unetd: update to Git HEAD (2025-01-26)

322500403615 service: add default group @ to match all nodes
5f7860306200 ubus: rename unetd_ubus_notify to unetd_ubus_network_notify
d13752814651 enroll: add PEX sub-protocol to support enrolling new nodes into a network

Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry picked from commit c0f06cb6ac002ce90c10839b3d98aed8b41a2da5)

ucode: update to Git HEAD (2025-02-10)

a362263595e6 vm: fix inverted condition in uc_vm_signal_handlers_setup()
1b8660d33fb7 tests: adjust testcases after previous commit
4efad5388184 main: fix crash on printing -p output
0b80ceb29311 vm: close signal pipe in uc_vm_signal_handlers_reset()
175686c97d82 uci: eliminate usage of global variables
0f69da9c3777 struct: fix memory leak in buffer.pull()
47f3a6b69908 debian/control: Build-Depend on cmake
01efa7b92100 debian/control: Fix Standards-Version, add Bugs
f50149d67fef debian/source/format: change to native
bd001ae4f830 debian/control: libucode Recommends ucode-modules
87776a0bc2b2 docs: Mention `libmd` in macOS docs
826a5c9de7d3 docs: Fix compilation command for macOS
1fa771fc91c0 debian/copyright: fix missing dot to separate paragraphs
60c0e5eae992 debian/control: pkg-config now called pkgconf
244570db4f8e debian/changelog: switch versioning to native package scheme
fa2326fc00a0 .gitignore: ignore files generated by debuild
faec62367524 debian: ignore lintian warning no-manual-page
af881af4b0dc main: add global SCRIPT_NAME variable
5cbd32514187 socket: fix AF_PACKET recvmsg() and sockaddr formatting
6f71a35c4cfa rtnl: properly handle runtime exceptions in listener callback
6dd08716205f ci: bump actions/upload-artifact to v4
001ced17982f socket: properly support SO_ATTACH_FILTER sockopt
f0e865fe3ddd ubus: add named parameter support in functions with many params
18a2ffaf51fa fs: ioctl: export constants for direction values
0ba75bea925d fs: ioctl: improve ioctl read to avoid allocating twice
9da8b8aa4ad0 ubus: add defer.await() method
17dbf0bef4b6 ubus: add request get_fd/set_fd methods
4acb960c90f7 ubus: add support for sending file descriptors via ubus.call/defer
22b9523565a5 ubus: add support for receiving file descriptors in call and defer
c0d1654a450b ubus: add support for channels
b32321beb112 vm: fix crash due to stale frame pointer

Fixes: https://github.com/jow-/ucode/issues/254
Fixes: https://github.com/jow-/ucode/issues/255
Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry picked from commit 0e672e980650d8f890e620d1c359b78ef3a524d2)

ucode: update to Git HEAD (2024-12-06)

47b54cf5a4b6 types: introduce `ucv_array_sort_r()` and `ucv_object_sort_r()`
efeb57806552 types, vm: refactor usage of global variables
f9d2faf67de6 vm: reset signals when freeing VM
4e86847d802d lib: utilize `ucv_array_sort_r()` and `ucv_object_sort_r()`
c71444ea301f types: ucv_resource_create(): rename `typename` parameter to `type`
373df7299c79 nl80211: properly support split_wiphy_dump for single phys
9bcd25f54708 lexer: Preserve keyword, regexp flags until processing non-comment tokens
0a7ff4715cb8 main: pretty-print `-p` output by default
4c3d5b469156 struct: Add new buffer API for incremental packing/unpacking
efc4122124cb struct: do not use global variables for caching types

Fixes: https://github.com/jow-/ucode/issues/248
Fixes: https://github.com/jow-/ucode/issues/250
Fixes: https://github.com/efahl/owut/issues/25
Link: https://github.com/openwrt/openwrt/pull/17191
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit af6f1a90bbc824bd22814568c5b6990aa082ce8b)

ucode: update to latest HEAD

b0b5d93 Merge pull request #234 from IdWV/fs
60e7a88 Merge pull request #232 from sebastianertz/lib-digest
1752779 digest: implement compile time option to exclude less common algorithms
c7268a1 ci: include libmd in MacOS CI builds
fcb6f70 lib: introduce digest library
1323a27 Merge pull request #246 from jow-/fix-upvalue-resolve
ed5ce8f types: resolve upvalue values in arrays and objects
a6e0641 vm: resolve upvalues before pushing them onto the stack
ef1baab ci: drop OpenWrt tests for now
63e18ea fs: eliminate the usage of global variables
b1bd7b5 types: add ucv_resource_create() helper
3408edf Merge pull request #244 from nbd168/nl80211
8af77e7 nl80211: add new attributes for multi-radio support
1423ad7 nl80211: cover extended feature and EHT rate info attributes
ee1d6d8 Merge pull request #237 from sebastianertz/math
4b18a9b Merge pull request #213 from jow-/improve-vector-macros
1f022c0 math: removed global variable for thread safety
e5fe6b1 treewide: refactor vector usage code
20307ee utils: improve vector utilities
aa18952 Merge pull request #241 from jow-/socket-local-fanout-decl
79ccd9c socket: provide local definition of `struct fanout_args`
402280d Merge pull request #239 from jow-/safe-insert-during-obj-iteration
07afe96 Merge pull request #240 from jow-/stricter-number-conversion
736d450 types: fix potential use after free on adding keys during iteration
4134e71 vallist: more thoroughly check for trailing garbage after numeric string
9cf53dd Merge pull request #226 from jow-/lexer-improvements
2b2e732 lexer: make api functions public
855854f lexer: emit comment and template statement block tokens
328a50f lexer: improve token position reporting
fa22732 Merge pull request #225 from jow-/compiler-fix-keyword-property-labels-after-spread
6e88c62 Merge pull request #224 from jow-/lib-fs-readline-leak
94d1211 compiler: properly treat property names after spread expressions
67cd123 fs: fix potential memory leak on i/o errors in .read()

Signed-off-by: John Crispin <john@phrozen.org>
(cherry picked from commit ba9cdbeea862fada261b0a874e464b63ffa860ac)

imx: coretexa53: add network configuration for gw82xx-2x

The GW82xx-2x has two network ports. By convention, the first
port (eth0) should be the WAN port and the second port (eth1)
should be the LAN port.

Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Link: https://github.com/openwrt/openwrt/pull/17717
Signed-off-by: Robert Marko <robimarko@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17762
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit d2eaaa90d502fa418d7d9bb8d4e6ccbe0131c832)

imx: 6.6: add a patch to work around a PCI init hang on IMX8M{M,P}

The act of attempting link at gen1 then trying to link at higher speeds
causes a hang with the specific PCIe switch used on the Gateworks Venice
boards. Work around this by linking at the highest speed first as is
common with all other PCI controller drivers.

Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Link: https://github.com/openwrt/openwrt/pull/17717
Signed-off-by: Robert Marko <robimarko@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17762
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 23a18a57cf750d9081353cdd34e44982ea10e191)

imx: 6.6: add some backported kernel patches for Gateworks Venice

Add several dt backports for the Gateworks Venice product family.

Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Link: https://github.com/openwrt/openwrt/pull/17717
Signed-off-by: Robert Marko <robimarko@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17762
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 1ce87cf25c19f030b81954639545db2ab7f17062)

generic: 6.6: add backported btsdio patch for cyw437

Add a backport of a btsdio patch for the cyw437 bluetooth controller.

Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Link: https://github.com/openwrt/openwrt/pull/17717
Signed-off-by: Robert Marko <robimarko@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17762
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit c9f11523dd5b8639b24b97043d70f2a197a00a84)

mediatek: fix nmbm configuration mismatch (Xiaomi AX3000t)

This commit fixes nmbm configuration mismatch error on Xiaomi AX3000t
with Winbond W25N01KVZEIR spi-nand:
'''
[    0.786783] NMBM configuration mismatch
'''

Root cause:
1. U-Boot W25N01KV spi-nand driver ia compiled with 64B OOB size for the
   chip and store this size in the nmbm signature;
2. Linux W25N01KV driver use 96B OOB.

The change doesn't affect AX3000t variants with other spi-nand chips
(ESMT, Foresee) because their Linux drivers use 64B OOB.

Fixes: openwrt#16972
Tested-by: Aleksandr Danilov <sc16me@gmail.com>
Signed-off-by: Mikhail Zhilkin <csharper2005@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17549
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 3299d19c01080c3320e3ff6964d7825da81bb8b4)
Signed-off-by: Mikhail Zhilkin <csharper2005@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17898
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel: nmbm: add mediatek,bmt-mtd-overridden-oobsize property

This commit adds new "mediatek,bmt-mtd-overridden-oobsize" property. The
property helps avoid "NMBM configuration mismatch" error if mtd "OOB size"
is not equal to the "spare size" which is stored in the nmbm signature.

Signed-off-by: Mikhail Zhilkin <csharper2005@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17549
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit e585ae70d4f2a563b3b0e430e15ee19ad25caeb5)
Signed-off-by: Mikhail Zhilkin <csharper2005@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17898
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

umdns: add firewall rule for configured interfaces

Makes it easier to enable MDNS on wan without having to edit the firewall
configuration for it

Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry picked from commit 05138fe898ce2d102795bdc979a6d874a3d24424)

build: conditionally build llvm bpf toolchain by default

Unless another toolchain is present (or selected), build the bpf toolchain
whenever a package is selected that needs it.

Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry picked from commit 6605e45676815156f0c47d4117ee401e6616fcac)

kernel: r8168: update to v8.055.00

Changelog: https://github.com/openwrt/rtl8168/compare/8.054.00...8.055.00

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
(cherry picked from commit 5dd763680e874b0082e4cfbbc729fa75ee056601)

mt76: update to Git HEAD (2025-01-22)

3e85822b9c66 page_pool: remove PP_FLAG_PAGE_FRAG
a22d59e4ad50 tools: fix allocation check and missing memory freeing

Signed-off-by: Felix Fietkau <nbd@nbd.name>
Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
(cherry picked from commit 17dc079)
Link: https://github.com/openwrt/openwrt/pull/17707
Signed-off-by: Robert Marko <robimarko@gmail.com>

generic: 6.6: backport upstream r8169 patches

b11bff90f2ad r8169: add support for RTL8125BP rev.b
b3593df26ab1 r8169: add support for RTL8125D rev.b
b299ea006928 r8169: adjust version numbering for RTL8126
bb18265c3aba r8169: remove support for chip version 11
2e20bf8cc057 r8169: remove unused flag RTL_FLAG_TASK_RESET_NO_QUEUE_WAKE
e340bff27e63 r8169: copy vendor driver 2.5G/5G EEE advertisement constraints

The EEE advertisement patch has been reworked for linux v6.6 because
phy_set_eee_broken() is only present on linux >= v6.13 and eee_broken_modes
declaration has been converted to a bitmap, so linkmode_set_bit() can't be
used either.
https://github.com/torvalds/linux/commit/e340bff27e63ed61a1e9895bed546107859e48a7
https://github.com/torvalds/linux/commit/ed623fb8e38e2a241da12864778ec9c9cf930c65
https://github.com/torvalds/linux/commit/721aa69e708b7432af83c4bb00a30e2b7c27da28

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
(cherry picked from commit 084618f8db7d0e413b47bab3e0ea70804bf84ee2)

OpenWrt v24.10.0: revert to branch defaults

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

OpenWrt v24.10.0: adjust config defaults

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

Revert "ramips: mt7620: fix patching mac address in caldata"

This reverts commit 3295f6f1c254cd7e5e5285a05581bf6abbde8999.

It looks like the eeprom gets broken after this change.
I think this change was not tested on a real device before it was
merged.
The MAC addresses will be broken again after this revert.

Fixes: #17818
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit f628467dfd6b32ead172d33a2593f04901395343)

ipq806x: fix broken WAN on Linksys EAX and Asrock G10

Fix broken WAN on Linksys EAX and Asrock G10 by incrementing the WAN
interface MAC address + 1. This caused conflicting entry in the FDB
table and caused the WAN port to malfunction with the DSA conversion.

Fixes: #17157
Fixes: #15585
Fixes: #16604
Link: https://github.com/openwrt/openwrt/pull/17839
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit eba2fbf63814bbcf8c99c98177818476fca808a7)

ramips: fix xtal clock registration issue on RT3883

The redundant xtal clock source caused clock registration failure.

Link: https://lore.kernel.org/all/20250108093636.265033-1-sergio.paracuellos@gmail.com/
Fixes: https://github.com/openwrt/openwrt/issues/17677
Signed-off-by: Shiji Yang <yangshiji66@qq.com>
Link: https://github.com/openwrt/openwrt/pull/17830
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit cfca9d689459426a119edbc02ecb7c778a3c2b77)

mediatek: mt7622: fix SATA on BPi-R64

Two commits which made their way into Linux stable broke the SATA
support on the BPi-R64.

Fix this by reverting a node rename which broke DT-overlay application
and import a (still pending) patch re-adding the 'syscon' compatible to
the pciesys clock-controller which also contains phy-mode bits
referenced by the ahci_mtk driver expecting to access them using
syscon_regmap_lookup_by_phandle().

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit 69890e16b37d59b55ba64633522c727f957cb2fd)

generic: net: phy: realtek: work-around hang on SerDes setup

On some but not all devices using the RTL8221B 2.5GBit/s PHY the SerDes
setup sequence may hang under some circumstances (eg. <2500M link
partner present during boot).

