generic: 6.12: add new config symbols

Add new config symbols to 6.12 kernel config and move common symbols from x86.

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

qualcommax: use upstream QMP pipe patch

Passing QMP pipe clocks was upstreamed, but it requires conversion single
QMP node, so backport both of those instead of our downstream patch.

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

generic: 6.12: add Qualcomm PBS symbol

Qualcomm PBS symbols popped up during 6.12 work.

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

generic: 6.12: add Qualcomm PM8916 symbols

Qualcomm PM8916 PMIC symbols popped up while doing 6.12 for qualcommax.

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

qualcommax: renumber upstreamed patches

In order to free up some more number space, simply renumber existing
upstreamed patches starting from 0 again.

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

qualcommax: ipq50xx: remove ECC user config from board files

With the spi-qpic-snand driver now supporting reading ECC requirements
from the NAND chip itself, there's no need to set those in the DTS
anymore avoiding issues with devices of the same revision using
different NAND types with varying ECC requirements.

Signed-off-by: George Moussalem <george.moussalem@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/18723
Signed-off-by: Robert Marko <robimarko@gmail.com>

spi: spi-qpic-snand: update driver with latest patches from upstream

Updating the spi-qpic-snand driver with the latest patches sent
upstream and remove custom patches.
These patches add error handling fixes, use of right read location in
read ops, removal of unused variable, ability to read user config and
nand chip requirements, and support for 8bits ECC strength.

Tested on: Gl.iNet GL-B3000 & Linksys MR5500, MX2000, MX5500, SPNMX56

Signed-off-by: George Moussalem <george.moussalem@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/18723
Signed-off-by: Robert Marko <robimarko@gmail.com>

qualcommax: use upstreamed multiple conf clock patches

Multiple conf support was upstreamed into kernel 6.10, so lets use the
upstreamed patches and mark them as so.

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

ath11k-firmware: fix legacy FW tarball hash

The legacy FW tarball hash is wrong, fix it.

Fixes: ea83f7de2b02 ("ath11k-firmware: add wifi firmware for IPQ5018")
Signed-off-by: Robert Marko <robimarko@gmail.com>

qualcommax: update kernel version for MP5496 patches

MP5496 patches are not part of 6.15, but rather will be part of 6.16 as
they are no in 6.15 RC releases but they are in linux-next.

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

mac80211: ath11k: fix broadcast failures during GTK rekeying

Revert key_cipher change to NONE when clearing key.
Setting to NONE triggers asserts in the firmware and causes broadcast
drops with ath11k firmware 2.9.0.1, so until a future blob
fixes this, revert this as a workaround.

Imported from
https://git.codelinaro.org/clo/qsdk/oss/system/feeds/wlan-open/-/blob/win.wlan_host_opensource.3.0.r24/patches/ath11k/350-ath11k-Revert-clear-the-keys-properly-when-DISABLE_K.patch

Related discussion links:

- https://github.com/openwrt/openwrt/issues/9555
- https://lore.kernel.org/linux-wireless/Z2Q9POuV-6MIdzRf@pilgrim/T/#t

Signed-off-by: Henry Tung <henryptung@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18705
Signed-off-by: Robert Marko <robimarko@gmail.com>

realtek: rtl838x: fix broadcast flooding with many multicast entries

When many multicast entries are installed broadcast flooding might
potentially stop working for several ports. This is because the layer
2 broadcast flood port mask index has the wrong offset. It should be
9 bits, matching the 2^9 = 512 indexes on rtl838x, not 12.

The wrong offset leads to L2_BC_FLD_PMSK being set to 504, not 511
((511 << 12) >> 9) & 511 = 504). So, as by default an unset PMSK
is set to all ports, the issue would only become noticeable once
many multicast entries are installed, causing the 504th entry to be set
to something other than all ports.

Fixing this by setting the offset to 9 bits, to correctly point to our
511th reserved entry for all ports.

Tested-on: ZyXEL GS1900-24HP v1

Fixes: 28e972b2ea2f ("realtek: Configure initial L2 learning setup")
Signed-off-by: Linus Lüssing <linus.luessing@c0d3.blue>
Link: https://github.com/openwrt/openwrt/pull/18733
Signed-off-by: Robert Marko <robimarko@gmail.com>

ucode-mod-uline: re-introduce polling loop to fix prompt after line end

Introduce a check for the resource data in order to avoid segfaulting
on exit

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

ath10k-ct: update to latest version 6.14 from 2025-04-30

In this new version channel 177 is supported.

Dropped patches:
[1] 001-patch-version.patch
[2] 205-ath10k-ct-silence-warning-caused-by-unsupported-retr.patch

[1] [2] https://github.com/greearb/ath10k-ct/commit/8adb310f84b0a90df1f65701b863321e7b2afecd

Signed-off-by: Paweł Owoc <frut3k7@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18731
Signed-off-by: Robert Marko <robimarko@gmail.com>

kernel: bump 6.12 to 6.12.27

Changelog: https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.12.27

No patches update required.

Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/18728
Signed-off-by: Robert Marko <robimarko@gmail.com>

qualcommax: ipq50xx: gl-b3000: add 03_set_oem_name

This additional board.d script creates the "/tmp/sysinfo/oem_name" file
and populates it with the proper OEM "supported_devices" metadata entry
to be used by the new fwtool.sh oem detection feature.

(#18554): add oem image dectection to fwtool.sh

Signed-off-by: Scott Mercer <TheRootEd24@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18656
Signed-off-by: Robert Marko <robimarko@gmail.com>

generic: 6.12: update block NVMEM driver

Update block NVMEM driver based on backported OF partition support.

Note the different MMC card DT binding compared to the previous
downstream solution: Instead of having a 'partitions' subnode inside
a 'block' node, the 'partitions' node now resides directly under the
node representing the card.

In order to make references to the 'boot0' or 'boot1' hw partitions
you will have to define 'partitions-boot0' or 'partitions-boot1', and
move the NVMEM layout into a partition (you may, of course, use
'fixed-partitions' for that, and cover the whole device, if needed).

