripts: download-check-artifact.sh: fix shellcheck and improve

Fixes following shellcheck's recommendations:

 In scripts/download-check-artifact.sh line 24:
exit $1
             ^-- SC2086 (info): Double quote to prevent globbing and word splitting.

 In scripts/download-check-artifact.sh line 53:
local sum="$(shasum -a 256 "$image_file")";
                      ^-^ SC2155 (warning): Declare and assign separately to avoid masking return values.

 In scripts/download-check-artifact.sh line 72:
 cd "/tmp/verify.$$"
 ^-----------------^ SC2164 (warning): Use 'cd ... || exit' or 'cd ... || return' in case cd fails.

 In scripts/download-check-artifact.sh line 114:
printf "Keyserver to use? [$keyserver_url] > "
                       ^-- SC2059 (info): Don't use variables in the printf format string. Use printf '..%s..' "$foo".

 In scripts/download-check-artifact.sh line 115:
read url; case "${url:-$keyserver_url}" in
                ^--^ SC2162 (info): read without -r will mangle backslashes.

While at it make it clear, that it is possible to download/check any
build artifacts like even SDK or ImageBuilder.

Link: https://github.com/openwrt/openwrt/pull/16871
(cherry picked from commit 27c2c140b16cb490f20c1c2259b3decf50ae5762)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

scripts: add download-check-artifact.sh script

While checking wiki documentation about GPG signatures checking I found
this nice script and I've thought, that it would be nice to provide it
officially instead of some random internet site.

Usage example:

  $ ./scripts/download-check-artifact.sh https://downloads.openwrt.org/snapshots/targets/mediatek/filogic/openwrt-mediatek-filogic-openwrt_one-squashfs-sysupgrade.itb

  1) Downloading image file
  =========================
  ########################################### 100.0%

  2) Downloading checksum file
  ============================
  ########################################### 100.0%

  3) Downloading the GPG signature
  ================================
  ########################################### 100.0%

  4) Verifying GPG signature
  ==========================
  gpg: Signature made Tue 05 Nov 2024 05:21:50 PM UTC
  gpg:                using EDDSA key 92C561DE55AE6552F3C736B82B0151090606D1D9
  gpg: Good signature from "OpenWrt Build System (Nitrokey3) <contact@openwrt.org>" [ultimate]
  Primary key fingerprint: 8A8B C12F 46B8 36C0 F9CD  B36F 1D53 D187 7742 E911
       Subkey fingerprint: 92C5 61DE 55AE 6552 F3C7  36B8 2B01 5109 0606 D1D9

  5) Verifying SHA256 checksum
  ============================
  openwrt-mediatek-filogic-openwrt_one-squashfs-sysupgrade.itb: OK

  Verification done!
  ==================
  Downloaded artifact placed in '/var/home/ynezz/dev/openwrt/openwrt.git/openwrt-mediatek-filogic-openwrt_one-squashfs-sysupgrade.itb'

  Cleaning up.

Adding file in a state as downloaded from https://www.abitare.org/bin/download.sh

References: https://openwrt.org/docs/guide-quick-start/verify_firmware_checksum?s[]=gpg#linux
Signed-off-by: David S. H. Rosenthal <dshr@abitare.org>
Link: https://github.com/openwrt/openwrt/pull/16871
(cherry picked from commit 4c9031fda271ed7d4503642f63ff824faa6cc3a0)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

build: add comments to makefiles

Add comments to build system makefile functions and variables to help
developers in understanding build system internals and ease the
development process.

This patch adds some documentation examples with proposed doxygen-like
syntax. Hopefully, this would start the discussion and result in
generation of the makefile documentation guidelines.

Signed-off-by: Sergey Matsievskiy <matsievskiysv@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16888
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit a72c8c770584075782959f2aa5c1487d68517edd)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

qualcommax: ipq60xx: add NSS reserved memory

It seems that despite NSS not being supported in OpenWrt the memory it
usually uses needs to be reserved anyway for stability reasons.

Link: https://github.com/openwrt/openwrt/pull/16928
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 8729d2a64d282d37b69c0853269af401a0102517)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

cryptodev-linux: change home-page URL

Most links and info on the org page are broken or defunct, and all new
versions are available on the github based repo anyway.

