openwrt/staging/dangole.git
2 years agoat91: kernel v5.15: copy config and patches from 5.10
Claudiu Beznea [Thu, 15 Sep 2022 14:09:28 +0000 (17:09 +0300)]
at91: kernel v5.15: copy config and patches from 5.10

Copy kernel config and patches from 5.10. Along with it
individual targets' config-default from 5.10 has been moved to
config-5.10.

Signed-off-by: Claudiu Beznea <claudiu.beznea@microchip.com>
2 years agooxnas: shuttle-kd20: fix thermal zone
Daniel Golle [Fri, 21 Oct 2022 19:47:26 +0000 (20:47 +0100)]
oxnas: shuttle-kd20: fix thermal zone

Fix typo ('_' vs '-') and add #cooling-cells to gpio-fan to get
thermal zone into functional state.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2 years agooxnas: switch to Linux 5.15 and drop Linux 5.10
Daniel Golle [Fri, 21 Oct 2022 17:59:37 +0000 (18:59 +0100)]
oxnas: switch to Linux 5.15 and drop Linux 5.10

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2 years agoCI: kernel: fix deprecation of set-output
Christian Marangi [Fri, 21 Oct 2022 14:09:19 +0000 (16:09 +0200)]
CI: kernel: fix deprecation of set-output

From [0], github deprecated set-output with a better approach of
appending variables to $GITHUB_OUTPUT

[0] https://github.blog/changelog/2022-10-11-github-actions-deprecating-save-state-and-set-output-commands/

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
2 years agokernel: disable CONFIG_CPU_LITTLE_ENDIAN in generic config
Aleksander Jan Bajkowski [Mon, 22 Feb 2021 18:48:42 +0000 (19:48 +0100)]
kernel: disable CONFIG_CPU_LITTLE_ENDIAN in generic config

Endianness depends on CPU architecture. CONFIG_CPU_(BIG/LITTLE)_ENDIAN should
be enabled on target or subtarget based on SoC architecture.

Fixes warning:
$ make kernel_oldconfig CONFIG_TARGET=subtarget
...
.config:1008:warning: override: CPU_LITTLE_ENDIAN changes choice state
....

Summary:
- ARC - only the CONFIG_CPU_BIG_ENDIAN symbol is defined for this architeture.
  If it is disabled then the processor operates in LITTLE_ENDIAN mode (default),
- ARM32 - CONFIG_CPU_LITTLE_ENDIAN symbol available since kernel 5.19. This
  option should be enabled after OpenWRT moves to kernel 6.x. After refreshing
  the kernel, the symbol disappears,
- ARM64 - enabled CONFIG_CPU_LITTLE_ENDIAN,
- MIPS - enabled relevant symbols,
- POWERPC -  enabled CONFIG_CPU_BIG_ENDIAN,
- UML - Symbols are not defined for this architecture,
- X86 - always little endian. Symbols are not defined for this architecture.

Signed-off-by: Aleksander Jan Bajkowski <olek2@wp.pl>
2 years agolantiq: add rgmii delays on BT Home Hub 5A
Aleksander Jan Bajkowski [Sun, 16 Oct 2022 18:05:26 +0000 (20:05 +0200)]
lantiq: add rgmii delays on BT Home Hub 5A

This comit fixes warnings that occur on kernel 5.15:
...
[    2.269736] Intel XWAY PHY11G (PEF 7071/PEF 7072) v1.5 / v1.6 1e108000.switch-mii:00:
               PHY has delays (e.g. via pin strapping), but phy-mode = 'rgmii'
[    2.269736] Should be 'rgmii-id' to use internal delays txskew:1500 ps rxskew:1500 ps
...

Ref: https://github.com/torvalds/linux/commit/be393dd685d215d44a43f5c5dcb8f7e57668d00e
Signed-off-by: Aleksander Jan Bajkowski <olek2@wp.pl>
2 years agobcm53xx: only fixup seama on D-Link DIR-885L
Linus Walleij [Thu, 20 Oct 2022 20:43:45 +0000 (22:43 +0200)]
bcm53xx: only fixup seama on D-Link DIR-885L

Just one device builds seama images so let's just fix up
seama on that one device. I guess the tool errors out but
this feels cleaner.

Cc: Hauke Mehrtens <hauke@hauke-m.de>
Cc: Rafał Miłecki <zajec5@gmail.com>
Signed-off-by: Linus Walleij <linus.walleij@linaro.org>
[rmilecki: drop "fixtrx" from D-Link case]
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
2 years agobcm53xx: enable Broadcom 4366b1 firmware for Asus RT-AC88U
Arınç ÜNAL [Tue, 18 Oct 2022 08:54:20 +0000 (11:54 +0300)]
bcm53xx: enable Broadcom 4366b1 firmware for Asus RT-AC88U

On some of the hardware revisions of Asus RT-AC88U, brcmfmac detects the
4366b1 wireless chip and tries to load the firmware file which doesn't
exist because it's not included in the image.

Therefore, include firmware for 4366b1 along with 4366c0. This way, all
hardware revisions of the router will be supported by having brcmfmac use
the firmware file for the wireless chip it detects.

Signed-off-by: Arınç ÜNAL <arinc.unal@arinc9.com>
2 years agomediatek: filogic: consolidate adc '32k' clock
Daniel Golle [Thu, 20 Oct 2022 19:20:41 +0000 (20:20 +0100)]
mediatek: filogic: consolidate adc '32k' clock

Add dependency to '32k' ADC clock so it is always enabled for thermal
and raw access to ADC values. This allows to remove the patch for the
ADC driver and reduce the patch adding thermal support for MT7986 to
only add the new efuse layout and temperature decoding for V3.

Suggested-by: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2 years agoscripts/download.pl: pass aria2 config in ENV only
Zhang Hua [Sun, 2 Oct 2022 11:04:12 +0000 (19:04 +0800)]
scripts/download.pl: pass aria2 config in ENV only

The aria2c command tries to load config from
${XDG_CONFIG_HOME:-${HOME}/.config}/aria2/aria2.conf by default,
which may result unexpected behavior.

As a replacement, people can use environment variable ARIA2C_OPTIONS
to custom arguments passed to aria2c like curl and wget below.
Including --conf-path=/path/to/config.conf in ARIA2C_OPTIONS can
also set a custom config file path easily if needed.

