Port 21 definition was missed during addition of LGS328C.
Add it to the dts.
Fixes: 853d73f ("realtek: add support for Linksys LGS328C")
Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
Link: https://github.com/openwrt/openwrt/pull/21793
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 75fd2497ac)
Link: https://github.com/openwrt/openwrt/pull/21900
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Commit 78bf3a5f44 ("realtek: dsa: Fix rate control initialization") enabled
code setting up the "storm control" feature. This casued a speed regression
on rtl838x, reducing the effective max speed per port from line rate to around
500 Mbits/s.
Storm control is a policy feature with a number of input parameters depending
on use case and environment. It is not possible to define a meaningful static
policy in the driver. The problem isn't just the arbitrary limits in the
current code. Such features require userspace interfaces.
Drop this code for now. It wasn't missed while it was disabled.
Fixes: 78bf3a5f44 ("realtek: dsa: Fix rate control initialization")
Link: https://github.com/openwrt/openwrt/issues/21692
Signed-off-by: Bjørn Mork <bjorn@mork.no>
Link: https://github.com/openwrt/openwrt/pull/21877
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 3972aeaf7c)
Link: https://github.com/openwrt/openwrt/pull/21900
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Setup for DSA QOS on RTL839x accesses unitialized memory. For some
reason the handover of the priv structure was realized via global
intermediate variable switch_priv. During refactoring for adbb9a6
("realtek: dsa: rtl83xx: fix init section mismatch") this was not
noticed. Since then RTL839x devices crash during startup.
Fix this by using standard handover via function parameters.
Fixes: e81affb9 ("realtek: dsa: rtl83xx: fix init section mismatch")
Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
Link: https://github.com/openwrt/openwrt/pull/21703
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit ac96463944)
Link: https://github.com/openwrt/openwrt/pull/21900
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Compilation currently spits this message:
WARNING: modpost: vmlinux: section mismatch in reference:
rtl83xx_sw_probe+0x6a4 (section: .text.rtl83xx_sw_probe)
-> rtl83xx_setup_qos (section: .init.text)
That means that we have a "normal" function caller (can be
called during the whole uptime) and a "initialization" function
callee (only available during init.
Fix this and directly fix the unwanted family checks.
Fixes: 012e0091 ("realtek: dsa: avoid use-after-free")
Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
Link: https://github.com/openwrt/openwrt/pull/21690
(cherry picked from commit adbb9a6423)
Link: https://github.com/openwrt/openwrt/pull/21900
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The realtek target uses some functions marked __init for initialization.
However, that means they can only be called once when compiled in and
afterwards the memory occupied by them is freed and potentially reused.
Some "impossible" (code at a given location can't crash in the way it
does) crashes can be caused by this because upon re-execution of those
functions, garbage gets executed. Such re-execution can happen for
deferred probes or repeated probes.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Link: https://github.com/openwrt/openwrt/pull/21504
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
(cherry picked from commit a91c3abe83)
Link: https://github.com/openwrt/openwrt/pull/21900
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The mac_addr variable was not zero-initialized, causing weird side effects
when the memory contents were a valid MAC address.
Signed-off-by: Carl-Daniel Hailfinger <c-d.hailfinger.devel.2006@gmx.net>
Link: https://github.com/openwrt/openwrt/pull/21504
Signed-off-by: Stijn Tintel <stijn@linux-ipv6.be>
(cherry picked from commit 19b8d391a9)
These patches fix bugs in a patch we backported.
These patch were cherry picked from upstream Linux because it references
a patch we backported in the fixes tag.
The patches were reordered to match the ordering in the upstream Linux kernel.
Fixes: 93173aee96 ("qualcommbe: ipq95xx: Add initial support for new target")
Link: https://github.com/openwrt/openwrt/pull/21366
(cherry picked from commit 5230157a16)
Link: https://github.com/openwrt/openwrt/pull/21390
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This patch fixes a bug in a patch we backported.
This patch was cherry picked from upstream Linux because it references a
patch we backported in the fixes tag.
The first two patches are providing function needed by the last patch.
Fixes: 813ecda1f3 ("generic: backport phylink patches for PCS/PHY caps OPs")
Link: https://github.com/openwrt/openwrt/pull/21366
(cherry picked from commit 69d76e3a14)
Link: https://github.com/openwrt/openwrt/pull/21390
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
This patch fixes a bug in a patch we backported.
This patch was cherry picked from upstream Linux because it references a
patch we backported in the fixes tag.
The function phy_id_compare_vendor() is needed by the fix.
Refresh 782-05-v6.16-net-phy-Add-support-for-Aeonsemi-AS21xxx-PHYs.patch with the upstream code.