See also https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/tree/Documentation/devicetree/bindings/mmc/mmc-card.yaml?h=v6.13

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

generic: 6.12: update fitblk driver for Linux 6.12

Update the uImage.FIT filesystem subimage block driver for use
with Linux 6.12.

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

ramips: update drivers to be compatible with kernel 6.12

Update drivers to be compatible with kernel 6.12:
 - change deprecate function strlcpy to strscpy
 - fix kernel warning "no previous prototype", by static declaration
 - add missing of.h headers

Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/18711
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

mac80211: rt2x00: use remove_new

Easy compability fix for kernel 6.12.

Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/18719
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

iw: fix WiFi 6 (HE) Scan on big endian platforms

On big endian platforms, scans show invalid info for WiFi 6 APs.
This commit backports patches to fix this issue.

Signed-off-by: Aleksander Jan Bajkowski <olek2@wp.pl>
Link: https://github.com/openwrt/openwrt/pull/18717
Signed-off-by: Robert Marko <robimarko@gmail.com>

bcm27xx: bcm2712: support all devices

- RPi 500
- RPi 5 Compute Module
- RPi 5B (bcm2712 d0 rev)

Signed-off-by: Dave Marquard <dave-atx@users.noreply.github.com>
Link: https://github.com/openwrt/openwrt/pull/18622
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>

linux-firmware: add Raspberry Pi 5 CM NVRAM

The RPi 5 Compute Module expects the same NVRAM as the one from RPi 4
on a different file.

Signed-off-by: Dave Marquard <dave-atx@users.noreply.github.com>
Link: https://github.com/openwrt/openwrt/pull/18722
Signed-off-by: Álvaro Fernández Rojas <noltari@gmail.com>

toolchain/musl: fix build regression on x86_64

Fix whitespace mangling which broke matching opcodes in the CFI patch

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

generic: fix MIPS -mno-long-calls patchset for 6.12 kernel

It seems that we need to override all execmem alloc/free callbacks,
not just for module. This is the default patch behavior in the 6.6
kernel. Fix the warning when the module is loaded:

root@OpenWrt:~# insmod mtd-rw i_want_a_brick=1
[ 1404.954813] mtd-rw: mtd0: setting writeable flag
[ 1404.959643] mtd-rw: mtd5: setting writeable flag
[ 1404.966396] ------------[ cut here ]------------
[ 1404.971119] WARNING: CPU: 0 PID: 8 at mm/vmalloc.c:3361 vfree+0x1ac/0x2c4
[ 1404.978146] Trying to vfree() nonexistent vm area (74cc6c73)
[ 1404.983901] Modules linked in: mtd_rw(O) ath9k(O) ath9k_common(O)...
[ 1405.043696] CPU: 0 UID: 0 PID: 8 Comm: kworker/0:1 Tainted: G        W  O       6.12.25 #0
[ 1405.052118] Tainted: [W]=WARN, [O]=OOT_MODULE
[ 1405.056536] Hardware name:
[ 1405.061222] Workqueue: events do_free_init
[ 1405.065408] Stack : 807865d8 80850000 81823f80 81857df8 00000000 00000d21 81823fd0 800ca130
[ 1405.073924]         81839e48 807865d8 808d20bf 807865d8 81857d1c 00000001 81857ce8 4951640f
[ 1405.082434]         00000000 00000000 807865d8 81857bf8 ffffefff 00000000 ffffffea 00000b5d
[ 1405.090944]         81857c04 00000b5d 808537b0 ffffffff 00000001 00000000 807865d8 81857df8
[ 1405.099453]         00000000 00000d21 81823fd0 8085119c 00000018 803f4828 00000000 80a00000
[ 1405.107963]         ...
[ 1405.110454] Call Trace:
[ 1405.112935] [<80066910>] show_stack+0x28/0xf0
[ 1405.117392] [<8069f340>] dump_stack_lvl+0x48/0x7c
[ 1405.122186] [<80084ab8>] __warn+0x9c/0x118
[ 1405.126357] [<80084bc0>] warn_slowpath_fmt+0x8c/0xac
[ 1405.131399] [<801f02ec>] vfree+0x1ac/0x2c4
[ 1405.135570] [<800dde54>] do_free_init+0x50/0x84
[ 1405.140172] [<8009f4a0>] process_one_work+0x1b0/0x3dc
[ 1405.145312] [<800a022c>] worker_thread+0x308/0x478
[ 1405.150178] [<800a81c0>] kthread+0xf4/0x11c
[ 1405.154455] [<80061b58>] ret_from_kernel_thread+0x14/0x1c
[ 1405.159938]
[ 1405.161772] ---[ end trace 0000000000000000 ]---

Fixes: a9c0f28951fd ("generic: 6.12: move MIPS reloc patch from pending to hack and rework")
Ref: https://lore.kernel.org/all/20240505160628.2323363-1-rppt@kernel.org/
Signed-off-by: Shiji Yang <yangshiji66@outlook.com>
Tested-by: Tony Ambardar <itugrok@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/18721
Signed-off-by: Robert Marko <robimarko@gmail.com>

imx: coretexa53: add network config for single-port Gateworks Venice boards

Add network config for single-port Gateworks venice boards such that the
ethernet port is the WAN port instead of the default being a LAN port.

Signed-off-by: Tim Harvey <tharvey@gateworks.com>
Link: https://github.com/openwrt/openwrt/pull/18629
Signed-off-by: Robert Marko <robimarko@gmail.com>

realtek: fix cpu port link type

Some DTS files have a qsgmii link mode for the CPU port. This does
not harm but it is wrong. The CPU port of the realtek switch is always
directly connected to the switch by some unknown wiring and should
therefore be described as internal. Align the wrongly defined DTS
files to the standard.

Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
Link: https://github.com/openwrt/openwrt/pull/18691
Signed-off-by: Robert Marko <robimarko@gmail.com>

generic: 6.12: sync Realtek PHY patches with upstream

- Fix order of patches (3d483a10327f was merged before 34d5a86ff7bb).
- Reorganize patch numbers now that < 6.12 patches are no longer needed.