Signed-off-by: Zhang Hua <zhanghuadedn@gmail.com>
2 years agouboot-fritz4040: build FritzBox 7520 variant
Andre Heider [Fri, 15 Jul 2022 05:59:15 +0000 (07:59 +0200)]
uboot-fritz4040: build FritzBox 7520 variant

Support was added as variant of 7530 (DEVICE_ALT0_*) in:
cb6f4be1 "ipq40xx: add support for FRITZ!Box 7520"

u-boot has a distinct config for it [0], built it.

[0] https://github.com/chunkeey/FritzBox-4040-UBOOT/pull/6

Signed-off-by: Andre Heider <a.heider@gmail.com>
Tested-by: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
2 years agokernel: disable stack validation for external module builds as needed
Mark Mentovai [Sun, 11 Sep 2022 17:48:17 +0000 (13:48 -0400)]
kernel: disable stack validation for external module builds as needed

c3e31b6a9b04 and 5f8e5872406d disable stack validation when the build
host is not running Linux, as the objtool kernel build tool required for
stack validation is not portable to other build host environments. This
was achieved by setting CONFIG_STACK_VALIDATION= in KERNEL_MAKEOPTS, and
by setting SKIP_STACK_VALIDATION=1 in the environment. KERNEL_MAKEOPTS
only has effect for the kernel build, not for external module builds,
but through kernel 5.14, SKIP_STACK_VALIDATION worked to disable this
feature too, so stack validation was disabled for external module builds
as well. Since kernel 0d989ac2c90b, the kernel build no longer considers
SKIP_STACK_VALIDATION, so the feature will be disabled for the kernel
build, but not for external module builds.

When building OpenWrt on a non-Linux build host targeting x86 (the only
target architecture for which OpenWrt enables the kernel
CONFIG_STACK_VALIDATION) and using kernel 5.15 (such as via
CONFIG_TESTING_KERNEL), this caused a build failure during any external
module build, such as kmod-button-hotplug. This manifested as build
errors such as:

  make[4]: *** No rule to make target
  '.../build_dir/target-x86_64_musl/linux-x86_64/button-hotplug/button-hotplug.o',
  needed by
  '.../build_dir/target-x86_64_musl/linux-x86_64/button-hotplug/button-hotplug.mod'.
  Stop.

Although button-hotplug.c was present, the implicit rule to make
$(obj)/%.o from $(src)/%.c in the kernel's scripts/Makefile.build could
not be satisfied in this case, as it also depends on $(objtool_dep),
non-empty as a result of the failure to propagate disabling of stack
validation to external module builds, in a configuration where it is not
possible to build objtool.

KERNEL_MAKEOPTS is used for just the kernel build itself, while
KERNEL_MAKE_FLAGS is used for both the kernel build and for external
module builds. This restores the ability to build OpenWrt in such
configurations by moving the CONFIG_STACK_VALIDATION= make argument from
KERNEL_MAKEOPTS to KERNEL_MAKE_FLAGS where it is able to affect external
module builds properly.

Note that the kernel's objtool and related configuration have seen a
major overhaul since kernel 5.15, and may need more attention again
after 22922deae13f, in kernel 5.19.

Signed-off-by: Mark Mentovai <mark@mentovai.com>
2 years agoci: kernel: trigger build check on changes in kernel.mk as well
Petr Štetiar [Wed, 19 Oct 2022 21:02:43 +0000 (23:02 +0200)]
ci: kernel: trigger build check on changes in kernel.mk as well

So we can QA more parts of kernel build process.

Signed-off-by: Petr Štetiar <ynezz@true.cz>
2 years agopistachio: refresh kernel patches
Petr Štetiar [Thu, 20 Oct 2022 05:47:54 +0000 (07:47 +0200)]
pistachio: refresh kernel patches

Refresh dirty patches after commit 711f1a8bcbdd ("kernel: mtd: backport
SafeLoader parser").

Fixes: 711f1a8bcbdd ("kernel: mtd: backport SafeLoader parser")
Signed-off-by: Petr Štetiar <ynezz@true.cz>
2 years agoath79: refresh kernel patches
Petr Štetiar [Thu, 20 Oct 2022 05:47:54 +0000 (07:47 +0200)]
ath79: refresh kernel patches

Refresh dirty patches after commit 711f1a8bcbdd ("kernel: mtd: backport
SafeLoader parser").

Fixes: 711f1a8bcbdd ("kernel: mtd: backport SafeLoader parser")
Signed-off-by: Petr Štetiar <ynezz@true.cz>
2 years agomt7622: switch default CPU governor to ondemand
John Audia [Fri, 21 Jan 2022 19:12:53 +0000 (14:12 -0500)]
mt7622: switch default CPU governor to ondemand

The most common CPU governor in the OpenWRT project is currently ondemand (see
below).  Switch mt7622 over to it as well.

Audit the code by running the following and then analyzing the results:
find -name 'config-5.*' -print0 | xargs -0 grep 'CONFIG_CPU_FREQ_DEFAULT_GOV.*=y'

ondemand: 16
performance: 5
schedutil: 5
userspace: 2

Build system: x86_64
Build-tested: ramips/mt7621

Signed-off-by: John Audia <graysky@archlinux.us>
2 years agoscripts/download.pl: make the download tool configurable
Christian Marangi [Fri, 30 Sep 2022 23:36:48 +0000 (01:36 +0200)]
scripts/download.pl: make the download tool configurable

Introduce a new option in the "Advanced configuration options" to
configure a custom download tool.

By declaring a string in "Use custom download tool" an user can force
what command to use to download package. With the string empty the
default tool used is curl, with wget as a fallback if not available.

download.pl supports 3 tools officially aria2c, curl and wget.
If one of the tool is used in this config, download.pl will use the
default args to make use of them.

If the provided string is different than aria2c, curl or wget, the command
is used as is and the download url will be appended at the end of such command.

While at it also tweak the tool selection logic and chose the tool only
once when the script is called and move aria2c specific variables in the
relevant section.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
2 years agotools: remove implicit dependencies
Rosen Penev [Wed, 28 Sep 2022 01:57:07 +0000 (18:57 -0700)]
tools: remove implicit dependencies

Small cleanup.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
2 years agotools: add Host/Uninstall where possible
Rosen Penev [Wed, 28 Sep 2022 08:23:56 +0000 (01:23 -0700)]
tools: add Host/Uninstall where possible

This cleans staging_dir when calling tool/x/clean.