Fixes: 08a616b216 ("generic: backport support for Aeonsemi AS21xxx PHY")
Link: https://github.com/openwrt/openwrt/pull/21366
(cherry picked from commit b3318fe819)
Link: https://github.com/openwrt/openwrt/pull/21390
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The GS110TUP's lan9 port is connected via a QSGMII PHY to SERDES 2, and
therefore should use the SWITCH_PORT_SDS macro instead of SWITCH_PORT. This
was missed in e956adfe because the GS110TUP is not particularly well
documented and the old code was confusing.
lan10 is an SFP and doesn't have an onboard PHY, so also remove its
associated PHY references and update it to match other devices' SFP ports.
Fixes: https://github.com/openwrt/openwrt/issues/21324
Signed-off-by: Jacob Potter <jacob@j4cbo.com>
Link: https://github.com/openwrt/openwrt/pull/21346
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit b2c0182f26)
RTL930x devices have highmem starting address at 0x20000000.
The Linksys LGS328C highmem definition is wrongly shared with
the larger LGS352C RTL931x model and starts at 0x90000000.
Fix it by splitting the definition.
Fixes: 853d73f ("realtek: add support for Linksys LGS328C")
Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
Link: https://github.com/openwrt/openwrt/pull/21262
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
(cherry picked from commit 5d7470d4ca)
Use the correct identifier 'rtsds_of_match' instead of
'rtsds_mdio_of_match' because the latter doesn't exist.
This doesn't cause an error for 6.12. However, with 6.18 the
implementation of MODULE_DEVICE_TABLE has changed to use 'static' and
'used' [1] instead of 'extern' and 'unused' [2].
[1] 7d0a66e4bb/include/linux/module.h (L260)
[2] adc218676e/include/linux/module.h (L249)
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/21182
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit d519a3ee86)
As a first real usage of the new SerDes struct, move the polarity
configuration there. It was previously located in the global rtpcs_ctrl
struct as an array, indexed by SerDes id. Because this is per-SerDes
information, the new SerDes struct is the correct place to live in.
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/21146
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 56a71f3c82)
By using references to pre-initiated SerDes instances instead of plain
SerDes number, there is no need to check for the range anymore in
various places. During driver/pcs init it is ensured that only valid
SerDes will reach the configuration functions.
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/21146
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 18eea05483)
Also switch set_autoneg (and related helper rtpcs_sds_modify) to the
SerDes struct instead of the plain SerDes id by using just the reference
to the SerDes instance instead of (ctrl, sds_id) tuple. This completes
the transition.
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/21146
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit d8dd8bd88f)
Make use of the previously added SerDes struct in SerDes setup and all
functions in its call path by removing (ctrl, sds_num) being passed to
every function call and instead just pass the reference to the
corresponding SerDes instance.
Various SerDes calculations for even, odd and neighbor are unified by
switching to previously introduced helpers.
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/21146
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 1fccb0eb16)
Drop usage of the to-be-phased-out SerDes id stored in rtpcs_link and
use the reference to the SerDes instance to use the embedded id in
rtpcs_serdes instead.
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/21146
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 4d9400f86b)
Upon creation of a phylink_pcs instance by calling rtpcs_create, assign
a reference to the corresponding SerDes to the link structure. In the
next step, this should be used everywhere instead of the plain SerDes
number.
Rename the field used to hold the SerDes number from 'sds' to 'sds_num'
and name the new field 'sds' to make clear what is what.
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/21146
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 69bbcc685b)
Add dedicated helpers to get references to even, odd and neigbor SerDes
if needed. This should replace the various calculations scattered
throughout the code, providing a unified way to work with adjacent
SerDes.
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/21146
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 232c1fb14a)
Add a separate structure for a SerDes. This is needed to appropriately
store per-SerDes information, which in turn is needed for future work.
Additionally, it's intended to reduce boilerplate and several
inconsistencies.
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/21146
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 6b3f8fb16f)
Use a separate configuration field for the number of SerDes for each
variant of the Realtek Otto family. Add this field to the config
structure, assign it and use it during driver probe. This narrows
possible error cases and is needed for upcoming extensions.
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/21146
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 6cf33aacfe)
The Realtek SerDes mode capabilities do not map 1:1 to the
PHY_INTERFACE_MODE_* modes used in the kernel and passed to the PCS.
For example, some PHY chips use the proprietary XSGMII mode for which
there isn't an equivalent in the kernel, or HSGMII.
In the past, this led to problems and confusion using kernel's XGMII to
handle the XSGMII mode, and needed a downstream patch for HSGMII. They
have been solved/worked around for now, but XSGMII is currently not
implemented at all. And who knows what might come in the future.
