openwrt/staging/wigyori.git
6 months agoipq40xx: drop 6.1 support
Christian Marangi [Fri, 26 Apr 2024 10:30:26 +0000 (12:30 +0200)]
ipq40xx: drop 6.1 support

Drop support for ipq40xx for kernel 6.1.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
6 months agoipq40xx: switch default to 6.6
Christian Marangi [Fri, 26 Apr 2024 10:29:30 +0000 (12:29 +0200)]
ipq40xx: switch default to 6.6

Switch default kernel version for ipq40xx to 6.6.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
6 months agoipq806x: drop 6.1 support
Christian Marangi [Fri, 26 Apr 2024 10:27:21 +0000 (12:27 +0200)]
ipq806x: drop 6.1 support

Drop support for ipq806x for kernel 6.1.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
6 months agoipq806x: switch default to 6.6
Christian Marangi [Fri, 26 Apr 2024 10:26:53 +0000 (12:26 +0200)]
ipq806x: switch default to 6.6

Switch default kernel version for ipq806x to 6.6.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
6 months agoath79: update 900-unaligned_access_hacks.patch after recent GRO change
Felix Fietkau [Fri, 26 Apr 2024 10:24:21 +0000 (12:24 +0200)]
ath79: update 900-unaligned_access_hacks.patch after recent GRO change

Fixes build error

Signed-off-by: Felix Fietkau <nbd@nbd.name>
6 months agosunxi: update the 6.6 DTS_DIR hack
Zoltan HERPAI [Fri, 26 Apr 2024 09:06:17 +0000 (11:06 +0200)]
sunxi: update the 6.6 DTS_DIR hack

Update and simplify the 6.1 vs. 6.6 DTS_DIR hack until 6.1 support is gone.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
6 months agonetifd: add flow steering mode to the packet steering script
Felix Fietkau [Wed, 24 Apr 2024 17:23:11 +0000 (19:23 +0200)]
netifd: add flow steering mode to the packet steering script

This allows directing processing of locally received packets to the CPUs
of the tasks receiving them

Signed-off-by: Felix Fietkau <nbd@nbd.name>
6 months agonetifd: add a packet steering mode matching the old script
Felix Fietkau [Wed, 24 Apr 2024 17:16:05 +0000 (19:16 +0200)]
netifd: add a packet steering mode matching the old script

This spreads packet processing across all cores

Signed-off-by: Felix Fietkau <nbd@nbd.name>
6 months agokernel: improve GRO performance
Felix Fietkau [Tue, 23 Apr 2024 10:37:05 +0000 (12:37 +0200)]
kernel: improve GRO performance

For packets not belonging to a local socket, use fraglist GRO instead of
regular GRO. This make segmenting packets very cheap and avoids the need for
selectively disabling GRO

Signed-off-by: Felix Fietkau <nbd@nbd.name>
6 months agokernel: backport flow offload pppoe fix
Felix Fietkau [Tue, 23 Apr 2024 10:08:45 +0000 (12:08 +0200)]
kernel: backport flow offload pppoe fix

Signed-off-by: Felix Fietkau <nbd@nbd.name>
6 months agotools/elfutils: refresh portability patch for macOS
Michael Pratt [Mon, 8 Apr 2024 19:15:31 +0000 (15:15 -0400)]
tools/elfutils: refresh portability patch for macOS

Quilt refresh combined two sets of changes to the same file.

The switch from using libgen.h to dirname.h because of function poisoning
from gnulib's import of basename() was added as a new patch hunk instead
of an edit to the original one.

The original patch hunk was to fix build errors on an earlier version of
elfutils before the "dirname" module was being imported to fix further
build errors with the 0.191 version.

Tested-by: Georgi Valkov <gvalkov@gmail.com> # MacOS
Signed-off-by: Michael Pratt <mcpratt@pm.me>
Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
6 months agotools/elfutils: remove no-op copy of tdestroy()
Michael Pratt [Mon, 8 Apr 2024 07:54:13 +0000 (03:54 -0400)]
tools/elfutils: remove no-op copy of tdestroy()

A false tdestroy() function was added in order to make elfutils build on
macOS again. A previous commit added declarations for a real version of
tdestroy() into gnulib, which is already imported, as well as the
preprocessor flags and the triggers for the Makefile.am conditional in
order to include the source to be built.

Tested-by: Georgi Valkov <gvalkov@gmail.com> # MacOS
Signed-off-by: Michael Pratt <mcpratt@pm.me>
Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
6 months agotools/elfutils: use locally declared static reallocarray()
Michael Pratt [Mon, 8 Apr 2024 03:28:51 +0000 (23:28 -0400)]
tools/elfutils: use locally declared static reallocarray()

On macOS, stdlib.h in the standard include paths does not provide
reallocarray() while both elfutils and gnulib do, however they are
declared differently, leading to an error:

  ./system.h:101:1: error: static declaration of 'reallocarray' follows non-static declaration
  reallocarray (void *ptr, size_t nmemb, size_t size)

A normal "configure && make" build cycle results in both declarations
being enabled as a result of both elfutils and gnulib having completely
separate configure checks where gnulib uses an internal placeholder symbol
HAVE_REALLOCARRAY, and elfutils uses a standard autoconf macro
HAVE_DECL_REALLOCARRAY.

Fix this by excluding the import of the reallocarray module which causes
gnulib checks in the configure stage to not even consider whether to
declare reallocarray later on, so the decision is only between the
standard include stdlib.h and the elfutils header.

Tested-by: Georgi Valkov <gvalkov@gmail.com> # MacOS
Signed-off-by: Michael Pratt <mcpratt@pm.me>
Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
6 months agotools/elfutils: organize gnulib import build stage
Michael Pratt [Fri, 5 Apr 2024 08:42:14 +0000 (04:42 -0400)]
tools/elfutils: organize gnulib import build stage

Organize the Makefile lines involved in gnulib importing and its
workarounds. It improves readability and keeps git history organized.

Tested-by: Georgi Valkov <gvalkov@gmail.com> # MacOS
Signed-off-by: Michael Pratt <mcpratt@pm.me>
Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
6 months agotools/elfutils: override SUBDIRS variable of Makefile
Michael Pratt [Sun, 7 Apr 2024 19:28:03 +0000 (15:28 -0400)]
tools/elfutils: override SUBDIRS variable of Makefile

Instead of editing the SUBDIRS variable with a patch, it can be overriden
at the end of the command line when invoking Make.

This tool has a series of recursive Makefiles in each subdirectory,
therefore SUBDIRS is set to a pattern of Make functions so that the result
is variable depending on the current subdirectory that Make is being
invoked in.