Signed-off-by: Rosen Penev <rosenp@gmail.com>
2 years agomediatek: clean up platform kernel modules
Daniel Golle [Wed, 19 Oct 2022 22:03:07 +0000 (23:03 +0100)]
mediatek: clean up platform kernel modules

Remove kmod-sdhci-mtk as the mtk-sd driver is built-in anyway for the
relevant subtargets in order to support mounting rootfs from eMMC or
SD card.

Add kmod-iio-mt6577-auxadc to support reading the raw values from the
auxadc unit used as in-SoC thermal sensor. This driver was previously
built-in, but as thermal itself works well without it there is no use
for it in every day use of a device. Build the module to still allow
access to the raw values for those who need it.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2 years agokernel: modules: iio: create and use AddDepends
Daniel Golle [Wed, 19 Oct 2022 22:00:37 +0000 (23:00 +0100)]
kernel: modules: iio: create and use AddDepends

Just like for other subsystems, create a reusable AddDepends call.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2 years agoramips: kndrt31r19: set up wan interface by default
Andrey Butirsky [Fri, 9 Sep 2022 22:57:20 +0000 (01:57 +0300)]
ramips: kndrt31r19: set up wan interface by default

The only WAN port of the device is it's modem, so set it up as such

Signed-off-by: Andrey Butirsky <butirsky@gmail.com>
2 years agoipq40xx: Convert mikrotik,wap-ac to DSA
Mark Mentovai [Thu, 11 Nov 2021 17:24:24 +0000 (12:24 -0500)]
ipq40xx: Convert mikrotik,wap-ac to DSA

As done previously, this preserves the MAC addresses of they physical
Ethernet ports. The interfaces are renamed as eth0 is in use for the
native GMAC; the new interface naming matches the physical port labels.
 - sw-eth1 corresponds to the physical port labeled ETH1 and has the
   base MAC address. This port can be used to power the device.
 - sw-eth2 corresponds to the physical port labeled ETH2 and has a MAC
   address one greater than the base.

As this device has 2 physical ports, they are each connected to their
respective PHYs, allowing the link status to be visible to software.
Since they are not marked on the case with any role (such as LAN or
WAN), both are bridged to the lan network by default, although this can
easily be changed if needed.

Signed-off-by: Mark Mentovai <mark@mentovai.com>
2 years agoath79: improve MAC address configuration of ELECOM devices
INAGAKI Hiroshi [Wed, 19 Oct 2022 15:00:27 +0000 (00:00 +0900)]
ath79: improve MAC address configuration of ELECOM devices

Get MAC address of WAN from HW.WAN.MAC.Address in hwconfig partition
instead of calculated one from wlan's address.
And added label_mac.

Signed-off-by: INAGAKI Hiroshi <musashino.open@gmail.com>
2 years agoath79: use NVMEM for wlan caldata on ELECOM devices
INAGAKI Hiroshi [Wed, 19 Oct 2022 14:57:34 +0000 (23:57 +0900)]
ath79: use NVMEM for wlan caldata on ELECOM devices

Use NVMEM "calibration" implementation for ath9k/ath10k(-ct) on ELECOM
WRC-300GHBK2-I and WRC-1750GHBK2-I/C instead of mtd-cal-data property
or user-space script.

Signed-off-by: INAGAKI Hiroshi <musashino.open@gmail.com>
2 years agoath79: use ARTIFACTS for initramfs-factory of ELECOM devices
INAGAKI Hiroshi [Mon, 1 Aug 2022 04:40:10 +0000 (13:40 +0900)]
ath79: use ARTIFACTS for initramfs-factory of ELECOM devices

Use ARTIFACTS to generate factory image of the following ELECOM devices
instead of redundant recipe which generate on KERNEL_INITRAMFS.

- ELECOM WRC-300GHBK2-I
- ELECOM WRC-1750GHBK2-I/C

Signed-off-by: INAGAKI Hiroshi <musashino.open@gmail.com>
2 years agobuild: move Build/edimax-header to image-commands.mk
INAGAKI Hiroshi [Mon, 1 Aug 2022 04:30:39 +0000 (13:30 +0900)]
build: move Build/edimax-header to image-commands.mk

To use from the following devices in ath79 target, move edimax-header to
image-commands.mk.

- ELECOM WRC-300GHBK2-I
- ELECOM WRC-1750GHBK2-I/C

Signed-off-by: INAGAKI Hiroshi <musashino.open@gmail.com>
2 years agowireless-tools: add package CPE ID
Petr Štetiar [Wed, 19 Oct 2022 19:38:08 +0000 (21:38 +0200)]
wireless-tools: add package CPE ID

Common Platform Enumeration (CPE) is a structured naming scheme for
information technology systems, software, and packages.

Suggested-by: Steffen Pfendtner <s.pfendtner@ads-tec.de>
Signed-off-by: Petr Štetiar <ynezz@true.cz>
2 years agoncurses: add package CPE ID
Petr Štetiar [Wed, 19 Oct 2022 19:38:08 +0000 (21:38 +0200)]
ncurses: add package CPE ID

Common Platform Enumeration (CPE) is a structured naming scheme for
information technology systems, software, and packages.

Suggested-by: Steffen Pfendtner <s.pfendtner@ads-tec.de>
Signed-off-by: Petr Štetiar <ynezz@true.cz>
2 years agoarm-trusted-firmware-sunxi: add package CPE ID
Petr Štetiar [Wed, 19 Oct 2022 19:38:08 +0000 (21:38 +0200)]
arm-trusted-firmware-sunxi: add package CPE ID

Common Platform Enumeration (CPE) is a structured naming scheme for
information technology systems, software, and packages.

Suggested-by: Steffen Pfendtner <s.pfendtner@ads-tec.de>
Signed-off-by: Petr Štetiar <ynezz@true.cz>
2 years agolibnftnl: add package CPE ID
Petr Štetiar [Wed, 19 Oct 2022 19:38:08 +0000 (21:38 +0200)]
libnftnl: add package CPE ID

Common Platform Enumeration (CPE) is a structured naming scheme for
information technology systems, software, and packages.