Signed-off-by: Paul Donald <newtwen+github@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16910
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit ce0b843ac7d27e124cb9ca13e9a1aeb115fcaf23)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

cryptodev-linux: give a coherent description

For the last ten years, nobody has noticed that this driver doesn't
actually say what it is for. ;)

Signed-off-by: Paul Donald <newtwen+github@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16910
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit 3df5966f9453ecefcabc893f6449c8c4f1054aef)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

realtek: rtl83xx: use devm for mutex_init

mutex_destroy is missing in remove.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16926
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 1125ed408c7f0d130b5606ab39e4dd98a41d5f0a)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

ramips: mt7621: fix typo

Missing u caused the regulator to fail probe.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16925
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 3e42526404f021094ae3fa41524c396d753f4c6c)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

mdadm: backport musl patch

Needed to fix compilation with GCC14.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16912
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 1f3d5858b441cf4e1195b96242dd08d68538e83d)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

lantiq: use devm for mutex_init

It's common to avoid calling mutex_destroy when done. It's not correct strictly speaking.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16832
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit fccff963c0a6ba035455daea61566f4d121643e3)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

lantiq: use more devm for i2c

i2c_add_numbered_adapter is the wrong function to use here. It requires
setting nr to some value, otherwise it behaves the same as
i2c_add_adapter. nr is not set.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16825
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 44824f0b787b13e2f2fccc58256b2c42fa3f9f5d)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

treewide: remove THIS_MODULE assignment

Matches upstream coccinelle check: api/platform_no_drv_owner.cocci.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16846
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit cc98cfafd799d502fbe09ddb62bab68633d570f5)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

hostapd: split long lines

These two were getting rather long.

Signed-off-by: Leon M. Busch-George <leon@georgemail.eu>
Link: https://github.com/openwrt/openwrt/pull/16849
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit 8b6d5874b88bbe2e7a53bcab0e102b8781480880)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

menuconfig: add help description for Stack Smashing Protection options

Descriptions taken from:
https://wiki.osdev.org/Stack_Smashing_Protector

Signed-off-by: Marius Dinu <m95d+git@psihoexpert.ro>
Link: https://github.com/openwrt/openwrt/pull/16897
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit cbf7be9fb37baaaf7ee5c2b6315028bd47c764bc)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

ramips: mt7621: add missing regulator-boot-on

What seems to be happening is that the kernel requests an ACTIVE_LOW
gpio initially and sets it to high later based on gpios in dts.

This seems to break some devices where the bootloader sets it to high.

Fixes: e612900ae0 ("ramips: mt7621: convert usb power to regulators")
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16877
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit a7bc6bf7dba98c77724cb8284d21a9c77c263d56)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

rampis: mt7621: Edgerouter-X adjust IMAGE_SIZE to reflect new kernel size

IMAGE_SIZE was previously set to kernel1 + ubi size = 256768k, now
kernel1 is 6MB adjust this value to add 3072k to total image size.

Signed-off-by: Tim Lunn <tim@feathertop.org>
Link: https://github.com/openwrt/openwrt/pull/15194
(cherry picked from commit 5b406bccaccadb494cad139b3f69e17ab62d3968)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

ramips: mt7621: Edgerouter-X drop factory image

Edgerouter-X factory images have not built automatically since 19.x due
to images being over 3MB. While it was possible to build custom images
with very stripped down config, this is no longer possible with the size
increases of linux 6.1 and 6.6.

Drop code for generation of factory images, if some dev later wishes to
try custom images they can revert this commit.

Signed-off-by: Tim Lunn <tim@feathertop.org>
Link: https://github.com/openwrt/openwrt/pull/15194
(cherry picked from commit 4d90b79704efba57ab158c47ae6936ce4c1ef185)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

rampis: mt7621: Re-enable Edgerouter-X image builds

With the new layout providing 6MB for kernel there will be no issues
with kernel size affecting build of images.

Re-enable image builds for Edgerouter-X and X-SFP.

Signed-off-by: Tim Lunn <tim@feathertop.org>
Tested-by: Mauri Sandberg <maukka@ext.kapsi.fi>
Link: https://github.com/openwrt/openwrt/pull/15194
(cherry picked from commit 71b4842e3bdaed91cd2daa4b28a208559c3de7c6)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

rampis: mt7621: Edgerouter-X update compat version to 2.0

Use compat version to indicate that the new layout for larger kernels
is in place. This handled by the patch to ubnt.sh to always select
the kernel1 slot for flashing and as active kernel slot.

Signed-off-by: Tim Lunn <tim@feathertop.org>
Tested-by: Mauri Sandberg <maukka@ext.kapsi.fi>
Link: https://github.com/openwrt/openwrt/pull/15194
(cherry picked from commit 1bdbd511b27017157e92ffe128e66853d169dcbf)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

rampis: mt7621: Edgerouter-X set compat version message

Refer user to the wiki page for instructions on how to migrate to the
new kernel layout.

