Some sysctls currently are completely useless, as they only exist if
specific kernel configurations are enabled, which we have not.
To hide the error message and prevent them from interfering
unintentionally, if new kernel configurations are activated in the
future, they are removed.
Fixes: #42 (gitea)
Signed-off-by: Fabian Bläse <fabian@blaese.de>
Reviewed-by: Christian Dresel <freifunk@dresel.systems>
The net.ipv4.netfilter.ip* sysctls have been moved to
net.netfilter.nf* a long time ago, so they have been useless in our
firmware for quite a while.
It probably originally has been added because it was included in the
OpenWrt defaults and in earlier versions of our firmware the OpenWrt
defaults file got overwritten by our own one.
Because there does not seem to be any obvious reason to keep them (they
have been added without a comment in the commit or file) and they have
been inactive ever since they were moved in the kernel, they are removed
completely instead of using the correct path.
Fixes: #42 (gitea)
Signed-off-by: Fabian Bläse <fabian@blaese.de>
Reviewed-by: Christian Dresel <freifunk@dresel.systems>
Bump main repo and packages. (No changes for routing.)
Refresh patches (no diff returned).
This is a small release containing mostly kernel and package updates
and security fixes.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Reviewed-by: Fabian Bläse <fabian@blaese.de>
vxlan support in 19.07.5 is very limited.
This set of patches adds
- more flexible source ip selection
- control over most options
- multiple remote endpoint configuration
List of patches backported:
- 5222aadbf3 vxlan: remove mandatory peeraddr
- 65e9de3c33 vxlan: add capability for multiple fdb entries
- 036221ce5a vxlan: add extra config options
- ad3044c424 vxlan: fix rsc config option
- 3f5619f259 vxlan: allow for dynamic source ip selection (FS#3426)
- a3c033e2af netifd: vxlan: handle srcport range
- 226566b967 netifd: vxlan: refactor mapping of boolean attrs
- 11223f5550 netifd: vxlan: add most missing boolean options
- 55a7b6b7f2 netifd: vxlan: add aging and maxaddress options
Signed-off-by: Johannes Kimmel <fff@bareminimum.eu>
Reviewed-by: Robert Langhammer <rlanghammer@web.de>
Tested-by: Robert Langhammer <rlanghammer@web.de>
Reviewed-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
[refresh patches and remove some bloat]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
The package 'realpath' isn't available anymore on Debian 10, it is
part of coreutils now.
Reported-by: Felix Luber <Felix.Luber@servercreator.de>
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
The only part that is changed regularly inside buildscript are
the OpenWrt and package revisions and the selected packages.
Move them up and put them into dedicated variables so it is more
obvious what to change and easier to do so.
While at it, remove outdated COMPAT_VERSION comment from Gluon
package; we don't pull it from there anyway. Update comment on
feed definition syntax as well.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Reviewed-by: Robert Langhammer <rlanghammer@web.de>
The cake scheduler is a popular fair queuing scheduler, which is also
capable of shaping traffic. Due to its sensible defaults it is very
easy to set up.
When tunnel traffic exceeds the capability of the transport connection,
firmware users might want to shape traffic, so meaningful queueing can be done
before tunnel packets are dropped. As this is typically combined with a fair
scheduler, cake provides a simple yet very powerful solution for both problems.
Therefore the cake kernel module is now included in the layer3 variant.
Signed-off-by: Fabian Bläse <fabian@blaese.de>
Reviewed-by: Christian Dresel <freifunk@dresel.systems>
When neither peer_ip nor ipaddr are set, no ipv4 address for peering
interfaces is available. Therefore, no IPv4 routes can be advertised.
Other than that, a configuration like this is perfectly valid and
configuration is already continued. Therefore, the "FATAL" message might
be misleading so it is replaced with "WARNING" and clarified slightly.
Signed-off-by: Fabian Bläse <fabian@blaese.de>
Reviewed-by: Johannes Kimmel <fff@bareminimum.eu>
[bump PKG_RELEASE]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Grepping `/proc/cpuinfo` does not yield the correct number of available
cpus when running in a docker container or setting the number of
available cpus with taskset.
```
$ taskset 1 grep -c processor /proc/cpuinfo
8
$ taskset 1 nproc
1
```
This will prevent using too many build jobs on environments where the
number of available cpus is reduced.
`nproc` is part of `coreutils`.
Signed-off-by: Johannes Kimmel <fff@bareminimum.eu>
Reviewed-by: Fabian Bläse <fabian@blaese.de>
Reviewed-by: Robert Langhammer <rlanghammer@web.de>
On call -t we write the pid on /tmp/configure-layer3-pid.