Suggested-by: Steffen Pfendtner <s.pfendtner@ads-tec.de>
Signed-off-by: Petr Štetiar <ynezz@true.cz>
2 years agox86: switch to 5.15 as default kernel
Nick Hainke [Sat, 8 Oct 2022 14:02:45 +0000 (16:02 +0200)]
x86: switch to 5.15 as default kernel

Set 5.15 as default kernel. Testing support was already added on
29.03.2021.

Testing support was added in
9b3bc5d11978 ("x86: enable 5.15 as testing kernel").

[0] - https://github.com/torvalds/linux/commit/edd602146507532c1714d8428f654b87205f492e

Tested-by: Mark Mentovai <mark@mentovai.com>
Tested-by: Stefan Lippers-Hollmann <s.l-h@gmx.de>
Tested-by: Thomas Huehn <thomas.huehn@hs-nordhausen.de>
Signed-off-by: Nick Hainke <vincent@systemli.org>
2 years agokernel: test for kernel version greater 5.13 to fix mmc-spi
Nick Hainke [Mon, 10 Oct 2022 06:04:41 +0000 (08:04 +0200)]
kernel: test for kernel version greater 5.13 to fix mmc-spi

The of_mmc_spi.o resource is provider agnostic in kernels greater 5.13
and does not depend anymore on CONFIG_OF [0].

[0] - https://github.com/torvalds/linux/commit/edd602146507532c1714d8428f654b87205f492e

Suggested-by: John Thomson <git@johnthomson.fastmail.com.au>
Signed-off-by: Nick Hainke <vincent@systemli.org>
2 years agouboot-envtools: mediatek/mt7622: don't rely on mapped rootfs
Daniel Golle [Wed, 19 Oct 2022 19:01:35 +0000 (20:01 +0100)]
uboot-envtools: mediatek/mt7622: don't rely on mapped rootfs

Similar to the implementation for the BPi-R3 use the same logic also
for determining the device to look for the U-Boot environment of the
BPi-R64.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2 years agomediatek: mt7622: don't rely on existing image for sysupgrade
Daniel Golle [Wed, 19 Oct 2022 18:58:12 +0000 (19:58 +0100)]
mediatek: mt7622: don't rely on existing image for sysupgrade

Don't reply on mapped rootfs partition but rather just take what ever
has been set to the kernel cmdline root= parameter as a hint to decide
which media to install sysupgrade to on the BananaPi BPi-R64.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2 years agomediatek: mt7622: kernel config fixes
Daniel Golle [Wed, 19 Oct 2022 17:08:09 +0000 (18:08 +0100)]
mediatek: mt7622: kernel config fixes

 * remove orphaned Kconfig symbol now that CONFIG_IIO is no longer
   selected after commit ef8b935c95 ("mediatek: clean up mt7622 kernel config")

 * select UBI fast-map feature to decrease boot time and keep the
   number of spare blocks required in sync with U-Boot's expectations
   (we got fast-map enabled in U-Boot)

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2 years agomediatek: update pending and rename merged patch
Daniel Golle [Wed, 19 Oct 2022 14:31:35 +0000 (15:31 +0100)]
mediatek: update pending and rename merged patch

Add patch headers and description for pending patch.
Add version tag to patch already merged upstream.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2 years agokernel: mtd: backport SafeLoader parser
Rafał Miłecki [Tue, 18 Oct 2022 20:37:46 +0000 (22:37 +0200)]
kernel: mtd: backport SafeLoader parser

Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
2 years agokernel: mtd: backport extended dynamic partitions support
Rafał Miłecki [Tue, 18 Oct 2022 20:12:27 +0000 (22:12 +0200)]
kernel: mtd: backport extended dynamic partitions support

This gets rid of "nvmem-cells" limitation. Dynamic partitions can be
defined for any (sub)partitions layout.

Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
2 years agokernel: mtd: fix unbalanced of_node_put() in dynamic partitions code
Rafał Miłecki [Tue, 18 Oct 2022 19:25:15 +0000 (21:25 +0200)]
kernel: mtd: fix unbalanced of_node_put() in dynamic partitions code

Fixes: 1a9ee367343ed ("kernel: backport mtd dynamic partition patch")
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
2 years agokernel: disable kmod-fs-ntfs3 for Linux 5.10
Rafał Miłecki [Wed, 12 Oct 2022 09:42:36 +0000 (11:42 +0200)]
kernel: disable kmod-fs-ntfs3 for Linux 5.10

This new NTFS driver was added in kernel 5.15. Avoid building empty
package for kernel 5.10.

Fixes: bd0db6017b192 ("kernel: 5.15: add new module")
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
2 years agogeneric: 5.15: add missing tag for some backports patches
Christian Marangi [Tue, 18 Oct 2022 20:38:05 +0000 (22:38 +0200)]
generic: 5.15: add missing tag for some backports patches

Add for mtd-parsers-trx-allow-to-use-on-MediaTek-MIPS-SoCs.patch and
hwmon-lm70-Add-ti-tmp125-support.patch the 5.18 tag as it was merged
in kernel 5.18.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
2 years agoipq806x: 5.15: add 6.1 tag to upstream patch
Christian Marangi [Tue, 18 Oct 2022 20:19:34 +0000 (22:19 +0200)]
ipq806x: 5.15: add 6.1 tag to upstream patch

Add 6.1 tag to upstream patch now that 6.1 got tagged. This permits to
track patch in a better way and directly drop them on kernel bump.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
2 years agouboot-envtools: mediatek_filogic: fix BPi-R3 when no OS is installed
Daniel Golle [Tue, 18 Oct 2022 19:01:23 +0000 (20:01 +0100)]
uboot-envtools: mediatek_filogic: fix BPi-R3 when no OS is installed

Fix accessing the environment in case no OS is installed on the flash
media selected for boot as this is possible when booting initramfs.
In case of relying on the device specified to be mounted as rootfs to
be present, rather just use the kernel cmdline 'root' variable as a
hint to decide where to read/write the U-Boot environment.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2 years agomediatek: filogic: don't rely on image preset in flash or sysupgrade
Daniel Golle [Tue, 18 Oct 2022 18:47:25 +0000 (19:47 +0100)]
mediatek: filogic: don't rely on image preset in flash or sysupgrade

