The majority of our targets provide a default value for the variable
SUPPORTED_DEVICES, which is used in images to check against the
compatible on a running device:
SUPPORTED_DEVICES := $(subst _,$(comma),$(1))
At the moment, this is implemented in the Device/Default block of
the individual targets or even subtargets. However, since we
standardized device names and compatible in the recent past, almost
all targets are following the same scheme now:
device/image name: vendor_model
compatible: vendor,model
The equal redundant definitions are a symptom of this process.
Consequently, this patch moves the definition to image.mk making it
a global default. For the few targets not using the scheme above,
SUPPORTED_DEVICES will be defined to a different value in
Device/Default anyway, overwriting the default. In other words:
This change is supposed to be cosmetic.
This can be used as a global measure to get the current compatible
with: $(firstword $(SUPPORTED_DEVICES))
(Though this is not precisely an achievement of this commit.)
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
WNDR4700 uboot has an issue with decompressing kernel with default dictionary size (-d23 which is about 8MB).
Limiting lzma dictionary lowers memory footprint and allows device to boot the kernel.
The highest bootable dictonary size is 18, choosing 16 for an extra safety margin.
Kernel size befor and after:
-d23: 2663665 Bytes
-d16: 2892757 Bytes
Kernel size increased by 230kB (9%)
Fixes: FS#3258
Signed-off-by: Wiktor Stasiak <wiktor.stasiak@gmail.com>
All modifications made by update_kernel.sh run in a fresh clone
without any existing toolchains.
Build system: x86_64
Build-tested: ipq806x/R7800, ath79/generic, bcm27xx/bcm2711
Run-tested: ipq806x/R7800
No dmesg regressions, everything functional
Signed-off-by: John Audia <graysky@archlinux.us>
Move features common to all subtargets to the parent target.
Signed-off-by: Rui Salvaterra <rsalvaterra@gmail.com>
[split patch to make it target-specific, adjust commit title/message]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Manually rebased patches:
bcm27xx:
patches-5.4/950-0267-xhci-add-quirk-for-host-controllers-that-don-t-updat.patch
bcm53xx:
patches-5.4/180-usb-xhci-add-support-for-performing-fake-doorbell.patch
layerscape:
patches-5.4/802-can-0025-can-flexcan-add-LPSR-mode-support-for-i.MX7D.patch
patches-5.4/808-i2c-0002-MLK-10893-i2c-imx-add-irqf_no_suspend.patch
patches-5.4/820-usb-0016-MLK-16735-usb-host-add-XHCI_CDNS_HOST-flag.patch
Removed since could be reverse-applied by quilt:
mediatek:
patches-5.4/0700-arm-dts-mt7623-add-missing-pause-for-switchport.patch
All modifications made by update_kernel.sh
Build system: x86_64
Build-tested: ipq806x/R7800, ath79/generic, bcm27xx/bcm2711, x86_64
Run-tested: ipq806x/R7800, x86_64
No dmesg regressions, everything functional
Signed-off-by: John Audia <graysky@archlinux.us>
Tested-by: Curtis Deptuck <curtdept@me.com> [x86_64]
Rebase of 802-can-0025-can-flexcan-add-LPSR-mode-support-for-i.MX7D.patch
Signed-off-by: Yangbo Lu <yangbo.lu@nxp.com>
Signed-off-by: Kevin Darbyshire-Bryant <kevin@darbyshire-bryant.me.uk>
We use 5.4 on all targets by default, and 4.19 has never been released
in a stable version. There is no reason to keep it.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
This config option was moved to the generic kernel configuration.
Fixes: ab1bd57656 ("kernel: move F2FS_FS_XATTR and F2FS_STAT_FS symbols to generic")
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Extended attributes are required for overlayfs and have hence been long
ago enabled for jffs2, but should be enabled unconditionally for all
other filesystems which may potentially serve as overlayfs' upper
directory. Previously it was inconsistently added in multiple targets.
Add symbols to generic kernel config and remove all *_XATTR symbols
from target configs.
Signed-off-by: Paul Spooren <mail@aparcar.org>
[keep things as they are for squashfs, improve commit message]
Signed-off-by: Daniel Golle <daniel@makrotopia.org>
Devices with PCIe-Switches like the WNDR4700, MR24 and WNDAP660
need to have the interrupts property specified in the device-tree
for the legacy pci interrupt signaling method to work.
If the proper interrupt value is not specified, the default INTA
IRQ 12 is taken for all devices. This is especially bad, if the
device is setup to use INTC, because these interrupts will not
be serviced.