It's not necessary to have gnulib-cache.m4 in EXTRA_DIST since we don't
need to re-import after packaging this in the SDK, so get rid of the
entire patch hunk for ./Makefile.am

Tested-by: Georgi Valkov <gvalkov@gmail.com> # MacOS
Signed-off-by: Michael Pratt <mcpratt@pm.me>
Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
6 months agotools/elfutils: update to 1.91
Tony Ambardar [Mon, 4 Mar 2024 05:04:36 +0000 (21:04 -0800)]
tools/elfutils: update to 1.91

Release Notes:
https://sourceware.org/pipermail/elfutils-devel/2024q1/006876.html

Manually refresh:
- 100-portability.patch

Change:
- replace libgen.h with gnulib "dirname" module for compilation errors:
    In file included from ./../libdw/libdwP.h:38,
                     from eblobjnote.c:42:
    /usr/include/libgen.h:35:9: error: attempt to use poisoned "basename"
       35 | #define basename        __xpg_basename
          |         ^

Tested-by: Georgi Valkov <gvalkov@gmail.com> # MacOS
Co-Developed-by: Nick Hainke <vincent@systemli.org>
Signed-off-by: Nick Hainke <vincent@systemli.org>
Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
6 months agotools/coreutils: update to 9.5
Michael Pratt [Mon, 1 Apr 2024 22:36:49 +0000 (18:36 -0400)]
tools/coreutils: update to 9.5

Update to latest stable release.

The following commits in gnulib caused a conflict in locally bootstrapped
coreutils with stable gnulib:

  8f4b4e52c991de2233b471f8e35a068866b31f01
  2749234203959df8d72cd8638d4e00a9fff450db

A module (strftime) was marked deprecated and replaced by another module
(nstrftime) in the version of gnulib that coreutils was released with
compared to the stable branch that we use for importing. Conflicts from
the previous version of coreutils are now gone, so other imported headers
are now good.

Refresh patch:
 - 000-bootstrap.patch

Remove upstreamed patch:
 - 001-bootstrap-sync.patch

Link: https://lists.gnu.org/archive/html/coreutils/2024-03/msg00132.html
Tested-by: Georgi Valkov <gvalkov@gmail.com> # MacOS
Signed-off-by: Michael Pratt <mcpratt@pm.me>
Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
6 months agotools/coreutils: update to 9.4
Michael Pratt [Mon, 4 Sep 2023 22:59:51 +0000 (18:59 -0400)]
tools/coreutils: update to 9.4

Update to latest stable release.

Add configure option to disable support for the Year 2038 problem.
(for now, as some versions of GCC do not yet support it)

Syncing bootstrap script fails, backport an upstream patch which can be
removed at next coreutils update.

Several headers from the stable gnulib branch cause build failure because
the changes in the imported versions are incompatible with the Makefile
that gets generated for coreutils. This version of coreutils was released
after being bootstrapped and autoreconf'ed with a significantly different
version of gnulib compared to our local gnulib, so skip importing them
(and restore the backup).

While at it, organize restoring the originally shipped version of files
into a Make foreach function.

Refresh patch:
- 000-bootstrap.patch

New patch:
- 001-bootstrap-sync.patch

Link: https://lists.gnu.org/archive/html/coreutils/2023-08/msg00099.html
Tested-by: Georgi Valkov <gvalkov@gmail.com> # MacOS
Signed-off-by: Michael Pratt <mcpratt@pm.me>
Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
6 months agotools/bison: hardcode path to m4 with STAGING_DIR_HOST
Michael Pratt [Mon, 1 Apr 2024 19:51:16 +0000 (15:51 -0400)]
tools/bison: hardcode path to m4 with STAGING_DIR_HOST

Force bison to ignore the M4 environment variable and hardcode it to the
locally built m4 during build operations using the relocatable path
variable STAGING_DIR_HOST.

This allows bison to continue to function while we are forcefully avoiding
autoreconf and other autoconf and automake-like operations by giving a
fake path to m4 with the M4 environment variable.

The specific path can still be overridden independently from the
environment within the line of invocation that runs bison by setting
STAGING_DIR_HOST within the command, so document this in the help printout.

Tested-by: Georgi Valkov <gvalkov@gmail.com> # MacOS
Signed-off-by: Michael Pratt <mcpratt@pm.me>
Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
6 months agogettext-full: update to 0.22.5
Tony Ambardar [Wed, 6 Mar 2024 13:35:10 +0000 (05:35 -0800)]
gettext-full: update to 0.22.5

Release Announcement:
https://savannah.gnu.org/news/?group_id=425

Refresh:
- 200-libunistring-missing-link.patch

Tested-by: Georgi Valkov <gvalkov@gmail.com> # MacOS
Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
6 months agotools/gnulib: make tdestroy() fully portable
Michael Pratt [Mon, 8 Apr 2024 07:47:23 +0000 (03:47 -0400)]
tools/gnulib: make tdestroy() fully portable

The tdestroy() function, which is a GNU extension to the standard C
library, is defined in gnulib in tsearch.c but is missing it's
corresponding declaration in search.in.h by being completely missing...

This patch is large but upstreamable, including all of the macros and
conditionals and configure checks that upstream GNU would expect for
portable support, like using the @@ placeholder/substitution method to
determine whether or not to have declarations based on whether or not
tdestroy() is already declared within the standard headers of the default
include paths.

There were also some typedefs and aliases missing, along with the warnings
and preprocessor exceptions that need to be added for consistency with the
usage of the rest of the functions in the files.

Tested-by: Georgi Valkov <gvalkov@gmail.com> # MacOS
Signed-off-by: Michael Pratt <mcpratt@pm.me>
Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
6 months agotools/gnulib: add macros to skip reallocarray() functions
Michael Pratt [Sat, 13 Apr 2024 19:04:54 +0000 (15:04 -0400)]
tools/gnulib: add macros to skip reallocarray() functions

For modules that depend on the reallocarray module, like ialloc, xalloc,
and safe-alloc, it was not possible to skip importing the reallocarray
module as they all contained at least one function that called
reallocarray() and would cause build failure if the host system didn't
declare it.

This upstreamable patch adds macros that toggle whether to define
functions that depend on reallocarray() based on whether the reallocarray
module is being imported.

Tested-by: Georgi Valkov <gvalkov@gmail.com> # MacOS
Signed-off-by: Michael Pratt <mcpratt@pm.me>
Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
6 months agotools/gnulib: update to branch stable-202401
Tony Ambardar [Mon, 4 Mar 2024 04:15:25 +0000 (20:15 -0800)]
tools/gnulib: update to branch stable-202401

Patches refreshed automatically.

Tested-by: Georgi Valkov <gvalkov@gmail.com> # MacOS
Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
6 months agotools/missing-macros: add symlinks to makeinfo
Michael Pratt [Tue, 16 Apr 2024 01:08:06 +0000 (21:08 -0400)]
tools/missing-macros: add symlinks to makeinfo

There are other wrapper scripts released with makeinfo like texi2pdf which
are required by the build prerequisites of some tools, and have a similar
purpose and usage.

Let the makeinfo perl script handle all of these cases.

It's worth mentioning that "texi2any" is the actual program and "makeinfo"
is one of it's aliases. From upstream GNU:

  makeinfo: texi2any
rm -f $@
-$(LN_S) texi2any $@

Signed-off-by: Michael Pratt <mcpratt@pm.me>
Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
6 months agohost-build: fix stampfile name substitution per installed binaries
Michael Pratt [Thu, 4 Apr 2024 06:11:28 +0000 (02:11 -0400)]
host-build: fix stampfile name substitution per installed binaries

A funny bug was discovered where if the buildroot's path
has the name of the build target within it, it will also be substituted
along with the stampfile's name for each program,
causing an attempt to touch a file in a directory that doesn't exist.