Instead of trying to figure out the actual root device, just use the
kernel 'root' cmdline parameter as a hint to decide which device to
flash to.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2 years agofstools: update to git HEAD
Daniel Golle [Tue, 18 Oct 2022 18:42:35 +0000 (19:42 +0100)]
fstools: update to git HEAD

 ee54c6b libfstools: skip JFFS2 padding on block devices

Fixes config restore on the BPi R3 when using MMC storage.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2 years agouboot-mediatek: update to U-Boot 2022.10
Daniel Golle [Tue, 18 Oct 2022 13:28:53 +0000 (14:28 +0100)]
uboot-mediatek: update to U-Boot 2022.10

Remove patches adding support for MT7621 which have been merged upsteam.
Patches for MT7981 and MT7986 have been merged too, but not in time to
be included in the 2022.10 release, so we have to keep carrying them
until the 2023.01 release.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2 years agomediatek: filogic: use UBI fast map to speed up boot
Daniel Golle [Tue, 18 Oct 2022 18:48:50 +0000 (19:48 +0100)]
mediatek: filogic: use UBI fast map to speed up boot

Use UBI fast map feature to avoid scanning the whole flash on each
boot which takes several seconds.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2 years agomediatek: don't break auxadc without 32k clk
Daniel Golle [Sun, 16 Oct 2022 18:35:50 +0000 (19:35 +0100)]
mediatek: don't break auxadc without 32k clk

Make the newly added 32k clock optional for the auxadc driver also used
on pre-filogic platforms.

Signed-off-by: Daniel Golle <daniel@makrotopia.org>
2 years agomt7621: hiwifi_hc5962: fix reboot loop by using LZMA loader
Petr Štetiar [Tue, 18 Oct 2022 06:55:46 +0000 (08:55 +0200)]
mt7621: hiwifi_hc5962: fix reboot loop by using LZMA loader

This fixes a well known "LZMA ERROR 1" error, reported previously on
numerous of similar devices.

References: https://github.com/openwrt/openwrt/issues/10645#issuecomment-1282607274
Signed-off-by: Petr Štetiar <ynezz@true.cz>
2 years agoethtool: update to 6.0
Nick Hainke [Thu, 13 Oct 2022 21:47:26 +0000 (23:47 +0200)]
ethtool: update to 6.0

Release Notes:
https://lwn.net/Articles/910841/

Signed-off-by: Nick Hainke <vincent@systemli.org>
2 years agobase-files: add protocol qmi/mbim support for ucidef_set_interface()
Andrey Butirsky [Wed, 7 Sep 2022 13:54:00 +0000 (16:54 +0300)]
base-files: add protocol qmi/mbim support for ucidef_set_interface()

Make it possible to setup default WAN interface for devices with built-in LTE
modems, using QMI or MBIM.

Signed-off-by: Andrey Butirsky <butirsky@gmail.com>
Reviewed-by: Lech Perczak <lech.perczak@gmail.com>
2 years agortl8812au-ct: Fix compile
Hauke Mehrtens [Sat, 15 Oct 2022 10:25:19 +0000 (12:25 +0200)]
rtl8812au-ct: Fix compile

Replace the extern inline with a static inline. With extern inline the
compiler will generate the function in all compile units including this
file which breaks linking later.

Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
2 years agokernel: add missing config symbols for 5.15
Josef Schlehofer [Thu, 13 Oct 2022 13:09:45 +0000 (15:09 +0200)]
kernel: add missing config symbols for 5.15

When I enabled CONFIG_HOTPLUG_PCI for mvebu platform, it was asking for
more symbols.

Fixes:
Support for PCI Hotplug (HOTPLUG_PCI) [Y/?] y
  CompactPCI Hotplug driver (HOTPLUG_PCI_CPCI) [N/y/?] n
  SHPC PCI Hotplug driver (HOTPLUG_PCI_SHPC) [N/y/?] (NEW)

Signed-off-by: Josef Schlehofer <pepe.schlehofer@gmail.com>
2 years agokernel: bump 5.15 to 5.15.74
John Audia [Sat, 15 Oct 2022 09:51:54 +0000 (05:51 -0400)]
kernel: bump 5.15 to 5.15.74

All patches automatically rebased.

Build system: x86_64
Build-tested: bcm2711/RPi4B, mt7622/RT3200
Run-tested: bcm2711/RPi4B, mt7622/RT3200

Signed-off-by: John Audia <therealgraysky@proton.me>
2 years agokernel: bump 5.15 to 5.15.73
John Audia [Wed, 12 Oct 2022 09:40:31 +0000 (05:40 -0400)]
kernel: bump 5.15 to 5.15.73

All patches automatically rebased

Build system: x86_64
Build-tested: bcm2711/RPi4B, mt7622/RT3200
Run-tested: bcm2711/RPi4B, mt7622/RT3200

Signed-off-by: John Audia <therealgraysky@proton.me>
2 years agokernel: bump 5.10 to 5.10.149
John Audia [Mon, 17 Oct 2022 18:34:40 +0000 (14:34 -0400)]
kernel: bump 5.10 to 5.10.149

No patches required modification.

Signed-off-by: John Audia <therealgraysky@proton.me>
2 years agokernel: bump 5.10 to 5.10.148
John Audia [Sat, 15 Oct 2022 10:31:21 +0000 (06:31 -0400)]
kernel: bump 5.10 to 5.10.148

No patches required modification.

Signed-off-by: John Audia <therealgraysky@proton.me>
2 years agoucode: update to latest Git HEAD
Jo-Philipp Wich [Tue, 18 Oct 2022 09:06:31 +0000 (11:06 +0200)]
ucode: update to latest Git HEAD

00af065 fs: expose `getdelim()` functionality through `fd.read()`
21ace5e lexer: fixes for regex literal parsing

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
2 years agofirewall4: update to latest Git HEAD
Jo-Philipp Wich [Tue, 18 Oct 2022 07:44:42 +0000 (09:44 +0200)]
firewall4: update to latest Git HEAD

7ae5e14 fw4: gracefully handle `null` return values from `fd.read("line")`

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
2 years agomt7621: netgear_ex6150: fix reboot loop by using LZMA loader
Petr Štetiar [Tue, 18 Oct 2022 06:55:46 +0000 (08:55 +0200)]
mt7621: netgear_ex6150: fix reboot loop by using LZMA loader

This fixes a well known "LZMA ERROR 1" error, reported previously on
numerous of similar devices.