RTL8221B-VB-CG 2.5Gbps PHY (C45) mdio-bus:01: rtl822xb_config_init failed: -110

Work-around the issue by performing a hardware reset and subsequent
retry of the SerDes setup, which seems to always succeed.

Doing this requires moving ALDPS setup to config_init (which is anyway
the better place for that) as it otherwise doesn't survive the reset.

Also disable listening on MDIO address 0 which may be used by other PHYs
despite being spec'ed as "broadcast address", as bus activity on address
0 may otherwise confuse the RealTek PHY for good reasons.

Tested-by: Luis Mita <luis@luismita.com>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
(cherry picked from commit c87a767801ef375feadb0a5c41d5a674ad3a7d2c)
Link: https://github.com/openwrt/openwrt/pull/17790
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

x86: Add kmod-drm-i915 as default package

Add kmod-drm-i915 to the default packages. It was build into the kernel
before and is now build as a kernel module.

Fixes: 77cfe8fd15d3 ("x86: make i915 as a kmod with required firmware")
Link: https://github.com/openwrt/openwrt/pull/17781
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 8390599c9a29dbb0019447df9086cb8667bfd3d7)
Link: https://github.com/openwrt/openwrt/pull/17792
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

ramips: pinctrl: allow mux SDXC pins for mt76x8

The mt76x8 SDXC pin register definition is incompatible with the
mtmips generic pinctrl driver structure. This hack allows us to
mux the SDXC IO to different pin groups in device tree.

Signed-off-by: Shiji Yang <yangshiji66@qq.com>
Link: https://github.com/openwrt/openwrt/pull/17446
(cherry picked from commit 05ec3b50a8e673786dda0224b23f18b5d96a4a65)
Link: https://github.com/openwrt/openwrt/pull/17754
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

bmips: shg2500: add missing led controller cells

Fixes the following DT warnings:
../dts/bcm63168-sercomm-shg2500.dts:76.4-14: Warning (reg_format): /ubus/spi@10001000/led-controller@1/led@1:reg: property has invalid length (4 bytes) (#address-cells == 2, #size-cells == 1)
../dts/bcm63168-sercomm-shg2500.dts:75.9-78.5: Warning (avoid_default_addr_size): /ubus/spi@10001000/led-controller@1/led@1: Relying on default #address-cells value
../dts/bcm63168-sercomm-shg2500.dts:75.9-78.5: Warning (avoid_default_addr_size): /ubus/spi@10001000/led-controller@1/led@1: Relying on default #size-cells value

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
(cherry picked from commit fbe0bd5f6453a61fab871bee56883afc5c6308cf)

OpenWrt v24.10.0-rc7: revert to branch defaults

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

OpenWrt v24.10.0-rc7: adjust config defaults

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

dnsmasq: add fix related to DNSSEC verification from upstream

To find the DS record for a given zone the parent zone's nameserver must
be queried and not the nameserver for the zone. Otherwise DNSSEC
verification for unsigned delegations breaks.

Signed-off-by: Uwe Kleine-König <uwe+openwrt@kleine-koenig.org>
Link: https://patchwork.ozlabs.org/project/openwrt/patch/20250127151223.1420006-1-uwe+openwrt@kleine-koenig.org/
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 6dc0f0c50cf1072ec3751c0fb1fc152a0a86487d)

apm821xx: NETGEAR WNDR4700: Fix compat version

The definition for the netgear_wndr4700 had two different
DEVICE_COMPAT_VERSION definitions.

In commit 5815884c3a2 ("apm821xx: migrate to DSA"), an additional
DEVICE_COMPAT_VERSION := 3.0 attribute was added to the device
definition. The old one with version 2.0 stayed and was defined later
overwriting the new one.

Replace the old version 2.0 with the new version 3.0

Fixes: 5815884c3a2a ("apm821xx: migrate to DSA")
Link: https://forum.openwrt.org/t/openwrt-24-10-0-rc6-sixth-release-candidate/222466/43
Link: https://github.com/openwrt/openwrt/pull/17741
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit dab52c00d712a421d6e61eebc51359e5d38ad347)

wolfssl: Update to version 5.7.6

This fixes multiple bugs and also minor security problems.

Changelog:
https://github.com/wolfSSL/wolfssl/releases/tag/v5.7.4-stable
https://github.com/wolfSSL/wolfssl/releases/tag/v5.7.6-stable

The package size increases:
525814 bin/packages/mips_24kc/base/libwolfssl5.7.2.e624513f-5.7.2-r1.apk
549408 bin/packages/mips_24kc/base/libwolfssl5.7.6.e624513f-5.7.6-r1.apk

Link: https://github.com/openwrt/openwrt/pull/17742
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 0a7e92c2442bc09eec3875aae21996d9a5430806)

wolfssl: Adjust version for apk

Adjust wolfssl version for apk by removing the "-stable"
from the OpenWrt version, although it is still needed for
upstream download archive name.

Define PKG_BUILD_DIR accordingly.

Utilize new short version to simplify ABI_VERSION calculation.

Signed-off-by: Hannu Nyman <hannu.nyman@iki.fi>
Link: https://github.com/openwrt/openwrt/pull/16906
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit be952e98bc1d768a0da5b84e59a6e7c04a1cdab8)

stm32: enable CONFIG_SMSC_PHY

The STM32MP135F-DK board uses the LAN8742 PHY.
Enable CONFIG_SMSC_PHY to have full PHY support.

Signed-off-by: Thomas Richard <thomas.richard@bootlin.com>
Link: https://github.com/openwrt/openwrt/pull/17745
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit d981f28f762a559ae50386fa089b050aaaa78555)

ramips: restore full switch performance for USW-Flex

Since commit f1c9afd80138 ("ramips: mt7621-dts: mux phy0/4 to gmac1") the
USW-Flex lan1 port has been attached directly to the CPU. This improves
routing performance but hinders switching.

This is a generally accepted trade-off in that commit but for USW-Flex it
is a questionable choice. This switch is designed to deliver PoE to remote
places and using it as a router is unlikely. Meanwhile, the lan1 port is
also PoE-in and will often be the uplink, carrying most of the traffic.

Reverting f1c9afd80138 for USW-Flex restores full 1 Gbps switching
performance on all ports.

Signed-off-by: Anders Melchiorsen <amelchio@nogoto.net>
Link: https://github.com/openwrt/openwrt/pull/17703
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 62872f8bfd5c8bc80bb014417c6ed8057bbeb773)

airoha: an7581: replace clock patch with upstream version

Replace clock patch for eMMC with upstream version to add the tag and
flag them as upstreamed.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit a9b408a97f7bf8b1f98f355fefe3c9bfc959d5a6)

airoha: an7581: replace eMMC support with upstream version

Replace eMMC support with upstream version where we declare dummy clock
and dummy regulator instead of a specific compatible. Also drop the
downstream patch for it.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit b7edf88b0fe15b32be802c7de54036ff8e66dcc1)

airoha: an7581: refresh and fix cpufreq patch

Refresh cpufreq with merged upstream version. Also fix the PM Domain
rebased patch to correctly expose the symbol for non Mediatek target.