...
...
touch: cannot touch '/Volumes/touch/openwrt/staging_dir/host/stamp/.touch_installed': No such file or directory
touch: cannot touch '/Volumes/ln/openwrt/staging_dir/host/stamp/.ln_installed': No such file or directory
touch: cannot touch '/Volumes/chown/openwrt/staging_dir/host/stamp/.chown_installed': No such file or directory
make[2]: *** [Makefile:50: /Volumes/coreutils/openwrt/staging_dir/host/stamp/.coreutils_installed] Error 1
...
...

Split up the path with $(dir) and $(notdir) before substitution to fix
the syntax.

Reported-by: Georgi Valkov <gvalkov@gmail.com>
Tested-by: Georgi Valkov <gvalkov@gmail.com> # MacOS
Signed-off-by: Michael Pratt <mcpratt@pm.me>
Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
6 months agouboot-envtools: add env settings for Edgerouter-X
Tim Lunn [Mon, 22 Apr 2024 10:53:50 +0000 (20:53 +1000)]
uboot-envtools: add env settings for Edgerouter-X

uboot-envtools is currently missing config for Edgerouter-X
and its not immediately obvious what settings to manually
apply.

Provide default configuration for envtools on Edgerouter-X.

Signed-off-by: Tim Lunn <tim@feathertop.org>
6 months agosunxi: 6.6: set testing kernel
Zoltan HERPAI [Wed, 17 Apr 2024 11:49:15 +0000 (13:49 +0200)]
sunxi: 6.6: set testing kernel

Allow selecting 6.6 as testing kernel on sunxi.

Runtime-tested:
 - Linksprite pcDuino (cortexa8)
 - Olimex A20 Micro (cortexa7)
 - Pine64 SoM (cortexa53)

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
6 months agosunxi: 6.6: refresh patches
Zoltan HERPAI [Wed, 17 Apr 2024 11:45:49 +0000 (13:45 +0200)]
sunxi: 6.6: refresh patches

Refresh kernel patches.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
6 months agosunxi: update image Makefile to reflect updated DTS structure in 6.6
Zoltan HERPAI [Wed, 17 Apr 2024 11:48:41 +0000 (13:48 +0200)]
sunxi: update image Makefile to reflect updated DTS structure in 6.6

For the ARM arch on 6.6, DTS files are moved into their vendor directories,
mimicking arm64. Reflect this in the image Makefile.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
6 months agosunxi: 6.6: refresh kernel configs
Zoltan HERPAI [Wed, 17 Apr 2024 11:48:08 +0000 (13:48 +0200)]
sunxi: 6.6: refresh kernel configs

Refresh kernel config.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
6 months agosunxi: 6.6: remove upstreamed patches
Zoltan HERPAI [Wed, 17 Apr 2024 11:47:02 +0000 (13:47 +0200)]
sunxi: 6.6: remove upstreamed patches

Remove patches that have been upstreamed.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
6 months agokernel/sunxi: Restore kernel files for v6.1
Zoltan HERPAI [Tue, 16 Apr 2024 21:17:13 +0000 (23:17 +0200)]
kernel/sunxi: Restore kernel files for v6.1

This is an automatically generated commit which aids following Kernel patch history,
as git will see the move and copy as a rename thus defeating the purpose.

See: https://lists.openwrt.org/pipermail/openwrt-devel/2023-October/041673.html
for the original discussion.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
6 months agokernel/sunxi: Create kernel files for v6.6 (from v6.1)
Zoltan HERPAI [Tue, 16 Apr 2024 21:17:13 +0000 (23:17 +0200)]
kernel/sunxi: Create kernel files for v6.6 (from v6.1)

This is an automatically generated commit.

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

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
6 months agomxs: drop 6.1 support
Zoltan HERPAI [Thu, 25 Apr 2024 12:32:46 +0000 (14:32 +0200)]
mxs: drop 6.1 support

Now that 6.6 is the default, remove the 6.1 config and the hack that
was required for the arm32 DTS dir change.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
6 months agomxs: switch default to 6.6
Zoltan HERPAI [Thu, 25 Apr 2024 12:31:59 +0000 (14:31 +0200)]
mxs: switch default to 6.6

Switch the default kernel to 6.6.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
6 months agompc85xx: fix address config for ws-ap3825i
Martin Kennedy [Wed, 24 Apr 2024 23:38:59 +0000 (19:38 -0400)]
mpc85xx: fix address config for ws-ap3825i

In commit 6a8b83159389 ("mpc85xx: p1010: change wrapper address of
simple image devices"), we adjusted the wrapper address in the recipe
code for all mpc85xx simpleimage devices, including the Extreme
Networks WS-AP3825i. However, we did not also adjust the
KERNEL_LOADADDR and KERNEL_ENTRY config values for this board. This
broke the simpleimage wrapper loader, causing GitHub issue #15237.

Adjust those config values so we go back to pointing at the right
address. We don't exactly need the memory, but it's also not exactly a
punishment in this case.

Run-tested on a ws-ap3825i.

Fixes: commit 6a8b83159389 ("mpc85xx: p1010: change wrapper address of
simple image devices")

Tested-by: Martin Kennedy <hurricos@gmail.com>
Signed-off-by: Martin Kennedy <hurricos@gmail.com>
6 months agomodules: Add kernel module for MV88E6xxx DSA switch
Linus Walleij [Sat, 1 Jul 2023 14:00:50 +0000 (16:00 +0200)]
modules: Add kernel module for MV88E6xxx DSA switch

This adds a kernel module package for the Marvell
MV88E6XXX DSA switch and a separate module package for
the DSA tagger since it can in theory be used by multiple
DSA switches. Enable both DSA and EDSA tags in the
tagger.

We can't just compile this in because just a few devices
has this DSA, and it depends on e.g. the I2C and SFP
to be loaded as modules first.

We have no examples of DSA switches being packaged as
modules before, all seem to be compiled in, but it
actually works just fine to do this.

Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
6 months agowolfssl: Update to 5.7.0
Hauke Mehrtens [Sun, 21 Apr 2024 15:46:55 +0000 (17:46 +0200)]
wolfssl: Update to 5.7.0

This fixes multiple security problems:
 * [High] CVE-2024-0901 Potential denial of service and out of bounds
   read. Affects TLS 1.3 on the server side when accepting a connection
   from a malicious TLS 1.3 client. If using TLS 1.3 on the server side
   it is recommended to update the version of wolfSSL used.

 * [Med] CVE-2024-1545 Fault Injection vulnerability in
   RsaPrivateDecryption function that potentially allows an attacker
   that has access to the same system with a victims process to perform
   a Rowhammer fault injection. Thanks to Junkai Liang, Zhi Zhang, Xin
   Zhang, Qingni Shen for the report (Peking University, The University
   of Western Australia)."