Russell Senior reported his experience on the MR24:
"The symptom is client devices can't see the beacons.
Wifi ifaces appear, can scan and hear other networks,
but clients can't see the MR24's SSIDs."
(The interrupts-property on the WNDAP620 was optional since it
uses INTA by default. Likewise the MX60W is in the same category)
Reported-by: Russell Senior <russell@personaltelco.net>
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
This patch fixes various typos or tab-vs-space issues in
the APM821XX device targets Device-Tree source files.
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
This patch silences the following warnings:
>netgear-wndr4700.dts:168.3-13:Warning (reg_format): /plb/sata@bffd1800/sata-port@0:reg:property has invalid length (4 bytes) (#address-cells == 2, #size-cells == 1)
>netgear-wndr4700.dts:167.26-170.4: Warning (avoid_default_addr_size):/plb/sata@bffd1800/sata-port@0: Relying on default #address-cells value
>netgear-wndr4700.dts:167.26-170.4: Warning (avoid_default_addr_size):/plb/sata@bffd1800/sata-port@0: Relying on default #size-cells value
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
This patch adds the pcie-switch and bridge configuration of the
WNDR4700.
This allows to get rid of the legacy firmware monikers and drop
the usbport LED declaration.
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
In order to support SAE/WPA3-Personal in default images. Replace almost
all occurencies of wpad-basic and wpad-mini with wpad-basic-wolfssl for
consistency. Keep out ar71xx from the list as it won't be in the next
release and would only make backports harder.
Build-tested (build-bot settings):
ath79: generic, ramips: mt7620/mt76x8/rt305x, lantiq: xrx200/xway,
sunxi: a53
Signed-off-by: Petr Štetiar <ynezz@true.cz>
[rebase, extend commit message]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
This PR is a blend of several kernel bumps authored by ldir taken from his
staging tree w/ some further adjustments made by me and update_kernel.sh
Summary:
Deleted upstreamed patches:
generic:
742-v5.5-net-sfp-add-support-for-module-quirks.patch
743-v5.5-net-sfp-add-some-quirks-for-GPON-modules.patch
bcm63xx:
022-v5.8-mtd-rawnand-brcmnand-correctly-verify-erased-pages.patch
024-v5.8-mtd-rawnand-brcmnand-fix-CS0-layout.patch
mediatek:
0402-net-ethernet-mtk_eth_soc-Always-call-mtk_gmac0_rgmii.patch
Deleted patches applied differently upstream:
generic:
641-sch_cake-fix-IP-protocol-handling-in-the-presence-of.patch
Manually merged patches:
generic:
395-v5.8-net-sch_cake-Take-advantage-of-skb-hash-where-appropriate.patch
bcm27xx:
950-0132-lan78xx-Debounce-link-events-to-minimize-poll-storm.patch
layerscape:
701-net-0231-enetc-Use-DT-protocol-information-to-set-up-the-port.patch
Build system: x86_64
Build-tested: ath79/generic, bcm27xx/bcm2708, bcm27xx/bcm2711,
imx6, mvebu/cortexa9, sunxi/a53
Run-tested: Netgear R7800 (ipq806x)
No dmesg regressions, everything functional
Signed-off-by: John Audia <graysky@archlinux.us>
Tested-By: Lucian Cristian <Lucian.cristian@gmail.com> [mvebu]
Tested-By: Curtis Deptuck <curtdept@me.com> [x86/64]
[do not remove 395-v5.8-net-sch_cake-Take-advantage-... patch,
adjust and refresh patches, adjust commit message]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Tested-By: John Audia <graysky@archlinux.us> [ipq806x]
ifconfig is effectively deprecated for quite some time now. Let's
replace the remaining occurrences for our target setup by the
corresponding ip commands now.
Note that this does not touch ar71xx, as it will be dropped anyway,
and changing it would only make backports harder.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
The compressed image that the buildbots are building is too large for
the netgear uboot and it crashes and soft-bricks the device.
| Uncompressing Kernel Image ...
| LZMA: uncompress or overwrite error 1 - must RESET board to recover
The whole target likely needs to be switched zImage which is a major
hassle due to powerpc's legacy bootwrapper setup as compared to ARM.
So for now, disable the device.
Reported-by: Wiktor Stasiak (FS#3258)
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
The DEVICE_DTS variable always matches the device definition name,
just with "_" replaced by "-". Thus, create a DEVICE_DTS definition
in Device/Default and drop all the individual statements.