Update dtsi with new pm domain name.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit fbbfcbf150cd7c72f49c149e7f9b9f4d0bf43e4b)

octeon: ubnt-usg: add board name to supported devices

The on-device board name reported by 'ubus call system board' is not present
in the generated profiles.json.  This results in upgrade tools being unable
to match the image with the proper device.  Let's add a 'SUPPORTED_DEVICES'
entry for the board name to fix this.

Links: https://forum.openwrt.org/t/owut-openwrt-upgrade-tool/200035/441
Links: https://github.com/openwrt/openwrt/commit/2a07270180ed0e295d854d6e9e59c78c40549efc
Signed-off-by: Eric Fahlgren <ericfahlgren@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17736
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 7165937c3b6ebe5ac692fcde1cc69ba1c7a3d9c9)

ramips: mt7620: fix patching mac address in caldata

Fix usage of non-existent 'caldata_patch_mac' function
by using the 'caldata_patch_data' function.

Fixes: #17734
Signed-off-by: Paweł Owoc <frut3k7@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17737
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 3295f6f1c254cd7e5e5285a05581bf6abbde8999)

netifd: remove leftover wifi reload_legacy call

Commit d12753929165 removed support for it

Fixes: #17738
Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry picked from commit a3154cfeb6c381576682a8cc5427a8b1bd9daf96)

unetd: fix interface teardown

Pass the correct device name in the network_del ubus call

Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry picked from commit 80ba0d958dc96fb7aba26614f71325507fabd58a)

generic: fix probe issues with RealTek RTL8221B PHYs

Import patch "net: phy: realtek: mark existing MMDs as present"

    When using Clause-45 mode to access RealTek RTL8221B 2.5G PHYs some
    versions of the PHY fail to report the MMDs present on the PHY.
    Mark MMDs PMAPMD, PCS and AN which are always existing according to
    the datasheet as present to fix that.

Fixes: #16823, #17183, #17232
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Tested-by: Aleksander Jan Bajkowski <olek2@wp.pl>
Tested-by: Juan Pedro Paredes Caballero <juanpedro.paredes@gmail.com>
(cherry picked from commit 730db6b893748517949b1d3a0c0cb8b6c1288040)

kernel: Make kmod-usb-chipidea select kmod-phy-ath79-usb

The USB PHY on the ar9330 and similar SoCs needs the PHY driver. In
OpenWrt 23.05 it was compiled into the kernel. The kernel 6.6
configuration does not compile it in any more, make the
kmod-usb-chipidea driver select it to add it to the images.

Fixes: https://github.com/openwrt/openwrt/issues/17710
Fixes: 04bdf9b3323e ("ath79: disable ath79 USB phy drivers by default")
Link: https://github.com/openwrt/openwrt/pull/17720
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit b467e5a0afebbc4982213c03f90dfc6deb717b36)

mediatek: cleanup device tree for cudy devices

Fixes typo for led properties.
Delete the unused rfb compatible.

Signed-off-by: Chukun Pan <amadeus@jmu.edu.cn>
(cherry picked from commit 64d1b1089c25d778f5f22bfc0c1590a2e83b664e)

mediatek: update status led for Cudy TR3000

Use white led when running, consistent with the stock firmware.

Signed-off-by: Chukun Pan <amadeus@jmu.edu.cn>
(cherry picked from commit 87632219d4b8f0c96ea053e375f6f324fab0806f)

mediatek: cudy: fixes 2.5G PHY interrupt support

Fixed interrupt support for 2.5G PHY.
Removed useless phy-mode on phy node.

Tested on Cudy TR3000.
Signed-off-by: Chukun Pan <amadeus@jmu.edu.cn>
(cherry picked from commit 82b69dfaf6ca92eeffb7523295d3d81a1d54323a)

mediatek: cudy: fixes typo for spi and mtd properties

Same as commit 3674689, correct 'buswidth' to 'bus-width'.
Move the nmbm properties outside the partition definition.
Change uppercase to lowercase, add missing read-only flag.

Signed-off-by: Chukun Pan <amadeus@jmu.edu.cn>
(cherry picked from commit ab375a3484db67857450ced461498ad3d7942e05)

kernel: vrx518_tc: bump PKG_RELEASE

Bump PKG_RELEASE after the previous fixes.

Signed-off-by: Sergey Ryazanov <ryazanov.s.a@gmail.com>
Link: https://patchwork.ozlabs.org/project/openwrt/patch/20250122222654.21833-5-ryazanov.s.a@gmail.com/
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 78f908407e2824bc69a65acec270ffb94eca4038)

ipq40xx: fix compatibility with linux-atm tools

atm_qos struct should be the same both for user and kernel spaces. Via
the __SO_ENCODE() macro it is used to define the SO_ATMQOS socket IOC.

During the VRX518 support introduction, the atm_trafprm sturct nested
into the atm_qos stucture was update with newer fields that are
referenced by the ATM TC layer of the VRX518 TC driver. These new fields
are intended to communicate information for extra traffic classes
supported by the driver. But we are still using vanilla kernel headers
to build the toolchain. Due to the atm.h header incoherency br2684ctl
from linux-atm tools is incapable to configure the ATM bridge netdev:

  br2684ctl: Interface "dsl0" created sucessfully
  br2684ctl: Communicating over ATM 0.1.2, encapsulation: LLC
  br2684ctl: setsockopt SO_ATMQOS 22 <-- EINVAL errno
  br2684ctl: Fatal: failed to connect on socket; File descriptor in bad state

There are two options to fix this incoherency. (a) update the header
file in the toolchain to build linux-atm against updated atm_trafprm and
atm_qos structures, or (b) revert atm_trafprm changes.

Since there are no actual users of the extra ATM QoS traffic classes,
just drop these extra traffic classes from vrx518_tc ATM TC layer and
drop the kernel patch updating atm.h.

Besides fixing the compatibility with linux-atm tools, removing the
kernel patch should simplify kernel updates removing unneeded burden of
maintenance.

Run tested with FRITZ!Box 7530 with disabled extra traffic classes and
then removed them entirely before the submission.

CC: John Crispin <john@phrozen.org>
Fixes: cfd42a0098 ("ipq40xx: add Intel/Lantiq ATM hacks")
Suggested-by: Andre Heider <a.heider@gmail.com>
Reported-and-tested-by: nebibigon93@yandex.ru
Signed-off-by: Sergey Ryazanov <ryazanov.s.a@gmail.com>
Link: https://patchwork.ozlabs.org/project/openwrt/patch/20250122222654.21833-4-ryazanov.s.a@gmail.com/
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 6d6dc3a3c967174598a44503f4af281574660356)

kernel: vrx518_tc: fix ADSL/ATM operation

ATM TC layer have some issues which effectively prevent VRX518 from
being used as ADSL modem. Specifically, there one crash during the ATM
layer configuration and wrong PVC ID selection on packet receiving what
breaks RX path. Fix both of the issues. Make subif iface registration
optional to prevent the crash (see more details in the new patch) and
update the hardcoded PVC ID to match the first allocated channel.