 * [Med] Fault injection attack with EdDSA signature operations. This
   affects ed25519 sign operations where the system could be susceptible
   to Rowhammer attacks. Thanks to Junkai Liang, Zhi Zhang, Xin Zhang,
   Qingni Shen for the report (Peking University, The University of
   Western Australia).

Size increased a little:
wolfssl 5.6.6:
516880 bin/packages/mips_24kc/base/libwolfssl5.6.6.e624513f_5.6.6-stable-r1_mips_24kc.ipk
wolfssl: 5.7.0:
519429 bin/packages/mips_24kc/base/libwolfssl5.7.0.e624513f_5.7.0-stable-r1_mips_24kc.ipk

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
6 months agombedtls: Update to 2.28.8
Hauke Mehrtens [Sun, 21 Apr 2024 15:40:09 +0000 (17:40 +0200)]
mbedtls: Update to 2.28.8

This contains a fix for:
CVE-2024-28960: An issue was discovered in Mbed TLS 2.18.0 through 2.28.x
before 2.28.8 and 3.x before 3.6.0, and Mbed Crypto. The PSA Crypto
API mishandles shared memory.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
6 months agomediatek: sync cpufreq support with changed compatible string
Daniel Golle [Fri, 19 Apr 2024 17:17:01 +0000 (18:17 +0100)]
mediatek: sync cpufreq support with changed compatible string

The compatible string for the MediaTek MT7988 SoC ended up being
'mediatek,mt7988a' instead of 'mediatek,mt7988' in the now upstream
dtsi. Adapt the cpufreq driver so support for frequency scaling is
again usable.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
6 months agogeneric: 6.1, 6.6: mt7530: import pending patches
Daniel Golle [Wed, 24 Apr 2024 01:26:01 +0000 (02:26 +0100)]
generic: 6.1, 6.6: mt7530: import pending patches

 net: dsa: mt7530: explain exposing MDIO bus of MT7531AE better
 net: dsa: mt7530: do not pass port variable to mt7531_rgmii_setup()
 net: dsa: mt7530: use priv->ds->num_ports instead of MT7530_NUM_PORTS
 net: dsa: mt7530: get rid of mac_port_validate member of mt753x_info
 net: dsa: mt7530: refactor MT7530_PMEEECR_P()
 net: dsa: mt7530: get rid of function sanity check
 net: dsa: mt7530: define MAC speed capabilities per switch model
 net: dsa: mt7530: return mt7530_setup_mdio & mt7531_setup_common on error
 net: dsa: mt7530: move MT753X_MTRAP operations for MT7530
 net: dsa: mt7530: refactor MT7530_HWTRAP and MT7530_MHWTRAP
 net: dsa: mt7530: refactor MT7530_MFC and MT7531_CFC, add MT7531_QRY_FFP
 net: dsa: mt7530: rename mt753x_bpdu_port_fw enum to mt753x_to_cpu_fw
 net: dsa: mt7530: rename p5_intf_sel and use only for MT7530 switch
 net: dsa: mt7530: refactor MT7530_PMCR_P()
 net: dsa: mt7530: disable EEE abilities on failure on MT7531 and MT7988

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
6 months agogeneric: 6.1, 6.6: mt7530: import accepted patches
Daniel Golle [Tue, 23 Apr 2024 12:11:52 +0000 (13:11 +0100)]
generic: 6.1, 6.6: mt7530: import accepted patches

Import patches for the MT7530 DSA driver from net-next tree:
 cae425cb43fe net: dsa: allow DSA switch drivers to provide their own phylink mac ops
 dd0c9855b413 net: dsa: introduce dsa_phylink_to_port()
 7c5e37d7ee78 net: dsa: mt7530: simplify core operations
 868ff5f4944a net: dsa: mt7530-mdio: read PHY address of switch from device tree
 2c606d138518 net: dsa: mt7530: fix port mirroring for MT7988 SoC switch
 d59cf049c837 net: dsa: mt7530: fix mirroring frames received on local port
 62d6d91db98a net: dsa: mt7530: provide own phylink MAC operations

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
6 months agogeneric: mt7530: move accepted patches to backport-6.{1,6}
Daniel Golle [Tue, 23 Apr 2024 12:08:02 +0000 (13:08 +0100)]
generic: mt7530: move accepted patches to backport-6.{1,6}

Move two accepted patches for the MT7530 DSA driver from pending-6.{1,6}
to backport-6.{1,6}.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
6 months agomediatek: correct address of MT753x switch IC
Daniel Golle [Tue, 23 Apr 2024 11:14:07 +0000 (12:14 +0100)]
mediatek: correct address of MT753x switch IC

For all boards currently working with the mt7530 DSA driver we can
be sure that the address of the switch on the MDIO bus is 31 --
simply because that address is hard-coded in the driver and the
address from the Device Tree is being ignore.

An upcoming patch will add support for MT753x ICs which are programmed
to addresses different from 0x1f using bootstrap pins. As a result the
address from the Device Tree will then be taken into account, which
will break currently working boards which got the address set to
anything else than 31.

While at it also unify the syntax in Device Tree to always us a decimal
value for the 'reg' property.

 * mt7622-buffalo-wsr-3200ax4s.dts
   Cosmetic change 'reg = <0x1f>' -> 'reg = <31>'

 * mt7622-dlink-eagle-pro-ai-ax3200-a1.dtsi
   Wrong address: 0 -> 31

 * mt7622-elecom-wrc-x3200gst3.dts
   Wrong address: 0 -> 31

 * mt7622-linksys-e8450.dtsi
   Wrong address: 0 -> 31

 * mt7622-ruijie-rg-ew3200.dtsi
   Wrong address: 0 -> 31

 * mt7622-xiaomi-redmi-router-ax6s.dts
   Wrong address: 0 -> 31

 * mt7629-iptime-a6004mx.dts
   Wrong address: 2 -> 31

 * mt7981b-zbtlink-zbt-z8102ax.dts
   Cosmetic change 'reg = <0x1f>' -> 'reg = <31>'

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
6 months agooxnas: drop target
Nick Hainke [Wed, 24 Apr 2024 05:19:30 +0000 (07:19 +0200)]
oxnas: drop target

The Upstream Linux community has discontinued support for the target.
Maintaining support for it downstream would require too much effort.
Moreover, it seems that the supported hardware is no longer deemed worthy
of it.

Signed-off-by: Nick Hainke <vincent@systemli.org>
7 months agotools/zstd: enable position independent code compilation
Bryan Roessler [Wed, 24 Apr 2024 01:37:25 +0000 (21:37 -0400)]
tools/zstd: enable position independent code compilation

Currently, trying to compile LLVM-BPF will fail with:
[2225/3517] Linking CXX shared library lib/libLLVM-15.so
FAILED: lib/libLLVM-15.so
/usr/bin/ld: staging_dir/host/lib/libzstd.a(zstd_common.o): relocation R_X86_64_32 against `.rodata.str1.1' can not be used when making a shared object; recompile with -fPIC
/usr/bin/ld: failed to set dynamic section sizes: bad value
collect2: error: ld returned 1 exit status

So, to fix it enable PIC for the host ZSTD.