To make our life easier, introduce a dedicated internal representation
of SerDes modes which differs from kernel's PHY_INTERFACE_MODE_*. This
allows us to map "external" modes to different internal modes as needed
instead of carrying the PHY_INTERFACE_MODE_* through the whole SerDes
configuration code. The PCS driver needs to map PHY_INTERFACE_MODE_* to
RTPCS_SDS_MODE_* in pcs_config, and the latter should be used as the
only one.
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/21146
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit a4d6e10bf2)
Drop the unused and broken function rtpcs_930x_sds_clock_wait from the
PCS driver. The proper working variant is already some lines above and
called rtpcs_930x_sds_wait_clock_ready.
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/21146
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit decc4f6ba8)
Since the beginning, the PCS driver had the ability to call its
rtpcs_create without a reference to a valid PCS node. A comment in the
code mentions that this is done for RTL838X and its built-in octa-PHY
which is connected directly instead of via a SerDes. Further
explanations are not provided.
Drop this ability and make the rtpcs_create call in the dsa driver
conditional. As the built-in PHY of RTL838X isn't attached to a SerDes,
there is no obvious point of having the PCS driver in that chain. The
ports are marked as internal and have no pcs-handle, thus no phylink_pcs
instance should be created.
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/21146
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit a58e41e522)
Commit 3c073b5cb2 cleaned up the debugfs creation in
mdio-realtek-otto-serdes driver to not explicitly check if the root
directory already exists. This is fine because kernel handles the case
properly so there's no need to check anymore.
However, this pollutes the boot log with:
[..] debugfs: 'realtek_otto_serdes' already exists in '/'
[..] debugfs: 'realtek_otto_serdes' already exists in '/'
[..] debugfs: 'realtek_otto_serdes' already exists in '/'
[..] debugfs: 'realtek_otto_serdes' already exists in '/'
[..] debugfs: 'realtek_otto_serdes' already exists in '/'
[..] debugfs: 'realtek_otto_serdes' already exists in '/'
[..] debugfs: 'realtek_otto_serdes' already exists in '/'
[..] debugfs: 'realtek_otto_serdes' already exists in '/'
[..] debugfs: 'realtek_otto_serdes' already exists in '/'
[..] debugfs: 'realtek_otto_serdes' already exists in '/'
[..] debugfs: 'realtek_otto_serdes' already exists in '/'
Now, the root directory creation is attempted multiple times, causing
the kernel to print an error message because the directory already
exists.
Fix this by moving the SerDes loop into rtsds_debug_init and only try
to create the root debugfs directory once.
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/21179
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 8e4730fd60)
This will free memory automatically during driver unloading.
Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
Link: https://github.com/openwrt/openwrt/pull/21157
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit a4011797c2)
debugfs_create_dir() has a proper logic to handle existing directories.
Skip the manual test. Additionally quit early if directory creation fails.
Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
Link: https://github.com/openwrt/openwrt/pull/21157
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit 3c073b5cb2)
Calculating the backing serdes of a given frontend serdes does
not need any info about the control structure. Drop the reference.
Signed-off-by: Markus Stockhausen <markus.stockhausen@gmx.de>
Link: https://github.com/openwrt/openwrt/pull/21157
Signed-off-by: Robert Marko <robimarko@gmail.com>
(cherry picked from commit dcbc8722e9)
RTL838X SerDes is now completely managed by the PCS driver so it's time
to remove all the unused leftovers from DSA and PHY drivers to have that
finally separated.
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/20876
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Remove all pseudo-PHYs and phy-handle properties from DTS of RTL838X
devices. RTL838X SerDes is now handled by PCS driver and thus not
treated as PHY anymore.
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/20876
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
After having moved the configuration code and sequences from PHY and
DSA drivers to the PCS driver, add the hooks in PCS driver and remove
calls in PHY and DSA drivers to let PCS driver setup the SerDes
entirely on its own.
Also add pcs-handle to device tree definitions for most of the switch
ports because, due to the refactoring of the SerDes configuration, this
is needed now for all SerDes-attached ports.
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/20876
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The previous commit just imported some code as-is and commented it.
It needs heavy adjustments to compile and work within the PCS driver.
Do that now to that extent that it can be used within the driver. More
cosmetics and improvements will be done later.
Split the once-for-all SerDes configuration into the usual flow where
each SerDes is configured separately and on its own, as requested by the
PCS subsystem.
Move mode setting and patching into proper functions which are called
during SerDes configuration. Some configuration sequences are broken up
and moved into the SerDes configuration flow, e.g. reset sequences
because they were usually a single/few values applied to all SerDes at
once before.
Add proper configuration for SerDes 4 QSGMII to be able to setup this
mode properly on our own.