Run tested with FRITZ!Box 7530.

Fixes: 474bbe23b7 ("kernel: add Intel/Lantiq VRX518 TC driver")
Reported-and-tested-by: nebibigon93@yandex.ru
Signed-off-by: Sergey Ryazanov <ryazanov.s.a@gmail.com>
Link: https://patchwork.ozlabs.org/project/openwrt/patch/20250122222654.21833-3-ryazanov.s.a@gmail.com/
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 470335450e67002366fcbcd7334b15bdf008e44d)

kernel: vrx518_tc: fix RX desc phys to virt mapping

It looks like VRX518 returns phys addr of data buffer in the 'data_ptr'
field of the RX descriptor and an actual data offset within the buffer
in the 'byte_off' field. In order to map the phys address back to
virtual we need the original phys address of the allocated buffer.

In the same driver applies offset to phys address before the mapping,
what leads to WARN_ON triggering in plat_mem_virt() function with
subsequent kernel panic:

  WARNING: CPU: 0 PID: 0 at .../sw_plat.c:764 0xbf306cd0 [vrx518_tc@8af9f5d0+0x25000]
  ...
  Unable to handle kernel NULL pointer dereference at virtual address 00000000
  pgd = aff5701e
  [00000000] *pgd=00000000
  Internal error: Oops: 5 [#1] SMP ARM

Noticed in ATM mode, when chip always returns byte_off = 4.

In order to fix the issue, pass the phys address to plat_mem_virt() as
is and apply byte_off later for proper DMA syncing and on mapped virtual
address when copying RXed data into the skb.

Run tested with FRITZ!Box 7530 on both ADSL and VDSL (thanks Jan) links.

Fixes: 474bbe23b7 ("kernel: add Intel/Lantiq VRX518 TC driver")
Tested-by: Jan Hoffmann <jan@3e8.eu> # VDSL link
Reported-and-tested-by: nebibigon93@yandex.ru # ADSL link
Signed-off-by: Sergey Ryazanov <ryazanov.s.a@gmail.com>
Link: https://patchwork.ozlabs.org/project/openwrt/patch/20250122222654.21833-2-ryazanov.s.a@gmail.com/
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 7bd579689d2304c73c263be3e030d76c551d6e87)

mediatek-mt7622: netgear-wax206 fix wifi leds

the wifi leds of the wax206 were not reacting.
This patch enables the green leds to show activity, as the blue ones are very bright.
Also set the label-mac to the gmac0

Signed-off-by: Florian Maurer <f.maurer@outlook.de>
Link: https://github.com/openwrt/openwrt/pull/17694
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 9ada8578fd2768d077607b189987972ad600a804)

realtek: Fix old compatible for HPE 1920-8G PoE

Supported devices are listed in the metadata as the first part of the
DTS compatible. This normally follows the format "vendor,device".

When updating the device name of the 180W 1920-8G PoE an underscore was
used, instead of a comma, to join the vendor and device name. This will
lead to warnings for users wanting to sysupgrade a device with an older
compatible, as the device's info does not match the one the metadata.

Fixes: 987c96e88927 ("realtek: rename hpe,1920-8g-poe to match hardware")
Signed-off-by: Sander Vanheule <sander@svanheule.net>
(cherry picked from commit 6a7fa68569ac11bb2c2abb14026e8b84acd3a12f)

qualcommax: add missing WAN LED support to Spectrum SAX1V1K routers

Fixed an issue where both WAN LEDs light up before plugging in the
ethernet cable and no blinking regardless of WAN network activity.

Updated the LED configuration to reflect proper status:
Green indicates 2.5Gb connection speed.
Yellow indicates other connection speed and traffic activity.

This resolves inconsistent WAN LED behavior on Spectrum SAX1V1K routers.

Signed-off-by: Ivan Deng <hongba@rocketmail.com>
Link: https://github.com/openwrt/openwrt/pull/17623
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 8e78bc39a38d2a82818bcae467d3bcbbbe8531f3)

mediatek: add Airoha AN8855 gigabit switch driver

New revisions of Xiaomi AX3000T with 1.0.84+ stock firmware contain new hardware.
This commit add support for Airoha AN8855 gigabit switch driver with 6.6 kernel patches

Based on https://patchwork.kernel.org/project/netdevbpf/cover/20241209134459.27110-1-ansuelsmth@gmail.com/

Signed-off-by: Dim Fish <dimfish@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/16709
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit 0fd9d00cd6fc285b2a925eb03e6350a4b00fc279)

OpenWrt v24.10.0-rc6: revert to branch defaults

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

OpenWrt v24.10.0-rc6: adjust config defaults

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

generic: backport pending gpio-regmap ops patch

This patch has been accepted for linux v6.14 so we can move it from pending
to backport.

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
(cherry picked from commit 2f2e21a52bdfdcf83868e3169768fda134aa883b)

generic: fix version number on realtek phy patches

These patches have been accepted in linux v6.14 instead of v6.13.

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
(cherry picked from commit 12a07e934c013c87ea96ddda8ec18c9af0362c53)

generic: net: phy: realtek: expose temperature sensors

Expose the temperature sensor built-into RTL822x 2.5G and 5G PHYs.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
(cherry picked from commit 3a189e38317e3390ba5b17523f55bd2fb1b72327)

generic: refresh hack patches

Refresh hack patches with make target/linux/refresh.

Fixes: 9508ca44eb57 ("kernel: backport improvement to page pool fragment handling from 6.7")
Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
(cherry picked from commit ee730a66e9c670ec2ee37ee2b207accd539e1d15)

kernel: backport improvement to page pool fragment handling from 6.7

Makes it easier to keep drivers like mt76 in sync with newer versions

Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry picked from commit 9508ca44eb577a13e795e0a6c5683689efc61475)
(cherry picked from commit 478041997f068c87c6356108cf3dc3b667666dd5)

generic: mtk_eth_soc: reduce driver memory usage

1. Import pending patch to fix ramips/mt7621 64MB targets.

2. Do not enable CONFIG_PAGE_POOL_STATS by default.

Signed-off-by: Danila Romanov <pervokur@gmail.com>
Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry picked from commit 15887235c1e3b732d4c8e14afac253d00a7bfb36)

kernel: bump 6.6 to 6.6.73

Changelog: https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.6.73

No patches needed a rebase.