Signed-off-by: Tim Lunn <tim@feathertop.org>
Tested-by: Mauri Sandberg <maukka@ext.kapsi.fi>
Link: https://github.com/openwrt/openwrt/pull/15194
(cherry picked from commit 70c371d91cda300d7c655f019cb2076d2dd537fb)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

rampis: mt7621: resize Edgerouter-X kernel size to 6MB

With the updated partition layout set in dts, set the KERNEL_SIZE
parameter to 6MB allowing builds of Linux 6.1 and later to fit.

Signed-off-by: Tim Lunn <tim@feathertop.org>
Tested-by: Mauri Sandberg <maukka@ext.kapsi.fi>
Link: https://github.com/openwrt/openwrt/pull/15194
(cherry picked from commit 09a6bffb6e0a714f20d71043ae8382af4c5925b6)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

rampis: mt7621: Edgerouter-X disable toggling of kernel slots

Uboot selects which kernel slot to boot based on a flag in the factory
mtd partition. Patch ubnt.sh to ensure always flash to kernel1 slot and
update flag if required.

Signed-off-by: Tim Lunn <tim@feathertop.org>
Tested-by: Mauri Sandberg <maukka@ext.kapsi.fi>
Link: https://github.com/openwrt/openwrt/pull/15194
(cherry picked from commit 5e355f1f9084c1b6de3151fbcde948d736292d6c)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

rampis: mt7621: Convert Edgerouter-X dts to single kernel config

The OEM layout for Edgerouter X provides for two 3MB kernel slots.
As of linux 6.1 the kernel images no longer fit and as such
Edgerouter X builds have been disabled in Main.

Revise the layout to make kernel1 slot 6MB and drop kernel2 slot.

This patch applies the required changes to the dts file.

Signed-off-by: Tim Lunn <tim@feathertop.org>
Tested-by: Mauri Sandberg <maukka@ext.kapsi.fi>
Link: https://github.com/openwrt/openwrt/pull/15194
(cherry picked from commit dc51c4355d48e91ed8fd06d70f0bf71b0b4bc4d2)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

fritz-tools: fix mixed indentation

Signed-off-by: David Bauer <mail@david-bauer.net>
(cherry picked from commit 685e979d470a09361defe1612b001dfc3b7d9832)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

packages: drop DEVICE specific dependency for omnia

Both packages `ombnia-mcu-firmware` and `omnia-mcutool` would depend on
a specific device. The buildbots however build all devices and therefore
the package isn't build at all, due to unmet dependencies.

While this didn't cause issues with OPKG, APK fails actively due to the
missing packages. Drop the specific dependency, however wants to install
unrelated firmware on any device can do that anyway.

Signed-off-by: Paul Spooren <mail@aparcar.org>
(cherry picked from commit f35a29d63f445d784f1e266eaa6d3825a249cd6c)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

ath79: gpio-rb91x-key: use gpiochip_get_data

No need for a custom function that does the same thing.

Oversight from devm conversion.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16748
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit e99b5832e801fdf2ec366419797360e09c2348c6)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

ath79: gpio-latch-miktorik: use gpiochip_get_data

Oversight from devm conversion. No need for a custom static inline
function.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16747
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 6d2ed484fa2f62e355893f5e1fe3734c88201bfc)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

ramips: mt7621: convert usb power to regulators

These things are regulators. Should silence dmesg messages about using
dummy regulators.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16804
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit e612900ae0fd6c4624fa661e004536216409107f)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

kernel: can: package driver for Geschwister Schneider interface

Package gs_usb CAN driver, which is for Geschwister Schneider and
bytewerk.org candleLight compatible USB/CAN interfaces.

Signed-off-by: Chuanhong Guo <gch981213@gmail.com>
(cherry picked from commit ef4df2b5e9d66a6461982a0a81f928abb0f6a074)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

ipq40xx: use devm for mutex_init

Avoids having to call mutex_destroy in remove.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16755
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 57f2c72505f24b41fd9beaef3a8aa713b436cdb9)
Link: https://github.com/openwrt/openwrt/pull/17097
Signed-off-by: Petr Štetiar <ynezz@true.cz>

build: profiles.json: add kernel version information

Currently downstream tools like ASU lack information about kernel
version to find out the relevant kmod build folder on downloads server.