Fixes: #15247
Signed-off-by: Bryan Roessler <bryanroessler@gmail.com>
7 months agoscripts: ubinize-image.sh: fix support for static volume
Daniel Golle [Tue, 23 Apr 2024 10:39:34 +0000 (11:39 +0100)]
scripts: ubinize-image.sh: fix support for static volume

Apart from being misspelled ('dymamic' vs. 'dynamic') and a mismatch of
the number of parameters, there is also simply no need to add the 'dynamic'
volume type keyword as 6th parameter as that's the default anyway.
Fix and simplify the ubinize-image.sh script to work as intended.

Fixes: 6c17d71973 ("scripts: ubinize-image.sh: support static volumes, make size optional")
Reported-by: Hannu Nyman <hannu.nyman@iki.fi>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
7 months agomediatek: mt7988a: sync dts compatible string
Daniel Golle [Tue, 23 Apr 2024 22:34:30 +0000 (23:34 +0100)]
mediatek: mt7988a: sync dts compatible string

Use 'mediatek,mt7988a' instead of 'mediatek,mt7988' as compatible
string to be in-sync with upstream and no longer break the cpufreq
driver which was also kept in sync with upstream.

Fixes: 56dd6b473b ("mediatek: sync cpufreq support with changed compatible string")
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
7 months agobcm53xx: 6.6: fix ASM cache workaround linking
Robert Marko [Mon, 1 Apr 2024 17:46:34 +0000 (19:46 +0200)]
bcm53xx: 6.6: fix ASM cache workaround linking

Kernel 6.6 checks for orphan sections and prints a warning about them,
which in turn will make CI fails as we have Werror enabled there.

Issue is that cache-v7-min.S produces .init.text section which is an
orphan section since it is not being handled by the vmlinux.lds.S linker
script.

So, lets put the generated .init.text section under .text.

Fixes: f0d8ce4f482c ("bcm53xx: add testing support for kernel 6.6")
Signed-off-by: Robert Marko <robimarko@gmail.com>
7 months agokernel: bump 6.6 to 6.6.28
John Audia [Wed, 17 Apr 2024 19:01:37 +0000 (15:01 -0400)]
kernel: bump 6.6 to 6.6.28

Changelog: https://cdn.kernel.org/pub/linux/kernel/v6.x/ChangeLog-6.6.28

Removed upstream:
pending-6.6/796-net-dsa-mt7530-trap-link-local-frames-regardless-of-.patch

All oither patches automatically rebased.

1. https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/commit/?h=v6.6.28&id=8b6c4b62582606c62074a7e7c64156f56d2785f2

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

Signed-off-by: John Audia <therealgraysky@proton.me>
7 months agox86: generic: config-6.6: add/remove default syms
John Audia [Wed, 17 Apr 2024 19:05:00 +0000 (15:05 -0400)]
x86: generic: config-6.6: add/remove default syms

Add new default and remove deprecated symbols introduced by 6.6.28[1]

1. https://git.kernel.org/pub/scm/linux/kernel/git/stable/linux.git/diff/arch/x86/Kconfig?id=v6.6.28&id2=v6.6.27

Signed-off-by: John Audia <therealgraysky@proton.me>
7 months agoqualcommax: ipq807x: use ath11k_patch_mac for RAX120v2
Paweł Owoc [Mon, 4 Mar 2024 18:29:51 +0000 (19:29 +0100)]
qualcommax: ipq807x: use ath11k_patch_mac for RAX120v2

Use ath11k_patch_mac and ath11k_set_macflag functions for RAX120v2 (pre-caldata does not contain valid MAC addresses)

Signed-off-by: Paweł Owoc <frut3k7@gmail.com>
7 months agoqualcommax: ipq807x: use ath11k_patch_mac and ath11k_remove_regdomain for MX4200
Paweł Owoc [Mon, 29 Jan 2024 23:39:55 +0000 (00:39 +0100)]
qualcommax: ipq807x: use ath11k_patch_mac and ath11k_remove_regdomain for MX4200

Use ath11k_patch_mac, ath11k_remove_regdomain and ath11k_set_macflag functions for MX4200
(only v2 variant requires MAC patching)

Signed-off-by: Paweł Owoc <frut3k7@gmail.com>
7 months agobase-files: Add new functions for ath11k caldata
Paweł Owoc [Mon, 29 Jan 2024 22:36:58 +0000 (23:36 +0100)]
base-files: Add new functions for ath11k caldata

Add new functions for ath11k caldata:
- ath11k_patch_mac (from 0 to 5)
- ath11k_remove_regdomain
- ath11k_set_macflag (some pre-caldata have the nvMacFlag flag unset which is needed to change the MAC address)

Additionaly for ath10k caldata:
- ath10k_remove_regdomain

Signed-off-by: Paweł Owoc <frut3k7@gmail.com>
7 months agomwlwifi: update to version 10.4.10-20240419
Stefan Kalscheuer [Fri, 19 Apr 2024 15:31:24 +0000 (17:31 +0200)]
mwlwifi: update to version 10.4.10-20240419

This update contains a minor fix to resolve "detected write beyond size
of field" warning during compilation:
* "replace [0] with []" (1d0d08c)

All patches still apply.

References:
* https://github.com/openwrt/openwrt/issues/15108

Signed-off-by: Stefan Kalscheuer <stefan@stklcode.de>
7 months agoipq807x: add Spectrum SAX1V1K
Connor Yoon [Tue, 19 Mar 2024 13:10:51 +0000 (14:10 +0100)]
ipq807x: add Spectrum SAX1V1K

Spectrum SAX1V1K is a AX WIFI router with 3 1G and 1 2.5G ports.
The router is provided to Spectrum customers.
It is OEM of Askey RT5010W
https://forum.openwrt.org/t/spectrum-sax1v1k-askey-rt5010w-openwrt-support/149923

It continues the original work by @MeisterLone to get this device supported.

Specifications:
```
  â€¢  CPU: Qualcomm IPQ8072A Quad core Cortex-A53 2.2GHz
  â€¢  RAM: 2048MB of DDR3
  â€¢  Storage: 1024MB eMMC
  â€¢  Ethernet: 3x 1G RJ45 ports (QCA8075) + 1 2.5G Port (QCA8081)
  â€¢  WLAN:
     â€¢ 2.4GHz: Qualcomm QCN5024 4x4 802.11b/g/n/ax 1174 Mbps PHY rate
     â€¢ 5GHz: Qualcomm QCN5054 4x4 802.11a/b/g/n/ac/ax 2402 PHY rate

  â€¢  LED: 1 gpio-controlled dual color led (blue/red)
  â€¢  Buttons: 1x reset
  â€¢  Power: 12V DC jack
```

Notes:
```
  â€¢  This commit adds only single partition support, that means
     sysupgrade is upgrading the current rootfs partition.
  â€¢  Installation can be done by serial connection only.
  â€¢  A poulated serial header is onboard
     https://forum.openwrt.org/t/spectrum-sax1v1k-askey-rt5010w-openwrt-support/149923/6
  â€¢  RX/TX is working, u-boot bootwait is active, secure boot is enabled.
```

Installation Instructions:

**Most part of the installation is performed from an initramfs image.**

Boot initramfs : Using serial connection
1. Boot up the device and wait till it displays "VERIFY_IB: Success. verify IB ok"
2. Once that message appears,
    login with username 'root'
    password serial number of your router in uppercase.
3. Use vi to paste the 'open.sh' script from @MeisterLone github on your device
    https://github.com/MeisterLone/Askey-RT5010W-D187-REV6/blob/master/Patch/open.sh
4. chmod 755 open.sh
5. ./open.sh
6. Set your ip to 192.168.0.1
7. Run a TFTP server and host the initramfs image on the TFTP server and name it "recovery.img"
8. Reboot device. On boot it will try TFTP.