Build system: x86/64
Build-tested: bcm27xx/bcm2712, flogic/xiaomi_redmi-router-ax6000-ubootmod, ramips/tplink_archer-a6-v3
Run-tested: bcm27xx/bcm2712, 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/17687
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 761b244964f8d5da9fbd27a618565b4c38ae0cbc)

ath79: meraki-mr18: fix initramfs build

Now, that initramfs images built for ZTE devices work, by moving
LZMA_TEXT_START further up the available RAM - same fix works
successfully for Meraki MR18 too. Apply it and reenable initramfs
generation again.

Fixes: 1d49310fdb5e ("ath79: add Cisco Meraki MR18")
Signed-off-by: Lech Perczak <lech.perczak@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17680
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 7423e67e1b8c7e4c9f527d1300dbeaab72b1f95f)

mediatek: Xiaomi AX3000t: fix soft brick for the rd23 model

This commit fixes Xiaomi AX3000t soft bricks. Issue affects at least rd23
model (Global version) users:
Link: https://forum.openwrt.org/t/openwrt-support-for-xiaomi-ax3000t/180490/452
I also found that these nvram settings are wrong for the rd23 model and
U-Boot erase them:
'''
flag_try_sys1_failed 8
flag_try_sys2_failed 8
'''
As a result, platform.sh -> xiaomi_initial_setup() function sometimes ends
early without applying settings for the rd23 model.

RD03 model strategy:
1. Don't touch values those were set up by platform.sh ->
xiaomi_initial_setup() function

RD23 model strategy:
1. Apply correct nvram settings at every boot
2. Use bulk fw_setenv call

I didn't find opened issue for AX3000t. Similar AX3200 issue:
Link: https://github.com/openwrt/openwrt/issues/16347
So, other Xiaomi devices (e.g. Xiaomi WR30U) may also require fix.

Fixes: 7dbcc1215a56cc1da68c81bf92d7efa174c5e051 ("mediatek: filogic: add support for Xiaomi AX3000T")
Signed-off-by: Mikhail Zhilkin <csharper2005@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17580
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 42a253c7e8992e2274628b1770a0263ac433d0bc)

mediatek: replace multiple fw_setenv calls by the bulk one

This commit replaces multiple fw_setenv calls by the bulk one to reduce
flash writes.

Thanks @Linaro1985 for the idea.

Signed-off-by: Mikhail Zhilkin <csharper2005@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17580
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit a6b0abf4661e4dc1449494e46340f75c44df81de)

kernel: bump 6.6 to 6.6.72

Changelog: https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.6.72

Dropped to due being superseded by upstream patch:
mediatek/patches-6.6/870-drm-mediatek-only-touch-DISP_REG_OVL_PITCH_MSB-if-AF.patch[1]

All other patches automatically rebased.

1. https://github.com/gregkh/linux/commit/ac7f5641e9887354a0fc6f48190533af084bb0dc

Build system: x86/64
Build-tested: bcm27xx/bcm2712, flogic/xiaomi_redmi-router-ax6000-ubootmod, ramips/tplink_archer-a6-v3
Run-tested: bcm27xx/bcm2712, 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/17646
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 1d021458fd681dc4c27615a574cc09cc57d1c51e)

lantiq: xrx200: fb7430 set correct label-mac

the CWMP account mac is correctly set on the lan device but was not correctly as label-mac

Signed-off-by: Florian Maurer <f.maurer@outlook.de>
Link: https://github.com/openwrt/openwrt/pull/17618
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit f9813f41b2e9fccb644ae149eabe575fbae12fa3)

lldpd: fix config for build without LLDP-MED

The lldp_class and lldp_location config option are only valid when
compiled with LLDP-MED support. If not they will cause lldpd not to
start.

Signed-off-by: Sebastian Pflieger <sebastian@pflieger.email>
Link: https://github.com/openwrt/openwrt/pull/17571
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 2c22d7c3a4a7edcce1af656c8cddb1ab163e3d02)

lldpd: reset PKG_RELEASE

lldpd was updated, so reset PKG_RELEASE after the PKG_VERSION update.

Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit abbec429b40f149f9c5a99a64ee7bf4e804fbb7d)

lldpd: bump version to 1.0.18

Changes (breaking):
- Remove support for building 802.3bt TLVs (broken).

Fix:
- Fix memory leaks in EDP/FDP decoding when receiving some TLVs twice.
- Do not set interface description continuously.
- Use a different Netlink socket for changes and queries.

Signed-off-by: Sebastian Pflieger <sebastian@pflieger.email>
Link: https://github.com/openwrt/openwrt/pull/17570
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit a18385041eaeaf6d98ab79a30ce5fba4e712b765)

d1: Use correct module for devices with wifi

Devices with wifi (LicheePi RV and MangoPi MQ Pro) were using the
wrong module.  Also wpad was missing to enable using the WiFi.

Signed-off-by: Raylynn Knight <rayknight@me.com>
Link: https://github.com/openwrt/openwrt/pull/17576
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
(cherry picked from commit 8a17bae722a62954269f29dc48025e88c919dc24)

uci: update to Git HEAD (2025-01-20)

047b2efc1348 CMakeLists.txt: bump minimum cmake version
16ff0badbde7 CMakeLists: add support for including ABIVERSION in the library version number

Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry picked from commit e046f8c318618162fcdd768c4a9eadd53c239629)

mediatek-filogic: fix wax220 wifi leds

The WAX220 does have a 2.4GHz and 5GHz wifi led, which was set to trigger on netdev before.
This commit changes this to trigger on activity of the respective radio

Signed-off-by: Florian Maurer <f.maurer@outlook.de>
Link: https://github.com/openwrt/openwrt/pull/17627
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 99431e315039d7f9689b21a8aa220adadf52a886)

ath79: zte-mf28x-common: fix initramfs execution

Now that LZMA_TEXT_START is configurable per-target once again,
move the target above 32MB boundary for ZTE MF28* devices.

Signed-off-by: Lech Perczak <lech.perczak@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17616
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 32c6b9064ac7fc62cead37b1bc7d46ffba598c71)

ath79: lzma-loader: expose LZMA_TEXT_ADDRESS configuration again

initramfs for some devices grew so big, that it can't be loaded within
the previous 32MB RAM limit. Make the LZMA_TEXT_ADDRESS configurable
per-target once again, to fix it for bigger devices, while maintaining
compatibility with previous ones.

Signed-off-by: Lech Perczak <lech.perczak@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/17616
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 6720958659d226e71dca49657202e0502bb059bd)

mediatek: EX5601-T0 add Zyxel EX5601-T1 and T-56 alias

Define EX5601-T1 and T-56 as alternative name, to explicitly show
the device is supported using existing image. EX5601-T1 does not
have the option to switch between WAN/SFP port. The switch port
is hardwired to the WAN port. The Zyxel T-56 is the odido-branded
version of the EX5601-T1.