So lets fix it by providing a new `linux_kernel` JSON array which would
for the start provide Linux kernel version, revision and vermagic
information.

  "linux_kernel": {
     "release": "1",
     "vermagic": "b57450c07d3a786158c3601fc5cee57d",
     "version": "6.6.61"
   },

Fixes: openwrt/openwrt#17036
Fixes: efahl/owut#9
Co-developed-by: Petr Štetiar <ynezz@true.cz>
Signed-off-by: Eric Fahlgren <ericfahlgren@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17042
Signed-off-by: Petr Štetiar <ynezz@true.cz>
(cherry picked from commit c857145e034f623e31ab2028049a547ecd94ce85)

mtd-utils: add ubihealthd to nand-utils

Add ubihealthd to the nand-utils package, auto-create UCI config for
each UBI device and launch the daemon on boot.

The default time interval between scrubbing a random PED is 120 seconds
which means that a fully used 128 MiB flash chip gets scrubbed in about
a day and a half. The interval can be adjusted in UCI using the
'interval' option.

Suggested-by: Rodrigo Balerdi <lanchon@gmail.com>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Link: https://github.com/openwrt/openwrt/pull/16973
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit 7e287b563aa1043a1af54d5e2b0bd96326371732)

procd: update to git HEAD

 7330fa5 initd: mount /sys and /proc with MS_RELATIME

Fixes mounting /proc in unpriviledged user namespace.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit 1e8505ac4e88212106fe78486cfcbe9cae7660f9)

include: meson.mk: honor CONFIG_DEBUG

Set buildtype according to CONFIG_DEBUG.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit be3e7a84bd37f29e14e4c6a89b8c79bc9808c36c)

include: cmake.mk: honor CONFIG_DEBUG

Set CMAKE_BUILD_TYPE according to CONFIG_DEBUG.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit fc9ec5bd3d22d50654af154fc779517fa7aecb33)

kernel: modules: package driver for generic MIPI DBI TFT panels

Package kmod-drm-panel-mipi-dbi as well as modules it depends on in
order to support a wide range of MIPI DBI complaint SPI-connected
TFT panels.

See https://github.com/notro/panel-mipi-dbi/ for more information on
how to use specific panels.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit 0a7b7ef27cfc00521a0f99de073a7168f1b79e8f)

ramips: mt7621: support openwrt,netdev-name for renaming interfaces

Edgerouter X currently has its eth1 port on the switch missing since there
is a naming conflict currently.

So, as the root cause is mixing kernel support for DSA interfaces having
predictable names set via "label" property vs others having it assigned
dynamically lets avoid the conflict by using our own custom property as
suggested upstream [1].

So, add support via "openwrt,netdev-name" property and use it on ERX.

Fixes: 2a25c6ace8d8 ("ramips: get rid of downstream network device label patch")
Fixes: #15643
Link: https://github.com/openwrt/openwrt/pull/17062
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 5695267847c2fe51290fdbde6ecd7114654f7988)

wifi-scripts: Fix parsing of Capabilities

Fixup capabilities parsing in iw output.

In addition to the normal capabilities iw now also outputs HE MAC, HE
PHY and EHT MAC and EHT PHY capabilities. Exclude them in the parsing.

The grep returns this with mac80211-hwsim:
```
root@OpenWrt:~# iw phy phy0 info | grep 'Capabilities:'
Capabilities: 0x107e
HE PHY Capabilities: (0x02bfce0000000000000000):
EHT PHY Capabilities: (0x7c0000feffff7f01):
HE PHY Capabilities: (0x02bfce0000000000000000):
EHT PHY Capabilities: (0x7c0000feffff7f01):
HE PHY Capabilities: (0x02bf000000000000000000):
Capabilities: 0x107e
HE PHY Capabilities: (0x1cbfce0000000000000000):
EHT PHY Capabilities: (0xfc1f3ffeffff7f37):
HE PHY Capabilities: (0x1cbfce0000000000000000):
EHT PHY Capabilities: (0xfc1f3ffeffff7f37):
HE PHY Capabilities: (0x1cbf000000000000000000):
HE PHY Capabilities: (0x1cbfce0000000000000000):
EHT PHY Capabilities: (0xfefffffeffffff7f):
HE PHY Capabilities: (0x1cbfce0000000000000000):
EHT PHY Capabilities: (0xfefffffeffffff7f):
HE PHY Capabilities: (0x1cbf000000000000000000):
Capabilities: 0x107e
```

With busybox 1.36.1 the ht_cap_mask variable will be set to
-72057598332895361. With busybox 1.37.0 it will be set to -1.
Both values are wrong, after this change it will be set to 4222
(0x107E).