If necessary in the future, local DEVICE_DTS will still overwrite
that default.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Fixes:
- CVE-2020-10757
The "mtd: rawnand: Pass a nand_chip object to nand_release()" commit was
backported which needed some adaptations to other code.
Run tested: ath79
Build tested: ath79
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This drops the shebang from all target files for /lib and
/etc/uci-defaults folders, as these are sourced and the shebang
is useless.
While at it, fix the executable flag on a few of these files.
This does not touch ar71xx, as this target is just used for
backporting now and applying cosmetic changes would just complicate
things.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
With several subtargets, the image/Makefile becomes crowded after a
while. Many targets have moved their device definitions to $subtarget.mk
files to have them more organized, let's do this here as well.
While at it, also move subtarget-specific build recipes.
Cc: Christian Lamparter <chunkeey@gmail.com>
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Acked-by: Christian Lamparter <chunkeey@gmail.com>
Since DEVICE_TYPE cannot be set per device, just set DEVICE_TYPE
to "nas" for the entire subtarget, which only contains this single
device.
Note that while this looks like a cosmetic change in combination
with the previous patches, this particular patch actually changes
the packages for the device.
Suggested-by: Christian Lamparter <chunkeey@gmail.com>
Cc: Christian Lamparter <chunkeey@gmail.com>
Cc: Sungbo Eo <mans0n@gorani.run>
Cc: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
DEVICE_TYPE is a target/subtarget variable, and it does not have
any effect when set in a device definition. It can only be set
in a target's or subtarget's Makefile.
Consequently, having it set anyway is misleading, so this drops
all cases.
This effectively reverts the following commits:
7a1497fd60 ("apm821xx: MBL: set DEVICE_TYPE to NAS")
5b4765c93a ("gemini: Classify Raidsonic NAS IB-4220-B as a NAS")
cdc6de460b ("gemini: D-Link DNS-313 is a NAS")
For the following commit, the variable was set when adding device
support:
27b2f0fc0f ("kirkwood: add support for Iomega Storcenter ix2-200")
Cc: Christian Lamparter <chunkeey@gmail.com>
Cc: Sungbo Eo <mans0n@gorani.run>
Cc: Linus Walleij <linus.walleij@linaro.org>
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
With the "real HZ" debate out of the way, let's actually
use the apm821xx's default upstream config file at
arch/powerpc/configs/44x/bluestone_defconfig. From what
I can tell, it sets NO_HZ (well, this platform was for
NAS, so this is a bit unexpected) and remove any specific
HZ_$VALUE symbol.
Sadly, Daniel Engberg didn't run any before/after netperf
tests, because it would have been such a slam dunk across
the boards. In case of the apm821xx the tcp tx/rx
performance improved ~14% (from 600Mbps to 700Mps).
This now causes the emac to drop frames too, so let's see
if this is causes more problems or not.
This patch includes a refresh of the configuration too.
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
Most targets upstream use 250Hz or even 1000Hz by default while
100Hz is hardcoded in OpenWrt's default config. Use upstream default
except for apm821xx which hardsets 1000Hz instead of platform default of
250Hz.
Signed-off-by: Daniel Engberg <daniel.engberg.lists@pyret.net>
[Apply same changes to 5.4]
Signed-off-by: Kevin Darbyshire-Bryant <ldir@darbyshire-bryant.me.uk>
This patch prepares the WNDR4700 to use the HDD sensor for
the thermal zone. While the kernel's thermal.txt device-tree
binding documentation files talks about supporting multiple
sensors for a zone. This sadly is NOT the case. Even the most
current upstream kernels (5.6-rc) supports just >one< sensor
per zone: (driver/base/of-thermal.c:886)
| * REVIST: for now, the thermal framework supports only
| * one sensor per thermal zone. Thus, we are considering
| * only the first two values as slope and offset.
I do hope that this warning will prevent others wasteing time
on trying to figure out why their multi-sensor thermal-zones
definitions are not working as specified.
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
This patch adds the hwmon-drivetemp to the device.
It also adds device-tree bindings. This can be useful to
automate external fans which can be controlled for example
by either an unused sata-port or by the usb-power regulator.
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
Now that check-size uses IMAGE_SIZE by default, we can skip the argument from
image recipes to reduce redundancy.
Signed-off-by: Sungbo Eo <mans0n@gorani.run>
[do not touch ar71xx]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
There is no such role as target maintainer anymore, one should always
send corresponding changes for the review and anyone from the commiters
is allowed to merge them or eventually use the hand break and NACK them.