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

generic: 6.12: backport hwmon_ops static visibility

This patch was added on linux v6.13 in order to avoid simple is_visible
functions on hwmon drivers.
See https://github.com/torvalds/linux/commit/79bc0af904db647979c735563299c9b0d820e432

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

generic: 6.12: sync r8169 patches with upstream

Upstream e2015942e90a couldn't be backported to 6.6 because the following
symbols were missing:
- disable_work
- disable_work_sync
- enable_work
See https://github.com/torvalds/linux/commit/e2015942e90a021151a5751776f35830ba063be7

Upstream 34e5ededf4b8 couldn't be backported to 6.6 because the following
symbol was missing:
- pcim_iomap_region
See https://github.com/torvalds/linux/commit/34e5ededf4b8ad4c9e58f0cab8596e26c8fa59a2

Reorganize patch numbers now that < 6.12 patches are no longer needed.

The following patches still differ from upstream:
- e340bff27e63
  phy_set_eee_broken symbol is missing in 6.12
  https://github.com/torvalds/linux/commit/e340bff27e63

The following patches can't be backported to 6.12 due to missing symbols:
- 5e7a74b6a357
  phy_disable_eee_mode symbol is missing in 6.12
  https://github.com/torvalds/linux/commit/5e7a74b6a357

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

unetmsg: add subscriber update callback to notify about publish events

When services start publishing on a topic, this can be used to allow
subscribers to query them.

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

unetmsg: fix issuing requests without previously subscribing

Perform ACL check if not subscribed. Helps with one-shot requests.

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

gdb: update to version 16.2

Sync with toolchain/gdb version

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

musl: add pending patch to fix CFI annotations

Fixes debugging at least on aarch64, possibly others.

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

libubox: update to Git HEAD (2024-12-19)

3868f47c8f6c blob: constify attr argument to blob_memdup

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

kernel: iio: iio-core: update dependencies for 6.12

With kernel 6.11 iio-core gained a dependency to DMA_SHARED_BUFFER,
so add the appropriate dependency.

Signed-off-by: Jonas Gorski <jonas.gorski@gmail.com>

kernel: bluetooth: btmtk: update dependencies for 6.12

With kernel 6.11 several mtk usb functions were moved from btusb to
btmtk, adding a usb core depdendency to btmtk (if enabled).

Signed-off-by: Jonas Gorski <jonas.gorski@gmail.com>

tools/b43-tools: update to latest version, fix C23 build

Update b43-tools to the latest upstream version to fix a
compilation error with C23:

  util.h:25:15: error: 'bool' cannot be defined via 'typedef'
     25 | typedef _Bool bool;
        |               ^~~~

Changelog:
c6fc53f replace custom bool typedef with <stdbool.h>
dadf30c fix format warning in compilation
2fe10ea b43-fwdump: Fix forwarding of arguments to disassembler

Link: https://github.com/openwrt/openwrt/pull/18708
Signed-off-by: Nick Hainke <vincent@systemli.org>

kernel: bump 6.12 to 6.12.26

Changelog: https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.12.26

All patches automatically rebased.

Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/18695
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

libiconv-full: update to 1.18

libiconv 1.17 fails to compile with GCC15 when doing the host build,
luckily version 1.18 already contains the fixes required so update to it.

New in 1.18:
* Many more transliterations, in particular also of Emoji characters.
* The iconv_open function is now POSIX:2024 compliant: it recognizes a
  suffix //NON_IDENTICAL_DISCARD in the 'tocode' argument, with the effect
  that characters that cannot be represented in the target character set
  will be silently discarded. Whereas the suffix //IGNORE in the 'tocode'
  argument has the effect of discarding not only characters that cannot be
  represented in the target character set, but also invalid multibyte
  sequences in the input.
  Accordingly, the iconvctl function accepts requests
    ICONV_GET_DISCARD_INVALID, ICONV_SET_DISCARD_INVALID,
    ICONV_GET_DISCARD_NON_IDENTICAL, ICONV_SET_DISCARD_NON_IDENTICAL.
* The iconv_open function and the iconv program now support multiple suffixes,
  such as //TRANSLIT//IGNORE, not only one.
* GB18030 is now an alias for GB18030:2005. A new converter for GB18030:2022
  is added. Since this encoding merely cleans up a few private-use-area
  mappings, you can continue to use the GB18030 converter, for backward
  compatibility. Its Unicode to GB18030 conversion direction has been
  enhanced, to help transitioning away from PUA code points.
* When converting from/to an EBCDIC encoding, a non-standard way of
  converting newlines can be requested
    - at the C level, by calling iconvctl with argument ICONV_SET_FROM_SURFACE
      or ICONV_SET_TO_SURFACE, or
    - from the iconv program, by setting the environment variable
      ICONV_EBCDIC_ZOS_UNIX to a non-empty value.
* Special support for z/OS: The iconv program adds a charset metadata tag to
  its output file. (Contributed by Mike Fulton.)
* For conversions from UCS-2, UCS-4, UTF-16, UTF-32, invoking
  iconv(cd,NULL,NULL,...) now preserves the byte order state.

Link: https://github.com/openwrt/openwrt/pull/18698
Signed-off-by: Robert Marko <robimarko@gmail.com>

libiconv-full: drop patch-libtool fixup

It seems that this was required many years ago (It was added in 2012),
probably due to relying on outdated or broken libtool versions, but 1.17
uses libtool 2.4.7.

After trying building both host and target libiconv versions on Fedora and
MacOS I cannot seem to find what patch-libtool is supposed to fixup.

So, I can only presume that it is completely uneeded these days and can
be dropped.

Link: https://github.com/openwrt/openwrt/pull/18698
Signed-off-by: Robert Marko <robimarko@gmail.com>

tools: add options to optimize host binaries

Mains goals are:
- reduce binary size of host tools;
- reduce i/o load on build host;
- increase performance of host tools being built.

Signed-off-by: Konstantin Demin <rockdrilla@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18659
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>

dns320l-mcu: fix compilation with GCC14

Bump version to fix compiling with GCC14.