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/20876
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Import functions 'rtl8380_sds_rst', 'rtl8380_sds_power',
'rtl8380_configure_serdes' and 'rtl83xx_config_interface' from DSA and
PHY driver respectively but comment the code for now.
The code needs heavy adjustments to make it compile and work. To make
this as transparent as possible, do that in two stages.
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/20876
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
In the PHY driver, firmware files were used to store configuration
values for the SerDes which need to be applied upon initialization.
There are several issues which prevent to just take that over into the
PCS driver:
* SerDes and PHY parts are mixed within a firmware file
* SerDes access in PHY driver is based on writing into the switch's
global register space; PCS driver uses access via MDIO interface
--> destination values do not match
* firmware file format is not SerDes-agnostic
* no documentation or script for the "old" firmware files
Unfortunately, there is no proper firmware format yet where to take over
the required sequences. Thus, extract the sequences needed for RTL838X
SerDes, transform them to work with the MDIO based access and put them
as functions in the PCS driver.
Note that this should just be a temporary solution. In a next step, a
proper firmware format should be established and all configuration
sequences currently in the code should be moved into firmware files.
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/20876
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Add a new hook called 'init_serdes_common' to be able to perform
initialisations or anything else subject to all SerDes. This hook is
called in the end of 'rtpcs_probe' after everything else is done.
This is meant primarily to support the transition of RTL83XX from PHY
driver to PCS driver. Thus, it may be removed later again or kept if
there is sufficient need for this.
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/20876
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
EWS2910P has two SFP slots of which only one was fully supported so far.
The issue so far was that both SFP slots share the same I2C SCL line but
neither the kernel nor any downstream driver was able to deal with this.
Thus, only one SFP slot was completely working (with detection etc.) but
the other one had to be enabled manually. Networking was functional in
both though.
Since acd7ecc9ed we have a driver which is able to deal with that. Thus,
we can fix the SFP support for this device.
Signed-off-by: Jonas Jelonek <jelonek.jonas@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/20687
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Unneeded semicolon
WARNING comparing pointer to 0
WARNING: NULL check before some freeing functions is not needed.
WARNING: casting value returned by memory allocation function to (u32 *)
ERROR: allocation function on line 378 returns NULL not ERR_PTR on failure
Signed-off-by: Rosen Penev <rosenp@gmail.com>
Link: https://github.com/openwrt/openwrt/pull/19932
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Tests with ingress shaping and enabled flow control showed really high
packet loss. It seems like the MAC pause frames are not created correctly
when both burst high off is set to the same value as burst high on.
By default, RTL930x has set the burst high values to:
* on: 64K
* off: 32K
Using the same 1:2 ratio seems to solve the high packet loss rate during
UDP tests.
Fixes: 2e74eb6d93 ("realtek: dsa: rtl93xx: Support per port throttling")
Signed-off-by: Sven Eckelmann <se@simonwunderlich.de>
Link: https://github.com/openwrt/openwrt/pull/21011
Signed-off-by: Robert Marko <robimarko@gmail.com>
The parameters must be aligned based on the last opened parenthesis
(+1). If this not a multiple of the tab size (8) then the rest
alignment must be done using spaces.
Signed-off-by: Sven Eckelmann <sven@narfation.org>
Link: https://github.com/openwrt/openwrt/pull/20906
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
It is written "multi" and not "mutli"
Signed-off-by: Sven Eckelmann <sven@narfation.org>
Link: https://github.com/openwrt/openwrt/pull/20906
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
It is written "initialization" and not "intialization"
Signed-off-by: Sven Eckelmann <sven@narfation.org>
Link: https://github.com/openwrt/openwrt/pull/20906
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
The Linux kernel coding style recommends not to add a space after
casts.
Signed-off-by: Sven Eckelmann <sven@narfation.org>
Link: https://github.com/openwrt/openwrt/pull/20906
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
It is recommended in the Linux kernel coding style not to add multiple
newlines after another.
Signed-off-by: Sven Eckelmann <sven@narfation.org>
Link: https://github.com/openwrt/openwrt/pull/20906
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
Instead of manually writing shift operations, it is preferred to
use BIT(b) or GENMASK(e, s).
Signed-off-by: Sven Eckelmann <sven@narfation.org>
Link: https://github.com/openwrt/openwrt/pull/20906
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
uint(8|16|32|64)_t and int(8|16|32|64)_t types should not be used in
kernel code. The shorter s(8|16|32|64) and u(8|16|32|64) or the
endianness specific versions (le*, be*) must be used instead.
Signed-off-by: Sven Eckelmann <sven@narfation.org>
Link: https://github.com/openwrt/openwrt/pull/20906
Signed-off-by: Hauke Mehrtens <hauke@hauke-m.de>