Lets make it clear, that it is solely a community doing the maintenance
tasks.
Signed-off-by: Petr Štetiar <ynezz@true.cz>
Acked-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Acked-by: Piotr Dymacz <pepe2k@gmail.com>
This patch removes the 4.14 kernel support from the apm821xx target.
The 4.19 kernel has been available and stable for a while and the 5.4
kernel support has been tested successfully on real hardware as well.
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
IMAGE_SIZE is widely used in many targets. Declare it in the default template to
clean up redundant code. This also prevents deriving IMAGE_SIZE unintentionally
from the previously defined device.
While at it, remove duplicate KERNEL_SIZE declaration.
Signed-off-by: Sungbo Eo <mans0n@gorani.run>
This hunk became obsolete the moment when our
gpio-button-hotplug learned how to deal with
interrupt-supported gpio buttons. The gpio driver
never supported interrupt handling, so these
properties never served any use (outside of a
enhanced ppc4xx-gpio driver that was dropped).
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
This patch adds important NAND config symbols.
These are necessary as otherwise the devices
won't find the rootfs on the NAND chips.
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
Signed-off-by: David Bauer <mail@david-bauer.net>
[refreshed]
Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
* Sync the patches with the changes done for kernel 4.19
* Use KERNEL_TESTING_PATCHVER
* Refresh the configuration
* Fix multiple compile bugs in the patches
* Only add own ag71xx files for kernel 4.19 and use upstream version for
5.4.
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Remove CC_HAS_ASM_GOTO, CC_IS_GCC and GCC_VERSION kernel config options
which are set at runtime and which should be ignored now.
Signed-off-by: Petr Štetiar <ynezz@true.cz>
This unifies MAC address patch functions and moves them to a
common script. While those were implemented differently for
different targets, they all seem to do the same. The number of
different variants is significantly reduced by this patch.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
This moves the almost identical calibration data extraction
functions present multiple times in several targets to a single
library file /lib/functions/caldata.sh.
Functions are renamed with more generic names to merge different
variants that only differ in their names.
Most of the targets used find_mtd_chardev, while some used
find_mtd_part inside the extraction code. To merge them, the more
abundant version with find_mtd_chardev is used in the common code.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
[rebase on latest master; add mpc85xx]
Signed-off-by: David Bauer <mail@david-bauer.net>
This replaces deprecated backticks by more versatile $(...) syntax.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
[decapitalized patch subject at submitter's request]
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
This patch removes a typo (extra "0") so that the 'cpu-alert6'
step is triggered once the system reaches 85°C.
Note: Unless the WNDR4700 is placed in an hot oven, the
hardware-monitor will never reach this value.
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
The allocation of LEBs to ubi volumes is handled by the sysupgrade script:
package/base-files/files/lib/upgrade/nand.sh
and the ubimkvol and or ubirsvol command. Therefore, padding of the
kernel blob is not needed at all, so use cat instead of dd. The
BLOCKSIZE variable was only used in the dd command. In any case, 63k
made no sense for the way BLOCKSIZE was being used.
63k (64512) does make sense for DTB_SIZE because of the offsets expected
by u-boot given extant u-boot-env variables.
Tested on Meraki MR24.
Signed-off-by: Russell Senior <russell@personaltelco.net>
It's a variable set by procd that should replace hardcoded
/tmp/sysupgrade.tgz.
This change requires the most recent procd with the commit 0f3c136
("sysupgrade: set UPGRADE_BACKUP env variable").
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
$CONF_TAR shouldn't be assumed to always point to the sysupgrade.tgz.
This change makes code more generic and allows refactoring $CONF_TAR.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
1) Add BACKUP_FILE and use it when copying an archive to be restored
after sysupgrade (on the next preinit).
2) Use CONF_TAR for copying backup prepared by the /sbin/sysupgrade
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
This explicitly lets stage2 know if partitions should be preserved. No
more "touch /tmp/sysupgrade.always.overwrite.bootdisk.partmap" hack.
Signed-off-by: Rafał Miłecki <rafal@milecki.pl>
This changes size and offset set for WiFi caldata extraction and
MAC address adjustment to hexadecimal notation.
This will be much clearer for the reader when numbers are big, and
will also match the style used for mtd-cal-data in DTS files.
Since dd cannot deal with hexadecimal notation, one has to convert
back to decimal by simple $(($hexnum)).
Acked-by: Alexander Couzens <lynxis@fe80.eu>
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
This changes the offsets for the MAC address location in
mtd_get_mac_binary* and mtd_get_mac_text to hexadecimal notation.