Install OpenWrt from initramfs image:
1. Use SCP (or other way) to transfer OpenWrt factory image
2. Connect to device using SSH (on a LAN port)
3. Flash firmware: sysupgrade
   # sysupgrade -n -v /tmp/openwrt_sysupgrade.bin
4. Set U-boot env variable: bootcmd
   # fw_setenv bootcmd "run fix_uboot; run setup_and_boot"
5. Reboot the device
   # reboot
6. Once device is booted, residue of previous firmware will prevent openwrt to work properly.
    Factory Reset is MUST required
    # Once serial console is displaying to login, hold reset button for 10 sec
7. Now everything should be operational.

        Note: this PR adds only single partition support, that means sysupgrade is
              upgrading the current rootfs partition

Signed-off-by: Connor Yoon <j_connor@taliaent.com>
7 months agoubox: update to Git HEAD (2024-03-02)
Tony Ambardar [Sat, 20 Apr 2024 01:23:00 +0000 (18:23 -0700)]
ubox: update to Git HEAD (2024-03-02)

d413903016c4 kmodloader: support duplicate module names and aliases
1c9aaefc119a kmodloader: fix memleak adding to avl tree
4c5c45c6beac kmodloader: fix invalid write during insmod, CodeQL warnings
6a59975afc2c kmodloader: improve memory accounting and deallocation

Signed-off-by: Tony Ambardar <itugrok@yahoo.com>
7 months agogeneric: 5.15: refresh kernel patches
Christian Marangi [Mon, 22 Apr 2024 09:26:18 +0000 (11:26 +0200)]
generic: 5.15: refresh kernel patches

Refresh kernel patches with make target/linux/refresh.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
7 months agopistachio: 6.1: refresh kernel patches
Christian Marangi [Mon, 22 Apr 2024 09:24:10 +0000 (11:24 +0200)]
pistachio: 6.1: refresh kernel patches

Refresh kernel patches with make target/linux/refresh.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
7 months agogeneric: 5.15: fix broken BoHong bh25q128as patch
Christian Marangi [Mon, 22 Apr 2024 09:21:08 +0000 (11:21 +0200)]
generic: 5.15: fix broken BoHong bh25q128as patch

Fix broken BoHong bh25q128as patch that used wrong define for kernel
5.15.

Fixes: 4cb814d403bb ("generic: 5.15: Make support for BoHong bh25q128as generic")
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
7 months agobridger: update to Git HEAD (2024-04-22)
Felix Fietkau [Mon, 22 Apr 2024 08:59:19 +0000 (10:59 +0200)]
bridger: update to Git HEAD (2024-04-22)

ec8c620fd5f4 split bridge-local disable into rx and tx
40b1c5b6be4e flow: do not attempt to offload bridge-local flows

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 months agoat91: 6.1: set testing kernel
Nick Hainke [Wed, 17 Apr 2024 22:18:47 +0000 (00:18 +0200)]
at91: 6.1: set testing kernel

Allow selecting 6.1 as testing kernel on at91.

Tested-by: Claudiu Beznea <claudiu.beznea@tuxon.dev>
Signed-off-by: Nick Hainke <vincent@systemli.org>
7 months agoat91: 6.1: update kernel config
Nick Hainke [Thu, 18 Apr 2024 08:18:23 +0000 (10:18 +0200)]
at91: 6.1: update kernel config

Update the kernel config by running make kernel_oldconfig.

Signed-off-by: Nick Hainke <vincent@systemli.org>
7 months agoat91: 6.1: import patch to fix pinctrl unused function warning
Nick Hainke [Fri, 19 Apr 2024 05:05:48 +0000 (07:05 +0200)]
at91: 6.1: import patch to fix pinctrl unused function warning

Fixes error in the form of [0]:
  drivers/pinctrl/pinctrl-at91.c:1668:12: error: 'at91_gpio_resume' defined but not used [-Werror=unused-function]
   1668 | static int at91_gpio_resume(struct device *dev)
        |            ^~~~~~~~~~~~~~~~
  drivers/pinctrl/pinctrl-at91.c:1650:12: error: 'at91_gpio_suspend' defined but not used [-Werror=unused-function]
   1650 | static int at91_gpio_suspend(struct device *dev)
        |            ^~~~~~~~~~~~~~~~~

[0] - https://lore.kernel.org/all/20221215164301.934805-1-arnd@kernel.org/

Signed-off-by: Nick Hainke <vincent@systemli.org>
7 months agoat91: 6.1: remove upstreamed patches
Nick Hainke [Thu, 18 Apr 2024 08:10:39 +0000 (10:10 +0200)]
at91: 6.1: remove upstreamed patches

Remove the upstreamed patches.

Signed-off-by: Nick Hainke <vincent@systemli.org>
7 months agokernel/at91: Restore kernel files for v5.15
Nick Hainke [Wed, 17 Apr 2024 22:15:30 +0000 (00:15 +0200)]
kernel/at91: Restore kernel files for v5.15

This is an automatically generated commit which aids following Kernel patch history,
as git will see the move and copy as a rename thus defeating the purpose.

See: https://lists.openwrt.org/pipermail/openwrt-devel/2023-October/041673.html
for the original discussion.

Signed-off-by: Nick Hainke <vincent@systemli.org>
7 months agokernel/at91: Create kernel files for v6.1 (from v5.15)
Nick Hainke [Wed, 17 Apr 2024 22:15:29 +0000 (00:15 +0200)]
kernel/at91: Create kernel files for v6.1 (from v5.15)

This is an automatically generated commit.

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

Signed-off-by: Nick Hainke <vincent@systemli.org>
7 months agoppp: add sourcefilter option support
Chen Minqiang [Mon, 22 Jan 2024 14:35:41 +0000 (22:35 +0800)]
ppp: add sourcefilter option support

This make source based IPv6 routing option available for
ppp/pptp/pppoe/pppoa

Signed-off-by: Chen Minqiang <ptpt52@gmail.com>
7 months agokernel: backport ipeth CDC NCM support
Georgi Valkov [Fri, 19 Apr 2024 07:19:00 +0000 (10:19 +0300)]
kernel: backport ipeth CDC NCM support

Fixes no communication with tethered iOS devices in CDC NCM mode.
Freshly booted iOS devices start in legacy mode, but are put into
NCM mode by the official Apple driver.