This fixes the following compile problem:
```
dns320l-daemon.c: In function 'main':
dns320l-daemon.c:740:18: error: implicit declaration of function 'isprint' [-Wimplicit-function-declaration]
  740 |         else if (isprint (optopt))
      |                  ^~~~~~~
dns320l-daemon.c:50:1: note: include '<ctype.h>' or provide a declaration of 'isprint'
   49 | #include "dns320l-daemon.h"
  +++ |+#include <ctype.h>
   50 |
dns320l-daemon.c:799:5: error: implicit declaration of function 'umask' [-Wimplicit-function-declaration]
  799 |     umask(0);
      |     ^~~~~
dns320l-daemon.c:864:5: error: 'return' with no value, in function returning non-void [-Wreturn-mismatch]
  864 |     return;
      |     ^~~~~~
dns320l-daemon.c:691:5: note: declared here
  691 | int main(int argc, char *argv[])
      |     ^~~~
```

Link: https://github.com/openwrt/openwrt/pull/18688
Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>

tools: libtool: update to 2.5.4

Update libtool to the current stable 2.5.4 release.

130-trailingslash.patch was upstream, 100 and 110 patches were
manually refreshed while rest was automatically refreshed via quilt.

Link: https://github.com/openwrt/openwrt/pull/18655
Signed-off-by: Robert Marko <robimarko@gmail.com>

realtek: add missing symbol

Commit d7e82c78d7a2 added a generic kernel patch that exposes a new
symbol REALTEK_PHY_HWMON when REALTEK_PHY and HWMON are enabled. The new
symbol was added to kmod-phy-realtek, but the kmod is not used in the
realtek target.

Fixes: d7e82c78d7a2 ("generic: backport Realtek PHY patches from upstream")
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>

toolchain: gcc: Backport patch to fix mips16 linking

Backport patch from upstream GCC 14 branch which fixes linking with
MIPS16 on the pistachio target.

This fixes the following link problem:
```
/builder/shared-workdir/build/staging_dir/toolchain-mipsel_24kc+24kf_gcc-14.2.0_musl/lib/gcc/mipsel-openwrt-linux-musl/14.2.0/../../../../mipsel-openwrt-linux-musl/bin/ld.bfd: ./liblua.so: undefined reference to `__mips16_ledf2'
/builder/shared-workdir/build/staging_dir/toolchain-mipsel_24kc+24kf_gcc-14.2.0_musl/lib/gcc/mipsel-openwrt-linux-musl/14.2.0/../../../../mipsel-openwrt-linux-musl/bin/ld.bfd: ./liblua.so: undefined reference to `__mips16_call_stub_df_2'
/builder/shared-workdir/build/staging_dir/toolchain-mipsel_24kc+24kf_gcc-14.2.0_musl/lib/gcc/mipsel-openwrt-linux-musl/14.2.0/../../../../mipsel-openwrt-linux-musl/bin/ld.bfd: ./liblua.so: undefined reference to `__mips16_muldf3'
```

Link: https://github.com/openwrt/openwrt/pull/18688
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

toolchain: gcc: Refresh patches

Refresh all GCC patches.

Link: https://github.com/openwrt/openwrt/pull/18688
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

dns320l-mcu: Fix compilation with GCC 14

This fixes the following compile problem:
```
dns320l-daemon.c: In function 'main':
dns320l-daemon.c:740:18: error: implicit declaration of function 'isprint' [-Wimplicit-function-declaration]
  740 |         else if (isprint (optopt))
      |                  ^~~~~~~
dns320l-daemon.c:50:1: note: include '<ctype.h>' or provide a declaration of 'isprint'
   49 | #include "dns320l-daemon.h"
  +++ |+#include <ctype.h>
   50 |
dns320l-daemon.c:799:5: error: implicit declaration of function 'umask' [-Wimplicit-function-declaration]
  799 |     umask(0);
      |     ^~~~~
dns320l-daemon.c:864:5: error: 'return' with no value, in function returning non-void [-Wreturn-mismatch]
  864 |     return;
      |     ^~~~~~
dns320l-daemon.c:691:5: note: declared here
  691 | int main(int argc, char *argv[])
      |     ^~~~
```

Link: https://github.com/openwrt/openwrt/pull/18688
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

boot-lantiq: Fix compilation with GCC 14

This fixes the following compile problem:
```
arv7506pw11.c: In function 'show_boot_progress':
arv7506pw11.c:59:24: error: 'return' with a value, in function returning void [-Wreturn-mismatch]
   59 |                 return 0;
      |                        ^
arv7506pw11.c:56:6: note: declared here
   56 | void show_boot_progress(int arg)
      |      ^~~~~~~~~~~~~~~~~~
arv7506pw11.c:71:16: error: 'return' with a value, in function returning void [-Wreturn-mismatch]
   71 |         return 0;
      |                ^
arv7506pw11.c:56:6: note: declared here
   56 | void show_boot_progress(int arg)
      |      ^~~~~~~~~~~~~~~~~~
```

Link: https://github.com/openwrt/openwrt/pull/18688
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

ltq-adsl-app: Fix compilation with GCC 14

This fixes the following compile problem:
```
checking for asm/types.h... dsl_cpe_linux.c: In function 'DSL_CPE_ThreadInit':
dsl_cpe_linux.c:779:25: error: assignment to 'DSL_CPE_Thread_t' {aka 'int'} from 'pthread_t' {aka 'struct __pthread *'} makes integer from pointer without a cast [-Wint-conversion]
  779 |          pThrCntrl->tid = tid;
      |                         ^
dsl_cpe_linux.c: In function 'DSL_CPE_ThreadDelete':
dsl_cpe_linux.c:862:44: error: passing argument 1 of 'pthread_cancel' makes pointer from integer without a cast [-Wint-conversion]
  862 |             switch(pthread_cancel(pThrCntrl->tid))
      |                                   ~~~~~~~~~^~~~~
      |                                            |
      |                                            DSL_CPE_Thread_t {aka int}
In file included from dsl_cpe_linux.h:35:
/builder/shared-workdir/build/staging_dir/toolchain-mips_mips32_gcc-14.2.0_musl/include/pthread.h:98:20: note: expected 'pthread_t' {aka 'struct __pthread *'} but argument is of type 'DSL_CPE_Thread_t' {aka 'int'}
   98 | int pthread_cancel(pthread_t);
      |                    ^~~~~~~~~
dsl_cpe_linux.c: In function 'DSL_CPE_ThreadIdGet':
dsl_cpe_linux.c:1123:11: error: returning 'pthread_t' {aka 'struct __pthread *'} from a function with return type 'DSL_CPE_Thread_t' {aka 'int'} makes integer from pointer without a cast [-Wint-conversion]
 1123 |    return pthread_self();
      |           ^~~~~~~~~~~~~~
```

While at it, fix also some additional build warnings.