Fixes: #10645
Signed-off-by: Petr Štetiar <ynezz@true.cz>
2 years agouhttpd: use acme hotplug
Glen Huang [Sun, 16 Oct 2022 04:55:07 +0000 (12:55 +0800)]
uhttpd: use acme hotplug

Reload uhttpd after certificates are renewed with acme.

Reviewed-by: Toke Høiland-Jørgensen <toke@toke.dk>
Signed-off-by: Glen Huang <i@glenhuang.com>
2 years agoarm-trusted-firmware-mvebu: add Methode eDPU support
Robert Marko [Wed, 21 Sep 2022 16:40:59 +0000 (18:40 +0200)]
arm-trusted-firmware-mvebu: add Methode eDPU support

Provide ATF support for Methode eDPU as well, this makes it easy for
OpenWrt users to update the included U-boot+ATF combo.

Signed-off-by: Robert Marko <robert.marko@sartura.hr>
2 years agouboot-mvebu: add Methode eDPU support
Robert Marko [Wed, 21 Sep 2022 16:39:22 +0000 (18:39 +0200)]
uboot-mvebu: add Methode eDPU support

Add support for building for Methode eDPU board, no patches are needed
as board has been upstreamed and is part of the 2022.10-rc releases.

Signed-off-by: Robert Marko <robert.marko@sartura.hr>
2 years agouboot-mvebu: update to 2022.10
Robert Marko [Tue, 27 Sep 2022 08:58:50 +0000 (10:58 +0200)]
uboot-mvebu: update to 2022.10

Update mvebu U-boot to 2022.10 to avoid backporting patches in order
to support Methode eDPU.

It also allows dropping existing patches as they are all backports.

Tested-by: Andre Heider <a.heider@gmail.com> # espressobin-v3-v5-1gb-2cs
Tested-by: Russell Morris <github@rkmorris.us> # espressobin-v3-v5-1gb-1cs
Tested-by: Josef Schlehofer <pepe.schlehofer@gmail.com> [Turris Omnia]
Signed-off-by: Robert Marko <robert.marko@sartura.hr>
2 years agoCI: kernel: check if patch are refreshed for each target
Christian Marangi [Sat, 15 Oct 2022 08:56:46 +0000 (10:56 +0200)]
CI: kernel: check if patch are refreshed for each target

Enforce refreshed patch for each target with kernel pr tests.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
2 years agoipq806x: 5.15: refresh target patches
Christian Marangi [Sun, 16 Oct 2022 22:39:45 +0000 (00:39 +0200)]
ipq806x: 5.15: refresh target patches

Refresh ipq806x target patches for kernel 5.15.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
2 years agobcm53xx: 5.10: refresh target patches
Christian Marangi [Sun, 16 Oct 2022 22:34:28 +0000 (00:34 +0200)]
bcm53xx: 5.10: refresh target patches

Refresh bcm53xx target patches for kernel 5.10.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
2 years agogeneric: 5.10: refresh hack patches
Christian Marangi [Sun, 16 Oct 2022 18:36:49 +0000 (20:36 +0200)]
generic: 5.10: refresh hack patches

Refresh hack patches for kernel 5.10.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
2 years agogeneric: 5.10: refresh backport patches
Christian Marangi [Sun, 16 Oct 2022 18:36:03 +0000 (20:36 +0200)]
generic: 5.10: refresh backport patches

Refresh backport patches for kernel 5.10.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
2 years agogeneric: 5.15: refresh hack patches
Christian Marangi [Sun, 16 Oct 2022 11:19:33 +0000 (13:19 +0200)]
generic: 5.15: refresh hack patches

Refresh hack patches for kernel 5.15.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
2 years agogeneric: 5.15: refresh backport patches
Christian Marangi [Sun, 16 Oct 2022 11:19:05 +0000 (13:19 +0200)]
generic: 5.15: refresh backport patches

Refresh backport patches for kernel 5.15.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
2 years agofirewall4: update to latest Git HEAD
Jo-Philipp Wich [Fri, 14 Oct 2022 15:32:28 +0000 (17:32 +0200)]
firewall4: update to latest Git HEAD

4fbf6d7 ruleset.uc: log forwarded traffic not matched by zone policies
c7201a3 main.uc: reintroduce set reload restriction
756f1e2 ruleset: fix emitting set_mark/set_xmark rules with masks
3db4741 ruleset: properly handle zone names starting with a digit
43d8ef5 fw4: fix formatting of default log prefix
592ba45 main.uc: remove uneeded/wrong set reload restrictions
b0a6bff tests: fix testcases
145e159 fw4: recognize `option log` and `option counter` in `config nat` sections
ce050a8 fw4: fall back to device if l3_device is not available in ifstatus

Fixes: #10639, #10965
Signed-off-by: Jo-Philipp Wich <jo@mein.io>
2 years agoucode: update to latest Git HEAD
Jo-Philipp Wich [Wed, 21 Sep 2022 21:33:56 +0000 (23:33 +0200)]
ucode: update to latest Git HEAD

4ae7072 fs: use `getline()` for line wise read operations
21ace5e lexer: fixes for regex literal parsing
00965fa lib: implement slice() function
76d396d main: implement print mode
7bbba78 compiler: optimize function return opcode generation
a45f2a3 lexer: improve regex literal handling
d64d5d6 vm: maintain export symbol tables per program
f4b4ded uloop: task: gracefully handle absent output callback
a58fe47 ubus: hold reference to underlying connection until deferred is concluded
e23b58a lib: uc_system(): retry waitpid() on EINTR

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
2 years agorpcd: update to latest Git HEAD
Jo-Philipp Wich [Wed, 21 Sep 2022 21:35:18 +0000 (23:35 +0200)]
rpcd: update to latest Git HEAD

8c852b6 ucode: write ucode runtime exceptions to stderr

Signed-off-by: Jo-Philipp Wich <jo@mein.io>
2 years agompc85xx: fix unset kernel symbol
David Bauer [Fri, 14 Oct 2022 19:58:46 +0000 (21:58 +0200)]
mpc85xx: fix unset kernel symbol

Signed-off-by: David Bauer <mail@david-bauer.net>
2 years agoimage: always rebuild kernel loaders
Matthias Schiffer [Sun, 27 Mar 2022 21:15:48 +0000 (23:15 +0200)]
image: always rebuild kernel loaders

Kernel loaders like the lzma-loader currently don't track changes to
their sources. This can lead to an old version of a loader to be used
when a build tree is not clean between builds.