[1] https://github.com/torvalds/linux/commit/a2d274c62e44b1995c170595db3865c6fe701226

Fixes: #12566
Tested-by: Georgi Valkov <gvalkov@gmail.com>
Signed-off-by: Foster Snowhill <forst@pen.gy>
Signed-off-by: Georgi Valkov <gvalkov@gmail.com>
[ better reference fixed issue ]
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
7 months agogithub: add labels to bug-report issue to better identify type
Christian Marangi [Sun, 21 Apr 2024 16:23:30 +0000 (18:23 +0200)]
github: add labels to bug-report issue to better identify type

Add labels to bug-report issue to better identify type.
Also introduce label to-triage to better handle issue by CI actions.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
7 months agogithub: permits blank issue
Christian Marangi [Sun, 21 Apr 2024 16:15:58 +0000 (18:15 +0200)]
github: permits blank issue

Blank issue might be used by devs to create tracking issue or by
advanced user to report more strange issue.

These kind of issue won't be parsed and won't be tagged. Normally bBug
report should be used for 99% of the cases.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
7 months agogeneric: 5.15: Make support for BoHong bh25q128as generic
R Maru [Fri, 8 Mar 2024 21:55:52 +0000 (15:55 -0600)]
generic: 5.15: Make support for BoHong bh25q128as generic

Move the patch for BoHong bh25q128as out of ramips to make it
generic. Not including 6.1.y since the mtd subsystem has changed,
and does not need these changes.

Patch was dropped with ramips updating to 6.1, hence we reintroudce it
here for 5.15 generic.

5.15.y functionality was verified on a Wavlink WL-WN586X3 Rev.a.

Signed-off-by: R Maru <deviantmaru@gmail.com>
[ rebase and add extra info in commit description ]
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
7 months agogeneric: 6.1: backport support for generic spi-nor from SFDP data
Christian Marangi [Sat, 2 Mar 2024 14:40:00 +0000 (15:40 +0100)]
generic: 6.1: backport support for generic spi-nor from SFDP data

Backport patches for support of generic spi-nor from SFDP data for
kernel 6.1.

Kernel 5.15 have major rework of the info flags and it's not trustable
to backport this amount of changes and expect correct function of it.

All affected patches automatically refreshed using make
target/linux/refresh.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
7 months agotoolchain/wrapper: fill and reference info.mk in staging_dir ext toolchain
Christian Marangi [Sat, 20 Apr 2024 16:29:23 +0000 (18:29 +0200)]
toolchain/wrapper: fill and reference info.mk in staging_dir ext toolchain

Currently when external toolchain is used, a info.mk is created (with
to-be-filled values) but is never actually filled with real values and
is never actually used in rules.mk.

This info.mk is used down the code with special packages like libgcc to
reference info like gcc version and package special library.

To mimic what is done with internal toolchain, add the include in
rules.mk for info.mk also for external library and in fix
toolchain/wrapper to fill the staging_dir info.mk.

The logic is to check if the external toolchain provide an info.mk and
if it doesn't we at least fill the GCC_VERSION with the value set in the
.config file.

With this special library like libgcc correctly reference and have set the
GCC_VERSION variable.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
7 months agolldpd: extended interface(s) parsing to handle patterns
Paul Donald [Fri, 12 Apr 2024 19:27:56 +0000 (21:27 +0200)]
lldpd: extended interface(s) parsing to handle patterns

For interface type parameters, the man page documents patterns:
```
*,!eth*,!!eth1

uses all interfaces, except interfaces starting with "eth",
but including "eth1".
```

* Renamed `_ifname` to `_l2dev`.
* get the l2dev via network_get_physdev (and not l3dev)
* Glob pattern `*` is also valid - use noglob for this

The net result is that now interface 'names' including globs '*' and '!'
inversions are included in the generated lldpd configs.

Temporarily `set -o noglob` and then `set +o noglob` to disable & enable
globbing respectively, because when we pass `*` as an interface choice,
other file and pathnames get sucked in from where the init script runs,
and the `*` never makes it to lldpd.

Tested extensively on: 22.03.6, 23.05.3

Signed-off-by: Paul Donald <newtwen+github@gmail.com>
[ squash with commit bumping release version ]
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
7 months agouclient: update to Git HEAD (2024-04-19)
Felix Fietkau [Fri, 19 Apr 2024 18:05:26 +0000 (20:05 +0200)]
uclient: update to Git HEAD (2024-04-19)

e8780fa7792a uclient: fix http regression

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 months agosunxi: move KERNEL_LOADADDR into subtargets
Zoltan HERPAI [Sat, 23 Mar 2024 22:58:36 +0000 (23:58 +0100)]
sunxi: move KERNEL_LOADADDR into subtargets

As preparation for the arm926ej-s support which has a different
load address, move the KERNEL_LOADADDR into the subtargets.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
7 months agosunxi: move fpu feature flag into the subtargets
Zoltan HERPAI [Sat, 23 Mar 2024 17:35:58 +0000 (18:35 +0100)]
sunxi: move fpu feature flag into the subtargets

As the upcoming arm926ej-s does not have an FPU, move this flag
into the subtargets.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
7 months agosifiveu: 6.6: set testing kernel
Zoltan HERPAI [Thu, 18 Apr 2024 21:44:23 +0000 (21:44 +0000)]
sifiveu: 6.6: set testing kernel

Allow selecting 6.6 as testing kernel on sunxi.

Runtime-tested: SiFive Unleashed

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
7 months agosifiveu: 6.6: refresh patches
Zoltan HERPAI [Thu, 18 Apr 2024 21:43:49 +0000 (21:43 +0000)]
sifiveu: 6.6: refresh patches

Refresh kernel patches.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
7 months agokernel/sifiveu: Restore kernel files for v6.1
Zoltan HERPAI [Thu, 18 Apr 2024 21:15:30 +0000 (23:15 +0200)]
kernel/sifiveu: Restore kernel files for v6.1

This is an automatically generated commit which aids following Kernel patch history,
as git will see the move and copy as a rename thus defeating the purpose.

See: https://lists.openwrt.org/pipermail/openwrt-devel/2023-October/041673.html
for the original discussion.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
7 months agokernel/sifiveu: Create kernel files for v6.6 (from v6.1)
Zoltan HERPAI [Thu, 18 Apr 2024 21:15:30 +0000 (23:15 +0200)]
kernel/sifiveu: Create kernel files for v6.6 (from v6.1)

This is an automatically generated commit.

During a `git bisect` session, `git bisect --skip` is recommended.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
7 months agomxs: add kernel 6.6 as testing
Zoltan HERPAI [Fri, 9 Feb 2024 10:12:54 +0000 (11:12 +0100)]
mxs: add kernel 6.6 as testing

Add new LTS kernel support.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
7 months agomxs: adjust DTS path for 6.6
Zoltan HERPAI [Thu, 18 Apr 2024 16:49:27 +0000 (16:49 +0000)]
mxs: adjust DTS path for 6.6

For the ARM arch on 6.6, DTS files are moved into their vendor directories,
mimicking arm64. Reflect this in the image Makefile.