Link: https://github.com/openwrt/openwrt/pull/18688
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

ramips: i2s: use devm for debugfs seqfile

Avoids some pointless boilerplate.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18675
Signed-off-by: Robert Marko <robimarko@gmail.com>

ramips: i2s: replace spaces with tabs

This seems to be some manual patch editing or rebasing bug.

Found with grep ' '.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18675
Signed-off-by: Robert Marko <robimarko@gmail.com>

mac80211: ath9k: fix missing prototypes warnings

Fix following warnings by including the corresponding headers:

backports-6.12.6/drivers/net/wireless/ath/ath9k/hsr.c:50:6: error: no previous prototype for 'ath9k_hsr_init' [-Werror=missing-prototypes]
   50 | void ath9k_hsr_init(struct ath_hw *ah)
      |      ^~~~~~~~~~~~~~
backports-6.12.6/drivers/net/wireless/ath/ath9k/hsr.c:165:5: error: no previous prototype for 'ath9k_hsr_disable' [-Werror=missing-prototypes]
  165 | int ath9k_hsr_disable(struct ath_hw *ah)
      |     ^~~~~~~~~~~~~~~~~
backports-6.12.6/drivers/net/wireless/ath/ath9k/hsr.c:177:5: error: no previous prototype for 'ath9k_hsr_enable' [-Werror=missing-prototypes]
  177 | int ath9k_hsr_enable(struct ath_hw *ah, int bw, int fq)
      |     ^~~~~~~~~~~~~~~~
backports-6.12.6/drivers/net/wireless/ath/ath9k/hsr.c:234:5: error: no previous prototype for 'ath9k_hsr_status' [-Werror=missing-prototypes]
  234 | int ath9k_hsr_status(struct ath_hw *ah)
      |     ^~~~~~~~~~~~~~~~

Signed-off-by: Shiji Yang <yangshiji66@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/18637
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

rtl8812au-ct: fix build issue on 6.12 kernel

Update USB shutdown callback for the 6.12 kernel. A sprintf
overlaps issue is also fixed in this patch. Unfortunately,
there are dozens of missing-prototypes warnings so it's hard
to fix them one by one. Let's just silence them.

Signed-off-by: Shiji Yang <yangshiji66@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/18637
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

ath10k-ct: fix build warnings on 6.12 kernel

This patch fixes a lot of missing-prototypes warnings for the
upcoming 6.12 kernel.

Signed-off-by: Shiji Yang <yangshiji66@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/18637
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

mt76: fix build error on 6.12 kernel

Convert platform driver .remove to .remove_new to fix the
incompatible pointer type error on 6.12 kernel.

Signed-off-by: Shiji Yang <yangshiji66@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/18637
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

mac80211: fix unaligned.h header location for 6.12 kernel

Fix mt76 build error on 6.12 kernel:

In file included from /workspaces/openwrt/build_dir/target-x86_64_musl/linux-x86_64/mt76-2025.04.11~be28ef77/mt76x0/eeprom.c:13:
/workspaces/openwrt/staging_dir/target-x86_64_musl/usr/include/mac80211-backport/asm/unaligned.h:3:15: fatal error: asm/unaligned.h: No such file or directory
    3 | #include_next <asm/unaligned.h>
      |               ^~~~~~~~~~~~~~~~~

Signed-off-by: Shiji Yang <yangshiji66@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/18637
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

mac80211: rt2x00: fix build warnings on 6.12 kernel

Move rt2x00lib_read_eeprom() function prototype from rt2800lib.h to
rt2x00.h and make rt6352_enable_pa_pin() static to fix the missing
prototypes warnings:

/workspaces/openwrt/build_dir/target-x86_64_musl/linux-x86_64/mac80211-regular/backports-6.12.6/drivers/net/wireless/ralink/rt2x00/rt2x00eeprom.c:213:5: error: no previous prototype for 'rt2x00lib_read_eeprom' [-Werror=missing-prototypes]
  213 | int rt2x00lib_read_eeprom(struct rt2x00_dev *rt2x00dev)
      |     ^~~~~~~~~~~~~~~~~~~~~
/workspaces/openwrt/build_dir/target-x86_64_musl/linux-x86_64/mac80211-regular/backports-6.12.6/drivers/net/wireless/ralink/rt2x00/rt2800lib.c:308:6: error: no previous prototype for 'rt6352_enable_pa_pin' [-Werror=missing-prototypes]
  308 | void rt6352_enable_pa_pin(struct rt2x00_dev *rt2x00dev, int enable)
      |      ^~~~~~~~~~~~~~~~~~~~

Signed-off-by: Shiji Yang <yangshiji66@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/18637
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel: trelay: fix build warning on 6.12 kernel

Make trelay_handle_frame() static to fix the build warning:

/workspaces/openwrt/build_dir/target-x86_64_musl/linux-x86_64/trelay/trelay.c:38:21: error: no previous prototype for 'trelay_handle_frame' [-Werror=missing-prototypes]
   38 | rx_handler_result_t trelay_handle_frame(struct sk_buff **pskb)
      |                     ^~~~~~~~~~~~~~~~~~~

Signed-off-by: Shiji Yang <yangshiji66@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/18637
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

arm-trusted-firmware-stm32: rework patch for the RTC configuration

A patch was added upstream to temporary enable RTC clock configuration only
for STM32MP15 boards. Use this patch instead of reverting commit 03a581e2.
Now for STM32MP135 boards RTC clock configuration is handled by optee-os.