As the loaders are tiny and the build times are insignificant, simply
force rebuilding them on every build to avoid this problem.

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
2 years agompc85xx: p1010: make TP-Link WDR4900 v1 build again
Matthias Schiffer [Sun, 27 Mar 2022 19:00:30 +0000 (21:00 +0200)]
mpc85xx: p1010: make TP-Link WDR4900 v1 build again

Add the spi-loader as a pre-kernel stage, so we can lift the kernel size
limit.

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
2 years agompc85xx: add SPI kernel loader for TP-Link TL-WDR4900 v1
Matthias Schiffer [Sun, 27 Mar 2022 18:57:01 +0000 (20:57 +0200)]
mpc85xx: add SPI kernel loader for TP-Link TL-WDR4900 v1

Similar to the lzma-loader on our MIPS targets, the spi-loader acts as
a second-stage loader that will then load and start the actual kernel.
As the TL-WDR4900 uses SPI-NOR and the P1010 family does not have support
for memory mapping of this type of flash, this loader needs to contain a
basic driver for the FSL ESPI controller.

Signed-off-by: Matthias Schiffer <mschiffer@universe-factory.net>
2 years agobusybox: nslookup: ensure unique transaction IDs for the DNS queries
Uwe Kleine-König [Fri, 14 Oct 2022 10:23:07 +0000 (12:23 +0200)]
busybox: nslookup: ensure unique transaction IDs for the DNS queries

On machines with a coarse monotonic clock (here: TP-Link RE200 powered
by a MediaTek MT7620A) it can happen that the two DNS requests (for A
and AAAA) share the same transaction ID. If this happens the second
reply is wrongly dropped and nslookup reports "No answer".

Fix this by ensuring that the transaction IDs are unique.

Signed-off-by: Uwe Kleine-König <uwe@kleine-koenig.org>
2 years agogeneric: 5.15: backport qca8k fixup patches for Big-Endian systems
Christian Marangi [Fri, 14 Oct 2022 13:10:52 +0000 (15:10 +0200)]
generic: 5.15: backport qca8k fixup patches for Big-Endian systems

Backport qca8k fixup patches for inband mgmt on Big-Endian systems.
This is needed for ath79 and mpc85xx targets that are Big-Endian and use
qca8k based switch.

Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
2 years agomediatek: mt7623: fix thermal zone
Bruno Umuarama [Fri, 14 Oct 2022 00:37:39 +0000 (00:37 +0000)]
mediatek: mt7623: fix thermal zone

Raising the temperatures for passive and active trips. @VA1DER
proposed at issue 9396 to remove passive trip. This commit relates to
his suggestion.

Without this patch. the CPU will be throttled all the way down to 98MHz
if the temperature rises even a degree above the trip point, and it was
further discovered that if the internal temperature of the device is
above the first trip point temperature when it boots then it will start
in a throttled state and even
$ echo disabled > /sys/class/thermal/thermal_zone0/mode
will have no effect.

The patch increases the passive trip point and active cooling map. The
throttling temperature will then be at 77°C and 82°C, which is still a
low enough temperature for ARM devices to not be in the real danger
zone, and gives some operational headroom.

Signed-off-by: Bruno Umuarama <anonimou_eu@hotmail.com>
2 years agomediatek: clean up mt7622 kernel config
Felix Fietkau [Wed, 12 Oct 2022 10:39:41 +0000 (12:39 +0200)]
mediatek: clean up mt7622 kernel config

Remove a few unused options to reduce kernel size

Signed-off-by: Felix Fietkau <nbd@nbd.name>
2 years agokernel: remove target specific CONFIG_IKCONFIG lines
Felix Fietkau [Tue, 11 Oct 2022 12:20:04 +0000 (14:20 +0200)]
kernel: remove target specific CONFIG_IKCONFIG lines

Use generic configuration for this feature

Signed-off-by: Felix Fietkau <nbd@nbd.name>
2 years agokernel: add missing lockdep config symbols for 5.15
Felix Fietkau [Tue, 11 Oct 2022 12:20:40 +0000 (14:20 +0200)]
kernel: add missing lockdep config symbols for 5.15

Signed-off-by: Felix Fietkau <nbd@nbd.name>
2 years agomediatek: fill wlan device entries in board.json on mt7622
Felix Fietkau [Fri, 30 Sep 2022 17:32:50 +0000 (19:32 +0200)]
mediatek: fill wlan device entries in board.json on mt7622

Signed-off-by: Felix Fietkau <nbd@nbd.name>
2 years agomac80211: use board.json provided phy names in generated default config
Felix Fietkau [Sat, 1 Oct 2022 14:04:03 +0000 (16:04 +0200)]
mac80211: use board.json provided phy names in generated default config

The phy will be automatically renamed on setup

Signed-off-by: Felix Fietkau <nbd@nbd.name>
2 years agomac80211: change the default config for a renamed wiphy
Felix Fietkau [Fri, 23 Sep 2022 12:39:59 +0000 (14:39 +0200)]
mac80211: change the default config for a renamed wiphy

use option phy to reference the device instead of path/macaddr

Signed-off-by: Felix Fietkau <nbd@nbd.name>
2 years agomac80211: fix detecting highest radio* config section index
Felix Fietkau [Mon, 10 Oct 2022 11:26:32 +0000 (13:26 +0200)]
mac80211: fix detecting highest radio* config section index

Deal with gaps by iterating over existing sections instead of counting

Signed-off-by: Felix Fietkau <nbd@nbd.name>
2 years agomac80211: rename phy according to board.json entries on bringup
Felix Fietkau [Fri, 30 Sep 2022 18:14:05 +0000 (20:14 +0200)]
mac80211: rename phy according to board.json entries on bringup