Link: https://github.com/openwrt/openwrt/pull/17043
(cherry picked from commit adf958c9192d94fbeeb0ec8e16800b0071455909)
Link: https://github.com/openwrt/openwrt/pull/17055
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel: bump 6.6 to 6.6.63

Changelog: https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.6.63

Manually rebased:
bcm27xx/patches-6.6/950-0297-staging-vchiq_arm-Add-36-bit-address-support.patch

All other patches automatically rebased.

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/16997
(cherry picked from commit eb395585ae1b7560bbcea9544f428db2a9105152)
Link: https://github.com/openwrt/openwrt/pull/17055
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel: bump 6.6 to 6.6.62

Changelog: https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.6.62

All patches automatically rebased.

Build system: x86/64
Build-tested: x86/64/AMD Cezanne, flogic/glinet_gl-mt6000, flogic/xiaomi_redmi-router-ax6000-ubootmod, ramips/tplink_archer-a6-v3
Run-tested: x86/64/AMD Cezanne, flogic/glinet_gl-mt6000, 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/16997
(cherry picked from commit 611bfa4d72c87e8163d7193d9557479fc7312467)
Link: https://github.com/openwrt/openwrt/pull/17055
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

debugcc: Mark it nonshared

This tool was build in the phase 2 build, there the TARGET dependencies
are probably not meat. Mark it as non shared to build it together with
the targets where this option is set.

Link: https://github.com/openwrt/openwrt/pull/16975
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit febe1db43d1a01f83e474bcb143dcb3997bd35a3)
Link: https://github.com/openwrt/openwrt/pull/17055
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

ath79: lzma-loader: fix syntax error

C compiler can't parse '#else if'.

Fixes: f84a9f7dc095 ("ath79: add support for Huawei AP6010DN")
Signed-off-by: Shiji Yang <yangshiji66@qq.com>
Link: https://github.com/openwrt/openwrt/pull/16989
(cherry picked from commit 98f26346cbbb45e5c0417f57846ce1783e884ea2)
Link: https://github.com/openwrt/openwrt/pull/17055
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel: bump 6.6 to 6.6.61

Changelog: https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.6.61

Manually rebased:
bcm27xx/patches-6.6/950-0998-i2c-designware-Add-support-for-bus-clear-feature.patch

All other patches automatically rebased.

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/16959
(cherry picked from commit 487ca61f917f9308222faf481ad8be4e2cef0c90)
Link: https://github.com/openwrt/openwrt/pull/17055
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

mediatek: fix Cudy AP3000 Outdoor label-mac

The label-mac does not match the one assigned to the ethernet interface.
Use the mac-address assigned to the wifi interface instead, as it
matches the one found on the device label.

Signed-off-by: David Bauer <mail@david-bauer.net>
(cherry picked from commit b8b658bc0d22df31e7ac2d07cdd2af7155d12630)

imagebuilder: fix OPKG package_whatdepends support

Trying to use 'package_whatdepends' feature of the ImageBuilder with OPKG
will currently fail as OPKG does not support "list --depends" call at all,
it seems that this is a mixup from the original APK support commit.

So, lets restore 'package_whatdepends' support for OPKG by calling
"whatdepends -A" instead as we used to before APK support.

Fixes: d788ab376f85 ("build: add APK package build capabilities")
Link: https://github.com/openwrt/openwrt/pull/17022
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 52519a59a859d349f154ed79e0a3d63457b6aa6a)

ipq806x: tr4400v2: fix dtc warnings by deleting stock partitions

The original set of stock partitions was kept in the TR4400 v2 port,
with the same partition numbers but their names prefixed with 'stock_'.
This allowed scripts (installation, back to stock, etc) to run on both
stock and OpenWrt firmware. But this triggers warnings in the device
tree compiler, as partitions of the old and new schemes overlap.

This commit fixes the dtc warnings by deleting the stock partitions,
also renumbering some of the remaining MTD partitions in the process.
Additionally, the 'fw_env' partition is set to read-only.

These changes can break existing scripts as well as user configurations
that utilize the 'extra' partition. Users wanting to run old scripts can
do so by reverting to the 23.05 series releases.

Signed-off-by: Rodrigo Balerdi <lanchon@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16958
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 8ec35a2a1533ccff8a4b0e6f4f0a0954d1bee6a6)