Signed-off-by: Thomas Richard <thomas.richard@bootlin.com>
Link: https://github.com/openwrt/openwrt/pull/18628
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

optee-os-stm32: bump to 4.6.0

Bump optee-os-stm32 to upstream release 4.6.0.

Signed-off-by: Thomas Richard <thomas.richard@bootlin.com>
Link: https://github.com/openwrt/openwrt/pull/18628
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

uboot-stm32: bump to 2025.04

Bump uboot-stm32 to upstream release 2025.04.

Signed-off-by: Thomas Richard <thomas.richard@bootlin.com>
Link: https://github.com/openwrt/openwrt/pull/18628
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel: bump 6.6 to 6.6.89

Changelog: https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.6.89

All patches automatically rebased.

Build system: x86/64
Build-tested: flogic/xiaomi_redmi-router-ax6000-ubootmod, ramips/tplink_archer-a6-v3
Run-tested: 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/18607
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

kernel: bump 6.6 to 6.6.88

Changelog: https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.6.88

Manually rebased:
 - bcm27xx/patches-6.6/950-0327-media-i2c-ov7251-Make-the-enable-GPIO-optional.patch[1]
 - bcm27xx/patches-6.6/950-0521-PCI-brcmstb-Add-BCM2712-support.patch[2]
 - generic/hack-6.6/610-net-page_pool-try-to-free-deferred-skbs-while-waitin.patch[3]
 - generic/pending-6.6/734-net-ethernet-mediatek-enlarge-DMA-reserve-buffer.patch[4]

All other patches automatically rebased.

1. https://web.git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.6.88&id=f249c05416ea0bef24c9dbed0e653d2fad87b127
2. https://web.git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.6.88&id=1fea7726276e5d6526ecd4e7ccb4c91a6135deb5
3. https://web.git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.6.88&id=95f17738b86fd198924d874a5639bcdc49c7e5b8
4. https://web.git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.6.88&id=a2874f0dff63829d1f540003e2d83adb610ee64a

Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/18607
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>

uboot-sifiveu: drop PWM-related patches

These are dropped from later SDK releases, and don't compile
with the recent GCC14 updates as well.

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

Revert "toolchain: gcc: make config consistent with glibc/musl"

This reverts commit 57841c83d9c1503a19212766639d17ae5019bb8c.

This is completely breaking the inital GCC configuration and most likely
was not even compile tested, so revert until fixed.

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

Revert "kernel: netdevices: Create Vitesse DSA switch packages"

This reverts commit 5687f448a411a570b7d8558b08a132938f3de0f5.

This unfortunately is currently breaking all targets during building
as in 6.6 kernel there is
no CONFIG_NET_DSA_TAG_VSC73XX_8021Q, it was added in 6.11 AFAIK.

So they will all fail due to $(LINUX_DIR)/net/dsa/tag_vsc73xx_8021q.ko missing.

So revert until kmod is fixed.

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

Revert "gemini: Add module for VSC73xx switches"

This reverts commit 1c993da8ff2ee9ed33af23657ef6c3d45d19ed2f.

VSC kmod is broken and is causing all targets to fail compiling so revert.

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

policycoreutils: update to 3.8.1 and fix build with GCC14

Update policycoreutils to latest 3.8.1 release and add a fix for building
with GCC14 and musl.

Fix will be sent upstream.

Link: https://github.com/openwrt/openwrt/pull/18674
Signed-off-by: Robert Marko <robimarko@gmail.com>

libsemanage: update to 3.8.1 and fix build with GCC14

Update libsemanage to latest 3.8.1 release and add an upstream backport[1]
that fixes compilation with GCC14 due to basename.

[1] https://github.com/SELinuxProject/selinux/commit/a339594da6f027aed5d66ec6798a3d732df235e4

Link: https://github.com/openwrt/openwrt/pull/18674
Signed-off-by: Robert Marko <robimarko@gmail.com>

secilc: update to 3.8.1

Update secilc to latest 3.8.1 release.

Link: https://github.com/openwrt/openwrt/pull/18674
Signed-off-by: Robert Marko <robimarko@gmail.com>

checkpolicy: update to 3.8.1

Update checkpolicy to latest 3.8.1 release.

Link: https://github.com/openwrt/openwrt/pull/18674
Signed-off-by: Robert Marko <robimarko@gmail.com>

libsepol: update to 3.8.1

Update libsepol to latest 3.8.1 release.

Link: https://github.com/openwrt/openwrt/pull/18674
Signed-off-by: Robert Marko <robimarko@gmail.com>

libselinux: update to 3.8.1

Update libselinux to latest 3.8.1 release.

In order to keep building on 32-bit targets with musl, one backport[1]
and one patch pending[2] upstream are required.

[1] https://github.com/SELinuxProject/selinux/commit/5c3fcbd931b7f9752b5ce29cec3b6813991d61c0
[2] https://lore.kernel.org/selinux/CAP+JOzQBaGv=74tNgczpjZVGpzZo93kxnHXO0isL+TzmOc4byg@mail.gmail.com/T/#t

Link: https://github.com/openwrt/openwrt/pull/18674
Signed-off-by: Robert Marko <robimarko@gmail.com>

toolchain: binutils: fix compilation with GCC15

GCC15 has switched the C language default from GNU17 to GNU23[1] and this
causes builds to fail with:
In file included from mips-opc.c:29:
mips-opc.c: In function 'decode_mips_operand':
mips-formats.h:86:7: error: expected identifier or '(' before 'static_assert'
   86 |       static_assert[(1 << (SIZE)) == ARRAY_SIZE (MAP)]; \
      |       ^~~~~~~~~~~~~
mips-opc.c:214:15: note: in expansion of macro 'MAPPED_REG'
  214 |     case 'z': MAPPED_REG (0, 0, GP, reg_0_map);
      |               ^~~~~~~~~~

So, backport upstream fix for this[2] to fix compilation with GCC15.
Patch for 2.40 was manually refreshed as part of the S390 code does not
exist in 2.40 as it was added after it.