This will be much clearer for the reader when numbers are big, and
will also match the style used for mtd-mac-address in DTS files.
(e.g. 0x1006 and 0x5006 are much more useful than 4102 and 20486)
Acked-by: Alexander Couzens <lynxis@fe80.eu>
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
The 4.19 code has been available for some time now.
Since the f2fs / loop issue has been under control
I think it's time the target gets to move to it.
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
This patch replaces the OpenWrt' favoured
CRC32 Sarwate algorithm with the kernel
recommended SLICE8 CRC32 version.
This change alone lifted btrfs writes from 40 MiB/s
to 54 MiB/s on my My Book Live (SLICE4 managed 48 MiB/s).
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
"There are often transient line events when
the system is powered up and initialized and it
is often necessary for the gpio_chip driver to
clear any interrupt flags in hardware before
setting up the gpio chip, especially the
irqchip portions of it."
<http://lists.infradead.org/pipermail/openwrt-devel/2019-June/017630.html>
This patch adds a fix for the APM821XX's interrupt
controller to clear any bogus pending toggled
interrupts that happens on various APM821XX boards
on boot.
The patch also changes the debouce-interval from the
default 5ms debounce interval to 60ms all around.
The default setting caused on occasions that the button
state became stuck in a pressed state, even though the
button was released.
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
Refreshed all patches.
This bump contains upstream commits which seem to avoid (not properly fix)
the errors as seen in FS#2305 and FS#2297
Altered patches:
- 403-net-mvneta-convert-to-phylink.patch
- 410-sfp-hack-allow-marvell-10G-phy-support-to-use-SFP.patch
Compile-tested on: ar71xx, cns3xxx, imx6, mvebu, x86_64
Runtime-tested on: ar71xx, cns3xxx, imx6, x86_64
Signed-off-by: Koen Vandeputte <koen.vandeputte@ncentric.com>
This patch fixes the active_low setting and
converts all of the physical keys on the wndr4700
to utilize the interrupt-driven gpio-keys driver
over the polled version.
The sdcard-insertion hack has been removed since the
block-subsystem will now be polling the device instead.
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
Commit "generic: ar8216: add mib_poll_interval switch attribute" sets
mib-poll-interval as disabled by default (was set to 2s), so it makes
switch LEDs trigger disfunctional on devices which don't have
mib-poll-interval set.
So this patch sets mib-poll-interval to 500ms on devices which have
ar83xx switch connected to emac0, as the same value was set for built in
switches in 443fc9ac35 ("ath79: use ar8216 for builtin switch").
Signed-off-by: Petr Štetiar <ynezz@true.cz>
The SD-Card polling is now implemented by default in the
fs-tools block-mount utility package. It might not be as
fast as the current detection method since the polling
time is 2 Seconds, but it's much less of an hack.
Signed-off-by: Christian Lamparter <chunkeey@gmail.com>
Recently, upgrade device autodetection has been added to the mvebu target.
This exposes some shortcomings of the generic export_bootdevice function,
e.g. on the Turris Omnia: export_bootdevice silently reports the root
partition to be the boot device. This makes the sysupgrade process fail at
several places.
Fix this by clearly distinguishing between /proc/cmdline arguments which
specify the boot disk, and those which specify the root partition. Only in
the latter case, strip off the partition, and do it consistently.
root=PARTUUID=<pseudo PARTUUID for MBR> (any partition) and root=/dev/*
(any partition) are accepted.
The root of the problem is that the *existing* export_bootdevice in
/lib/upgrade/common.sh behaves differently, if the kernel is booted with
root=/dev/..., or if it is booted with root=PARTUUID=...
In the former case, it reports back major/minor of the root partition,
in the latter case it reports back major/minor of the complete boot disk.
Targets, which boot with root=/dev/... *and* use export_bootdevice /
export_partdevice, have added workarounds to this behaviour, by specifying
*negative* increments to the export_partdevice function.
Consequently, those targets have to be adapted to use positive increments,
otherwise they are broken by the change to export_bootdevice.
Fixes: 4e8345ff68 ("mvebu: base-files: autodetect upgrade device")
Signed-off-by: Klaus Kudielka <klaus.kudielka@gmail.com>
Tested-by: Tomasz Maciej Nowak <tomek_n@o2.pl>
If the target supports a newer kernel version that is not used by default
yet, it can be enabled with this option
Signed-off-by: Felix Fietkau <nbd@nbd.name>