Signed-off-by: Aleksander Jan Bajkowski <olek2@wp.pl>
Link: https://github.com/openwrt/openwrt/pull/17615
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 3d63a41ffab69cb6f49c14cff99c3c3a84899363)

kernel: fix rootfs initramfs not updating on subsequent single build

There is currently a problem where the rootfs in an initramfs image for
single target build is not updated on subsequent run. This is caused by
a bug introduced in d78dec3e19e3 ("kernel: copy kernel build dir on
Per-Device Initramfs compilation") where the initramfs_data.cpio rm was
moved to PrepareConfigPerRootfs. This caused the side effect of dropping
the rm call for single target build making the kernel reusing the
previous generated initramfs_data.cpio.

To correctly handle this, restore the original location of this call
right after the touch /init call. This way the kernel will always
regenerate the initramfs embedded rootfs ALSO for single target build.

Fixes: d78dec3e19e3 ("kernel: copy kernel build dir on Per-Device Initramfs compilation")
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit 334c649a8d9b191c0282f9104b3997b090628f56)

fritz-tools: increase PKG_RELEASE

Increase PKG_RELEASE as follow-up for
("lantiq: fritz_cal_extract with reverse option for AVM FritzBox 7430").

Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 4a8717b5e76edbc8e1f63122f8b05685e0b779db)

lantiq: fritz_cal_extract with reverse option for AVM FritzBox 7430

This implementation of fritz_cal_extract can also retrieve firmware
data stored in reverse byte order, as found in the AVM 7430 device.

This is done by intermediate storage in a buffer presumably large enough
to hold the complete data set. Currently, this buffer size is 128kB + 1kB
(some extra space for skipped data).

In the usual case of "forward" data, this implementation should behave
like the original implementation in all common cases. limit [-l] will
determine the amount of data read and size of buffer allocated.

However, if you are reading reversed data or didn't set a limit, the buffer
may be too small to hold all data. In this case, you can choose a higher
limit [-l] to enforce a sufficient buffer size.

Signed-off-by: Dustin Gathmann <dzsoftware@posteo.org>
Link: https://github.com/openwrt/openwrt/pull/15501
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit b2cac2a9785ada7c30388349f5d9a9bdcff2e730)

ath79: add extended AR9344 reset sequence

According to datasheet, on AR9344 the switch and switch analog need to
be reset first before initiating a full reset.

Resetting these systems fixes spurious reset hangs on Atheros AR9344
SoCs.

Link: https://github.com/freifunk-gluon/gluon/issues/2904
Signed-off-by: David Bauer <mail@david-bauer.net>
(cherry picked from commit 144af32b47be6181f01815762159dff863f4616b)

Revert "ath79: reset ETH switch for AR9344"

This reverts commit 916af73fc388d638b1a717a2411792e0680dd8e6.

It was reported this change did not fix the reboot issues on AR9344.

Signed-off-by: David Bauer <mail@david-bauer.net>
(cherry picked from commit 937b5a3e0017f4f39c0f7727a910d2df78fc4ce9)

ipq4019: patch initialization of Lantiq DSL on FritzBox 7530.

Some VRX518 modems fail to initialize properly with the error message
"dc_ep_clk_on failed". As a result, the DSL data path doesn't work.

This hack, which is based on code from the FRITZ!Box 7530 GPL archive,
fixes the issue. It changes the PCIe vendor/device ID to values matching
a Lantiq SoC. It also appears to emulate a Lantiq CPU ID register for
connected PCIe devices, by remapping the matching address area to a
specially crafted buffer using the address translation unit.

A dedicated compatible is created to activate this in
the device tree, so this shouldn't affect any devices other than
FRITZ!Box 7530/7520.

Original investigation was done in https://github.com/janh/openwrt/commit/59f52125178146a9dc44290c11c63e5e029e8044 which used the "avm,host_magic" property to enabled the patch.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
Signed-off-by: Florian Maurer <f.maurer@outlook.de>
Link: https://github.com/openwrt/openwrt/pull/17622
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 676dcb1b2cf84142a61d51a15e7a8e1bdc937c23)

ramips: mt7621: add support for Confiabits MT7621 v1

Hardware:
 - SoC: MediaTek MT7621DAT
 - Flash: 16 MiB cFeon
 - RAM: 128 MiB MT7621DAT
 - WLAN: 2.4 GHz (MT7603EN), 5 GHz (MT7613BEN 802.11ac)
 - Ethernet: 1x 10/100/1000 Mbps WAN, 3x 10/100/1000 LAN (MT7621DAT)
 - Buttons: 1 Reset button, 1 WPS button
 - LEDs: 5x Green (POWER/WAN/2.4G/5G/WPS), 1x Red (WAN)
 - Serial console: unpopulated header, 57600 8n1
 - Power: 12 VDC, 1 A

MAC:
LAN MAC: label mac (eeprom@4)
WAN MAC: label mac (eeprom@4)
2.4G MAC: label mac (eeprom@4)
5G MAC: label mac + 1 (eeprom@8004)

Installation:
The stock firmware is OpenWrt-based. If you can reach LuCI or SSH, just use the sysupgrade image
with the 'Keep settings' option turned off.

Signed-off-by: Luis Mita <luis@luismita.com>
Link: https://github.com/openwrt/openwrt/pull/17534
(cherry picked from commit 1c5ca24da6cd727f4fcb8edbe00fa66543a6c678)
Link: https://github.com/openwrt/openwrt/pull/17653
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

qualcommax: dl-wrx36: fix 2.5G port LED-s

Currently, 2.5G port LED-s on Dynalink are incorrectly configured and thus
they will light up all of the time.

So, lets fix this by:
1. Current green LED is actually yellow, change the color
2. Fix its polarity as its actually active-low
3. The yellow LED that was registered as being connected to LED_1 pin on
the PHY is not actually connected at all, so remove it.
4. The actual green LED is connected to LED_2 on the PHY so add it.

Fixes: 75ad5c24142a ("qualcommax: switch to qca8081 upstream PHY driver")
Fixes: #14502
Link: https://github.com/openwrt/openwrt/pull/17656
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 2d077913cd53cad33123c8a9bf6da8a1c5576c69)

generic: backport pending Realtek PHY patches

These patches have been accepted, so we can move them from pending to
backported.

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
(cherry picked from commit 7924acdd63490aa436bf114d290cb56cd4eba008)

generic: add pending patches for RealTek Ethernet PHYs

Import patches to fix several issues with status reporting of RealTek
2.5G PHYs.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
(cherry picked from commit 41c164d32bedb6b27ad1fd022dec6d25338db6d6)

generic: backport Realtek PHY patches from upstream

Adds patches for the temperature sensor on RTL822x.