This allows phy names specified in board.json to be used directly instead of
the path option

Signed-off-by: Felix Fietkau <nbd@nbd.name>
2 years agomac80211: change default ifname to <phy>-<type><index>
Felix Fietkau [Fri, 23 Sep 2022 12:12:32 +0000 (14:12 +0200)]
mac80211: change default ifname to <phy>-<type><index>

This makes it clear, which phy a wlan device belongs to and also helps with
telling them apart by including the mode in the ifname.
Preparation for automatically renaming PHYs

Signed-off-by: Felix Fietkau <nbd@nbd.name>
2 years agobase-files: add helper functions for adding wlan device entries to board.json
Felix Fietkau [Fri, 30 Sep 2022 17:19:51 +0000 (19:19 +0200)]
base-files: add helper functions for adding wlan device entries to board.json

These will be used to give WLAN PHYs a specific name based on path specified
in board.json. The platform board.d script can assign a specific order based
on available slots (PCIe slots, WMAC device) and device tree configuration.

This helps with maintaining config compatibility in case the device path
changes due to kernel upgrades.

Signed-off-by: Felix Fietkau <nbd@nbd.name>
2 years agolibubox: update to the latest version
Felix Fietkau [Fri, 30 Sep 2022 11:54:58 +0000 (13:54 +0200)]
libubox: update to the latest version

ea56013409d5 jshn.sh: add json_add_fields function for adding multiple fields at once

Signed-off-by: Felix Fietkau <nbd@nbd.name>
2 years agoiwinfo: update to the latest version
Felix Fietkau [Fri, 14 Oct 2022 11:11:16 +0000 (13:11 +0200)]
iwinfo: update to the latest version

0496c722f1d7 nl80211: fix issues with renamed wiphy and multiple phy per device

Signed-off-by: Felix Fietkau <nbd@nbd.name>
2 years agomac80211: fix typo in netifd script
Felix Fietkau [Thu, 13 Oct 2022 19:49:38 +0000 (21:49 +0200)]
mac80211: fix typo in netifd script

Reported-by: Chad Monroe <chad.monroe@smartrg.com>
Fixes: 590eaaeed59a ("mac80211: fix issues in HE capabilities")
Signed-off-by: Felix Fietkau <nbd@nbd.name>
2 years agoairoha: fix imagebuilder generation by introducing generic target
Nick Hainke [Sun, 11 Sep 2022 19:53:20 +0000 (21:53 +0200)]
airoha: fix imagebuilder generation by introducing generic target

The generic imagebuilder does not have a generic in the name, although
this is the default naming scheme. Apply the same fix as for the octeon
target. Thanks to @dangowrt for reporting that the same issues applies
also for the airoha target.

Before the fix:
openwrt-imagebuilder-airoha.Linux-x86_64.tar.xz

After:
openwrt-imagebuilder-airoha-generic.Linux-x86_64.tar.xz

Signed-off-by: Nick Hainke <vincent@systemli.org>
2 years agohostapd: add measurement report value for beacon reports
Nick Hainke [Sat, 8 Jan 2022 07:55:21 +0000 (08:55 +0100)]
hostapd: add measurement report value for beacon reports

Add the measurement report value to the beacon reports send via ubus. It
is possible to derive from the measurement report if a station refused to
do a beacon report and why. It is important to know why a station refuses
to do a beacon-report. In particular, we should not request a beacon
report from a station again that refused a beacon-report before.

The rejection reasons can be found by looking at the bits defined by:
- MEASUREMENT_REPORT_MODE_ACCEPT
- MEASUREMENT_REPORT_MODE_REJECT_LATE
- MEASUREMENT_REPORT_MODE_REJECT_INCAPABLE
- MEASUREMENT_REPORT_MODE_REJECT_REFUSED

Suggested-by: Ian Clowes <clowes_ian@hotmail.com>
Signed-off-by: Nick Hainke <vincent@systemli.org>
2 years agobcm53xx: linksys ea6500-v2: extend RAM to full 256MB
Aleksey Nasibulin [Thu, 13 Oct 2022 08:22:53 +0000 (08:22 +0000)]
bcm53xx: linksys ea6500-v2: extend RAM to full 256MB

Linksys ea6500-v2 have 256MB of ram. Currently we only use 128MB.
Expand the definition to use all the available RAM.

Signed-off-by: Aleksey Nasibulin <alealexpro100@ya.ru>
[ wrap to 80 columns ]
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>
2 years agokernel: backport gnu11 upgrade
Olliver Schinagl [Sun, 4 Sep 2022 20:08:47 +0000 (22:08 +0200)]
kernel: backport gnu11 upgrade

Upstream has switched to gnu11 not too long ago. One advantage of
backporting this to these older kernels is, that we can encourage and
write better upstreamable kernels. E.g. the kernel devs prefer loop
declarations. Shrinking the master/local gap will be useful in these
cases.

Signed-off-by: Olliver Schinagl <oliver@schinagl.nl>
Acked-by: Christian Marangi <ansuelsmth@gmail.com>
2 years agobuild: image: use printf for ModelNameLimit16
Huangbin Zhan [Tue, 29 Mar 2022 16:15:29 +0000 (00:15 +0800)]
build: image: use printf for ModelNameLimit16

Fix syntax error on macos, for substr is undefined results according to
the POSIX standard.

From expr on macos:
 According to the POSIX standard, the use of string arguments length,
 substr, index, or match produces undefined results.  In this version of
 expr, these arguments are treated just as their respective string values.

By a simple test Makefile:

define ModelNameLimit16
$(shell expr substr "$(word 2, $(subst _, ,$(1)))" 1 16)
endef

define ModelNameLimit16_2
$(shell printf %.16s "$(word 2, $(subst _, ,$(1)))")
endef

hello:
echo $(call ModelNameLimit16, technicolor_tg582n-telecom-italia)
echo $(call ModelNameLimit16_2, technicolor_tg582n-telecom-italia)

The same output is produced.

echo tg582n-telecom-i
tg582n-telecom-i
echo tg582n-telecom-i
tg582n-telecom-i

Signed-off-by: Huangbin Zhan <zhanhb88@gmail.com>
[ wrap commit description to 80 columns and improve it ]
Signed-off-by: Christian Marangi <ansuelsmth@gmail.com>