[1] https://gcc.gnu.org/gcc-15/porting_to.html#c23
[2] https://sourceware.org/git/?p=binutils-gdb.git;a=commit;h=8ebe62f3f0d27806b1bf69f301f5e188b4acd2b4

Fixes: #18678
Link: https://github.com/openwrt/openwrt/pull/18681
Signed-off-by: Robert Marko <robimarko@gmail.com>

toolchain: gcc: make config consistent with glibc/musl

I've observed configuration drift for GCC between musl and glibc
(especially it's final stage):

  # musl
  lt_cv_prog_compiler_static_works=yes
  lt_cv_prog_compiler_static_works_CXX=yes
  lt_cv_sys_max_cmd_len=1572864

  # glibc
  lt_cv_prog_compiler_static_works=no
  lt_cv_prog_compiler_static_works_CXX=no
  lt_cv_sys_max_cmd_len=512

These changes should prevent this issue in future:

  export lt_cv_prog_compiler_static_works=yes
  export lt_cv_prog_compiler_static_works_CXX=yes
  export lt_cv_sys_max_cmd_len=1572864

Also:

- provide custom autotools/libtool variables via properly named
  variable ("GCC_CONFIGURE_VARS"),
- move variables from "GCC_MAKE" to "GCC_CONFIGURE_VARS"
  (at this moment only "gcc_cv_libc_provides_ssp=yes" for musl),
- propagate it's usage for both "./configure" and "make".

Signed-off-by: Konstantin Demin <rockdrilla@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18646
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>

make_ext4fs: Update to version 2025-05-02

13767a9 make_ext4fs: add missing space after LDFLAGS

Signed-off-by: Konstantin Demin <rockdrilla@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18682
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>

mold: update to 2.38.1

https://github.com/rui314/mold/releases/tag/v2.38.0
https://github.com/rui314/mold/releases/tag/v2.38.1

Signed-off-by: Carsten Schuette <schuettecarsten@googlemail.com>
Link: https://github.com/openwrt/openwrt/pull/18679
Signed-off-by: Robert Marko <robimarko@gmail.com>

gemini: Add module for VSC73xx switches

This adds the Vitesse VSC73xx DSA switch modules to the two
Gemini devices that have them.

Link: https://patchwork.ozlabs.org/project/openwrt/patch/20250410-gemini-rtl-dsa-module-v1-2-60af8219b2cc@linaro.org/
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>

kernel: netdevices: Create Vitesse DSA switch packages

This adds kernel packages for the Vitesse VSC73XX switches.

I have split the switch into explicit SPI and platform integrated
variants as it's quite a bit of code.

Link: https://patchwork.ozlabs.org/project/openwrt/patch/20250410-gemini-rtl-dsa-module-v1-1-60af8219b2cc@linaro.org/
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>

realtek: fix RTL8214FC probing on RTL839x

Probing of the RTL8214FC on RTL839x is currently very strange.

- On RTL8393 nothing is detected and only generic PHY is reported
- On RTL8392 the port 1 is not detected while port 2-4 seem to work

Someone left a special RTL8393 detection rules that seems to indicate
that the we probe the internal SerDes instead. That is not true. Since
upgrade to kernel 6.6 the RTL8218/RTL8214FC detection is 100% accurate
and probing functions are only called when really needed.

Fix the issue by removing the condition. For now do PHY patching only
on the RTL838x where it already worked before.

Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
Link: https://github.com/openwrt/openwrt/pull/18671
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>

realtek: use remove_new

Easy compability fix for kernel 6.12.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18660
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>

airoha: pwm: use devm

Removes the need for a remove function in platform_device.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18660
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>

qualcommax: remoteproc: use remove_new

Easy compability fix for kernel 6.12.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18660
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>

ipq40xx: ipqess: use remove_new

Easy compability fix for kernel 6.12.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18660
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>

ipq40xx: qca8k: use remove_new

Easy compability fix for kernel 6.12.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18660
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>

layerscape: ppfe: use remove_new

Easy compability fix for kernel 6.12.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18660
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>

qualcommax: pwm: use remove_new

Easy compability fix for kernel 6.12.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18660
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>

ramips: pwm: use remove_new

Easy way to add compatibility for kernel 6.12.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18660
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>

ramips: mt7621-i2c: use remove_new

Easy compability fix for kernel 6.12.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18660
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>

ramips: eip93: use remove_new

Compatibility fix for kernel 6.12

Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18660
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>

kernel: add KERNEL_STACKDEPOT_MAX_FRAMES

When KERNEL_SLUB_DEBUG is enabled, build fails due to the
KERNEL_STACKDEPOT_MAX_FRAMES missing. Add it.

Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
Link: https://github.com/openwrt/openwrt/pull/18676
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>

kernel: fix KERNEL_PROBE_EVENTS_BTF_ARGS

We no longer support kernels < 6.6, so drop the dependency. This fixes a
missing symbol with 6.12 when KERNEL_DEBUG_INFO_BTF and
KERNEL_KPROBE_EVENTS are enabled.

Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
Link: https://github.com/openwrt/openwrt/pull/18676
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>

qualcommax: ipq50xx: fix DTB warning in USB node of Linksys MR5500

Fix DTB warning when compiling an image by adding the missing hash sign
of adress-cells in the usb_dwc node.

Signed-off-by: George Moussalem <george.moussalem@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/18672
Signed-off-by: Robert Marko <robimarko@gmail.com>

qualcommax: ipq50xx: fix ECC strength for Linksys MR5500

Commit fc3ff2af0c57504a2e03d635692396c5bb099f17 introduced reading
ECC strength from the device tree. Linksys MR5500 uses the mx-base dtsi
which was originally developed for Linksys MX2000 and MX5500 where the
ECC strength is set to 8. Correct this and set it to strength of 4 as
per the NAND specs for Gigadevice GD5F2GQ5REYIH.