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
(cherry picked from commit d7e82c78d7a2a84404198dab8faf8e142939eb05)

generic: r8169: remove reverted temperature sensor patch

The patch adding temperature sensor support for r8169 has been removed upstream
and the functionality will be added to Realtek PHY instead:
https://github.com/torvalds/linux/commit/1f691a1fc4bef1c5cf5f503e14e1a22fc37c97e3

Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>
(cherry picked from commit 0de9999a7819acf74f6adeaa27be363cf72409ab)

airoha: an7581: add pending PCI patch

Add pending PCI patch that should correctly fix mediatek driver with
Airoha SoC.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit f22febae1a6b7c654eaf09ed02c0d0017be6d73a)

airoha: an7581: backport ETS patch for Airoha ethernet

Backport ETS patch for Airoha ethernet and refresh affected patches.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit 67635578fd19d0576c54bbe722a9664e8a6ab2a6)

uci: update to Git HEAD (2025-01-17)

fb3c2343b17b add support for an override config directory

Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry picked from commit 44c877f1973b923a2032ce9d880f496a2b077ef6)

uci: update to Git HEAD (2024-11-26)

10f7996ec294 file: Ignore config file with '.' in name

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
(cherry picked from commit 3929ef4c21844189d5db8d14e7e6dc6ba983d571)

mediatek: remove obsolete base-files scripts

Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry picked from commit 48e1427ffa867bddc822a463af862b9dacc5bde8)

netifd: improve packet steering on ipq40xx (and possibly others)

Some platforms a single ethernet device for all ports with multiple rx rings
and NAPI threading enabled. In this case, the steering script was limiting
performance by keeping all NAPI threads assigned to the same CPU.
Fix this by assigning each rx queue and the corresponding NAPI task separately.
Additionally, if the number of rx queues is at least as big as the number of
CPUs, skip weight based assignment and distribute the load across all CPUs
directly.

Fixes: https://github.com/openwrt/openwrt/issues/17611
Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry picked from commit acce25b789d40f2d88a2326a7d9921bd1743fe54)

mvebu: WD Cloud Mirror Gen2: Remove stty dependency

The package coreutils-stty is not part of the OpenWrt main packages. A
board can not depend on it. Remove the dependency to fix a build problem
seen in the build bots.

Fixes: 8fb805aa1f25 ("mvebu: Add support for WD Cloud Mirror Gen2")
(cherry picked from commit 717f62d256ccb52549a24c256615b28440ed1a86)
Link: https://github.com/openwrt/openwrt/pull/17589
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

mvebu: Add support for WD Cloud Mirror Gen2

Hardware
--------
Marvell Armada 385 (MV88F6820)
512B RAM
256MB NAND (Hynix H27U2G8F2CTR)
1x 1Gbit
2x USB 3.0
2x SATA-III
UART: 115200 8N1 3.3V
RTC
Weltrend MCU WT6703F connected via UART1 for Power LED / PWM Fan / hw reset / WoL

Installation
------------
Connect UART 3.3V adapter to JP2  pins: 1-RX / 2-GND / 5-TX

Use USB2.0 FAT32 pendrive with openwrt-mvebu-cortexa9-wd_cloud-mirror-gen2-initramfs-kernel.bin
 1. stop boot by pressing 1
 2. usb start
 3. fatload usb 0:1  0x02000000 openwrt-mvebu-cortexa9-wd_cloud-mirror-gen2-initramfs-kernel.bin;bootm 0x02000000 -
 4. do backup mtd1 mtd3
 5. use sysupgrade

Or tftp
 1. stop boot by pressing 1
 2. setenv ethact egiga2;setenv serverip 192.168.11.114;setenv ipaddr 192.168.11.113
 3. tftpboot 0x02000000 openwrt-mvebu-cortexa9-wd_cloud-mirror-gen2-initramfs-kernel.bin; bootm 0x02000000 -
 4. do backup mtd1 mtd3
 5. use sysupgrade

or Evgeny Kolesnikov <evgenyz@gmail.com> method from his failed PR 2040

- Using original firmware's network settings obtain SSH access to the device.
- Put *-image-cfs-factory.bin and *-uImage-factory.bin images into device's /tmp directory.
- Write kernel (uImage) image 'flash_eraseall /dev/mtd1 && nandwrite --markbad -p /dev/mtd1 /tmp/*-uImage-factory.bin'.
- Write rootfs (image-cfs) image 'ubiformat /dev/mtd3 -f /tmp/*-image.cfs-factory.bin -y'.
- Reboot the device.

Installation (upgrade):

Use *-sysupgrade.bin in a usual way.
Weltrend MCU control is done via uart1 19200

stty -F /dev/ttyS1 raw speed 19200
stty -F /dev/ttyS1 raw speed 19200

PWM Fan Control
off: 00
echo -n -e '\xfa\x02\x00\x00\x00\x00\xfb' > /dev/ttyS1
slow: 5F
echo -n -e '\xfa\x02\x00\x5f\x00\x00\xfb' > /dev/ttyS1
max: FF
echo -n -e '\xfa\x02\x00\xff\x00\x00\xfb' > /dev/ttyS1

Power LED Control
Blue
echo -n -e '\xfa\x26\x00\x11\x00\x01\xfb' > /dev/ttyS1
Red
echo -n -e '\xfa\x26\x00\x14\x00\x01\xfb' > /dev/ttyS1
Orange
echo -n -e '\xfa\x26\x00\x12\x00\x01\xfb' > /dev/ttyS1

more here: https://github.com/c-MM/mcm-daemon/blob/master/mcm.h

Signed-off-by: Robert Senderek <robert.senderek@10g.pl>
Link: https://github.com/openwrt/openwrt/pull/17046
(cherry picked from commit 8fb805aa1f25b69b7849c8b1ce787ed47b16d123)
Link: https://github.com/openwrt/openwrt/pull/17589
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

armsr: change image names to 'combined-efi'

luci-app-attendedsysupgrade expects images to be named
'combined-efi' when the system is using EFI images.

This came about as x86 has 'combined' images for legacy
(BIOS) boot and 'combined-efi' for EFI systems.

armsr images were originally named 'combined' only
as there was no 'legacy' image type.

To avoid special handling in the attendedsysupgrade
code, name EFI images consistent with other targets.

Signed-off-by: Mathew McBride <matt@traverse.com.au>
Link: https://github.com/openwrt/luci/pull/6430
Link: https://github.com/openwrt/openwrt/pull/12963
Signed-off-by: Petr Štetiar <ynezz@true.cz>
(cherry picked from commit c9ebd4fb30773ca55cc2da6b2c1962046a2e0fff)

netifd: fix napi process name matching in packet steering script

Fixes CPU usage imbalance on some devices using threaded NAPI

Signed-off-by: Felix Fietkau <nbd@nbd.name>
(cherry picked from commit 67a4aeef6c63e5836bd6b82faf6cc15934869efb)

wifi-scripts: allow per-IF mesh basic rate selection

Bringing up a mesh interface using wpa_supplicant already supports a
per-VIF basic rate selection. Add the same ability when creating a mesh
VIF without wpa_supplicant.

Signed-off-by: David Bauer <mail@david-bauer.net>
(cherry picked from commit 3deeb7805f69f54b9ab20c0e12f8dd916fec4de1)