Signed-off-by: Zoltan HERPAI <wigyori@uid0.hu>
7 months agouclient: update to Git HEAD (2024-04-19)
Felix Fietkau [Fri, 19 Apr 2024 15:56:17 +0000 (17:56 +0200)]
uclient: update to Git HEAD (2024-04-19)

704c78111a92 uclient-http: use ustream_ssl without ustream_fd

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 months agoustream-ssl: update to Git HEAD (2024-04-19)
Felix Fietkau [Fri, 19 Apr 2024 15:56:02 +0000 (17:56 +0200)]
ustream-ssl: update to Git HEAD (2024-04-19)

524a76e5af78 ustream-ssl: add support for using a fd instead of ustream as backing

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 months agomediatek: backport PWM drivers
Daniel Golle [Fri, 19 Apr 2024 09:55:46 +0000 (10:55 +0100)]
mediatek: backport PWM drivers

 * MT7981 and MT7988 backported to Linux 6.1
 * MT7988 backported to Linux 6.6

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
7 months agouboot-mvebu: update to version 2024.04
Robert Marko [Mon, 8 Apr 2024 08:57:58 +0000 (10:57 +0200)]
uboot-mvebu: update to version 2024.04

Lets update to 2024.04 in order to drop all of the patches as they have
been merged upstream.

Tested on Methode eDPU.

Signed-off-by: Robert Marko <robert.marko@sartura.hr>
7 months agouclient: update to Git HEAD (2024-04-18)
Felix Fietkau [Thu, 18 Apr 2024 20:27:28 +0000 (22:27 +0200)]
uclient: update to Git HEAD (2024-04-18)

c2bf660d88ec lib: add log_msg callback to get more detailed log messages
9adb4ca5219d uclient-fetch: add uclient / ustream-ssl log messages (enabled via -v)

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 months agoustream-ssl: update to Git HEAD (2024-04-18)
Felix Fietkau [Thu, 18 Apr 2024 20:27:33 +0000 (22:27 +0200)]
ustream-ssl: update to Git HEAD (2024-04-18)

f9a28a9ce864 ustream-ssl: poll connection on incomplete reads
3c49e70c4622 ustream-ssl: increase number of read buffers

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 months agoustream-ssl: update to Git HEAD (2024-04-18)
Felix Fietkau [Thu, 18 Apr 2024 11:18:21 +0000 (13:18 +0200)]
ustream-ssl: update to Git HEAD (2024-04-18)

60d8fbb5e669 mbedtls: handle session tickets for TLS 1.3
ac42af7981ae mbedtls: add ifdefs to fix legacy compatibility
af7c3532ad49 mbedtls: another cosmetic ifdef fix

Signed-off-by: Felix Fietkau <nbd@nbd.name>
7 months agobase-files: reduce IPv6 ULA prefix generation to a single call
Paul Donald [Fri, 29 Mar 2024 15:04:30 +0000 (16:04 +0100)]
base-files: reduce IPv6 ULA prefix generation to a single call

Reduce calls and pipes and read from urandom once directly with hexdump
for the necessary 5 bytes of random data to build the 48 bit ULA Prefix.

Fewer calls and forks; finish quicker; less memory used.

Tested on: 23.05.3

Signed-off-by: Paul Donald <newtwen+github@gmail.com>
7 months agomac80211: ath11k: re-enable ipq807x coldboot calibration
Robert Marko [Wed, 17 Apr 2024 10:01:02 +0000 (12:01 +0200)]
mac80211: ath11k: re-enable ipq807x coldboot calibration

It seems that move to kernel 6.6 somehow fixed the remoteproc restart so
now it properly restarts and thus coldboot calibration works as well.

ipq60xx still seems to be broken in a different way so keep it disabled.

Signed-off-by: Robert Marko <robimarko@gmail.com>
7 months agoqca-nss-dp: add patch fixing rmmod and insmod
Christian Marangi [Tue, 16 Apr 2024 14:44:41 +0000 (16:44 +0200)]
qca-nss-dp: add patch fixing rmmod and insmod

Add patch fixing rmmod and insmod. Lots of flawed logic fixed that
permits the module to correctly rmmod and insmod later.

Just to quote some change, use phy_detach instead of phy_disconnect, fix
exclusive reset_control that could only be used once, fix kernel panic
on second edma_cleanup, stop traffic before module exit...

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
7 months agoqoriq: add kernel 6.1 support
Stijn Tintel [Wed, 17 Apr 2024 21:29:00 +0000 (00:29 +0300)]
qoriq: add kernel 6.1 support

Add support for kernel 6.1 as testing kernel for qoriq. Refresh config
using `make kernel_oldconfig`.

Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
7 months agokernel/qoriq: Restore kernel files for v5.15
Stijn Tintel [Wed, 17 Apr 2024 21:23:36 +0000 (00:23 +0300)]
kernel/qoriq: Restore kernel files for v5.15

This is an automatically generated commit which aids following Kernel patch history,
as git will see the move and copy as a rename thus defeating the purpose.

See: https://lists.openwrt.org/pipermail/openwrt-devel/2023-October/041673.html
for the original discussion.

Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
7 months agokernel/qoriq: Create kernel files for v6.1 (from v5.15)
Stijn Tintel [Wed, 17 Apr 2024 21:23:36 +0000 (00:23 +0300)]
kernel/qoriq: Create kernel files for v6.1 (from v5.15)

This is an automatically generated commit.

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

Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
7 months agoqoriq: refresh kernel 5.15 config
Stijn Tintel [Wed, 17 Apr 2024 21:23:10 +0000 (00:23 +0300)]
qoriq: refresh kernel 5.15 config

Done with `make kernel_oldconfig`.

Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
7 months agoipq807x: add support for Zbtlink ZBT-Z800AX
Chukun Pan [Tue, 26 Mar 2024 15:16:05 +0000 (23:16 +0800)]
ipq807x: add support for Zbtlink ZBT-Z800AX

Hardware specifications:
  SoC: Qualcomm IPQ8072A
  RAM: 1GB of DDR4 600MHz
  Flash1: Winbond W25Q64DW 8MB
  Flash2: MX30UF2G28AD 256MB
  WiFi1: QCN5024 2.4GHz AX 4x4
  WiFi2: QCN5054 5GHz AX 4x4
  Ethernet: 5x 1G RJ45 port
  USB: 2x USB 3.0 (1x M.2)
  Button: Reset, WPS

Flash instructions:
  Upload factory.bin in stock firmware's
  upgrade page, do not preserve settings.

Signed-off-by: Chukun Pan <amadeus@jmu.edu.cn>
7 months agoipq-wifi: update to version 2024-04-16
Chukun Pan [Tue, 16 Apr 2024 15:10:02 +0000 (23:10 +0800)]
ipq-wifi: update to version 2024-04-16

Contains following updates:

 * ipq8074: add Spectrum SAX1V1K BDF
 * ipq8074: add Zbtlink ZBT-Z800AX BDF
 * ipq8074: update regdb in Spectrum SAX1V1K BDF
 * ipq6018: add Linksys MR7350 BDF

Signed-off-by: Chukun Pan <amadeus@jmu.edu.cn>