Signed-off-by: George Moussalem <george.moussalem@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/18672
Signed-off-by: Robert Marko <robimarko@gmail.com>

config: add new TRANSPARENT_HUGEPAGE choice for 6.12 kernel

Kernel 6.12 has a new selection for TRANSPARENT_HUGEPAGE.
Add them here to avoid missing symbols.

Ref: https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=linux-6.12.y&id=683ec99f12f4c386c23bed7f6a8ef44db5a4999a

Signed-off-by: Mieczyslaw Nalewaj <namiltd@yahoo.com>
Link: https://github.com/openwrt/openwrt/pull/18663
Signed-off-by: Robert Marko <robimarko@gmail.com>

Revert "qualcommax: ipq50xx: fix GE_PHY and Uniphy resets"

This reverts commit 34e5bc4af54d366147a83904c094d54fec7fb025.

While fixing support for 2.5Gbps mode on uniphy in phy to phy link
setups, it inadvertently broke fixed link scenario. So let's revert this
commit until there's a fix that doesn't impact fixed links.

Signed-off-by: George Moussalem <george.moussalem@outlook.com>
Link: https://github.com/openwrt/openwrt/pull/18668
Signed-off-by: Robert Marko <robimarko@gmail.com>

qualcommax: ipq5018: glinet_gl-b3000: fix kernel alignment
fixes the following warning seen in bootlog

"Kernel image misaligned at boot, please fix your bootloader!"

Fixes: 3307fe8 ("qualcommax: ipq50xx: add support for GL.iNET GL-B3000")
Signed-off-by: Scott Mercer <TheRootEd24@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18658
Signed-off-by: Robert Marko <robimarko@gmail.com>

toolchain: glibc: Update glibc 2.38 to recent HEAD

f510d75f nscd: Use time_t for return type of addgetnetgrentX
5aa4bb67 resolv: Fix some unaligned accesses in resolver [BZ #30750]
249646f1 Force DT_RPATH for --enable-hardcoded-path-in-tests
697ab62d i386: Disable Intel Xeon Phi tests for GCC 15 and above (BZ 31782)
c8cb4d2b misc: Add support for Linux uio.h RWF_NOAPPEND flag
71245363 s390x: Fix segfault in wcsncmp [BZ #31934]
7bfc3595 nptl: fix potential merge of __rseq_* relro symbols
d9d019d6 elf: Make dl-rseq-symbols Linux only
bb30bd21 Linux: Make __rseq_size useful for feature detection (bug 31965)
dc512364 resolv: Allow short error responses to match any query (bug 31890)
6cad0f54 resolv: Do not wait for non-existing second DNS response after error (bug 30081)
1bed6acf resolv: Track single-request fallback via _res._flags (bug 31476)
0301637b linux: Update the mremap C implementation [BZ #31968]
f0e21145 mremap: Update manual entry
6bb75212 Add mremap tests
3ac7ba61 Update syscall lists for Linux 6.5
9184c136 resolv: Fix tst-resolv-short-response for older GCC (bug 32042)
059f82c3 Fix name space violation in fortify wrappers (bug 32052)
c005d1bd x86: Fix bug in strchrnul-evex512 [BZ #32078]
28f358bc support: Add FAIL test failure helper
99ffa84b stdio-common: Add test for vfscanf with matches longer than INT_MAX [BZ #27650]
87a1968a Make tst-ungetc use libsupport
804d3c8d ungetc: Fix uninitialized read when putting into unused streams [BZ #27821]
b9f72bd5 ungetc: Fix backup buffer leak on program exit [BZ #27821]
15ca6630 posix: Use <support/check.h> facilities in tst-truncate and tst-truncate64
f30501ca nptl: Use <support/check.h> facilities in tst-setuid3
370be858 libio: Attempt wide backup free only for non-legacy code
4dd86414 Add crt1-2.0.o for glibc 2.0 compatibility tests
fa4ad104 elf: Change ldconfig auxcache magic number (bug 32231)
9423cc53 nptl: initialize rseq area prior to registration
37ded328 nptl: initialize cpu_id_start prior to rseq registration
cf067723 x86: Avoid integer truncation with large cache sizes (bug 32470)
5c9be512 x86_64: Sort fpu/multiarch/Makefile
49016f21 x86_64: Add log2 with FMA
b2a45f1e x86_64: Add expm1 with FMA
c92946d9 x86_64: Add log1p with FMA
58822f95 x86: Check the lower byte of EAX of CPUID leaf 2 [BZ #30643]
7772f935 elf: Fix slow tls access after dlopen [BZ #19924]
5a64f933 x86: Only align destination to 1x VEC_SIZE in memset 4x loop
0d14bf07 sysdeps/x86/Makefile: Split and sort tests
dc176211 x86_64: Fix missing wcsncat function definition without multiarch (x86-64-v4)
04b8d484 x86: Improve large memset perf with non-temporal stores [RHEL-29312]
12fec8aa x86/string: Fixup alignment of main loop in str{n}cmp-evex [BZ #32212]
48642ef1 elf: Avoid some free (NULL) calls in _dl_update_slotinfo
549e7f7c elf: Support recursive use of dynamic TLS in interposed malloc
c32fd593 Fix underallocation of abort_msg_s struct (CVE-2025-0395)
650a0aaa stdlib: Test using setenv with updated environ [BZ #32588]
f984e2d7 assert: Add test for CVE-2025-0395
9ca74b8a AArch64: Improve generic strlen
95aa2143 AArch64: Optimize memset
5fe151d8 AArch64: Remove zva_128 from memset
3de51123 math: Improve layout of expf data
52c2b155 AArch64: Add SVE memset
097299ff AArch64: Use prefer_sve_ifuncs for SVE memset
5a08d049 math: Improve layout of exp/exp10 data

Signed-off-by: Konstantin Demin <rockdrilla@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/18648
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>

ramips: make Genexis EX400 factory image dependent on initramfs

The factory image generation for the Genexis EX400 image currently fails
if CONFIG_TARGET_ROOTFS_INITRAMFS is disabled.

Create the factory image only if said config option is enabled to avoid
failing builds.

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