If the script exits from user we use trap to run the new function keep_changes()
If the connections to the router lost, the user can run configure-layer3 -k after
reconnect to keep changes manually
Signed-off-by: Christian Dresel <freifunk@dresel.systems>
Reviewed-by: Robert Langhammer <rlanghammer@web.de>
[wrap and rephrase exit comment, bump PKG_RELEASE]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
This adds a script to init.d to automatically call
the appropriate configure-layer3 commands after an upgrade,
so the configuration of the device is restored.
The changes are applied if configure-layer3 is successful.
Otherwise, they are reverted. Due to this, no additional checks for
the configuration are necessary: The configure-layer3 script
will fail if the config version is wrong or no configuration exists at all.
After executing the script destroys itself. With START=99,
the execution happens _after_ uci-defaults and configurenetwork,
and no interference is expected.
Signed-off-by: Fabian Bläse <fabian@blaese.de>
[convert to init.d, extend commit message, rebase]
Signed-off-by: Christian Dresel <freifunk@dresel.systems>
Reviewed-by: Robert Langhammer <rlanghammer@web.de>
Reviewed-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Creating a directory via a .keep subfile is not really nice.
Use the OpenWrt mechanism for this instead.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Reviewed-by: Robert Langhammer <rlanghammer@web.de>
Acked-by: Christian Dresel <freifunk@dresel.systems>
The former fff-web package is essentially for serving a user
interface (UI). Therefore, include the UI in the name to distinguish
it from other web packages.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Reviewed-by: Robert Langhammer <rlanghammer@web.de>
Acked-by: Christian Dresel <freifunk@dresel.systems>
There might be scenarios where a user only needs the hood file,
but no WebUI (e.g. for 4M devices and node firmware), or only
the WebUI, but no hood file (e.g. layer3 firmware).
This separates the HTTP server (section) for the hood file into a
separate package fff-web-hood. The new package is then only added
to the node firmware.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Reviewed-by: Robert Langhammer <rlanghammer@web.de>
Acked-by: Christian Dresel <freifunk@dresel.systems>
Build/Prepare and Build/Configure are not required for packages
which only contain local files and do not need any compilation.
Remove them.
Note that Build/Compile needs to be present and empty to overwrite
the defaults, though.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Reviewed-by: Robert Langhammer <rlanghammer@web.de>
The name br-mesh is actually quite misleading, since the bridge
actually includes the "client" interfaces. In order to make this
obvious, and to prevent confusion with the properly named wXmesh
interfaces, rename them to br-client.
Note that br-mesh is also particularly disturbing for the layer 3
firmware without batman-adv.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Acked-by: Fabian Bläse <fabian@blaese.de>
Acked-by: Christian Dresel <freifunk@dresel.systems>
Reviewed-by: Robert Langhammer <rlanghammer@web.de>
The term "gateway" is ambiguous, and we are using "layer 3" for this
flavor now. Reflect that in the package name as well.
This is cosmetic otherwise.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Reviewed-by: Christian Dresel <freifunk@dresel.systems>
configuregateway and it's gateway.d files represent a specific
functionality that other packages depend on. Thus, it is put into
a package of its own so dependencies can be expressed more properly.
While at, use the chance to get rid of the ambiguous term "gateway"
and rename the script to configure-layer3 and the folders to layer3.d.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Reviewed-by: Christian Dresel <freifunk@dresel.systems>
With this patch the router answers in traceroute over the fff table with
Freifunk IPs and not with the WAN IP. All other connections use the fff
table too.
We already have the rules with 5000 and 5001 so that wireguard does not use
the fff table anyway and connect the VPN only via the main table.
Signed-off-by: Christian Dresel <freifunk@dresel.systems>
Acked-by: Fabian Bläse <fabian@blaese.de>
Reviewed-by: Robert Langhammer <rlanghammer@web.de>
Some packages contain a whitespace at the start of TITLE, some
don't. This is completely irrelevant since Make strips leading
and trailing spaces anyway. Nevertheless, make it consistent.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Reviewed-by: Robert Langhammer <rlanghammer@web.de>
To add new menu items in other packages we need a modular construction:
- Simpel Babelweb
- Layer 3 configuration
and so on
Every menu item needs a file in /www/menu/ssl/ with the content
link,name
- link is the name of the html file without extension (.html)
- name is the text of the link in the webui
While at it, rename "Password" menu item to German "Passwort".
Signed-off-by: Christian Dresel <freifunk@dresel.systems>
[bump PKG_RELEASE, use /www/menu, use script_file for link,
improve use of class_active, commit message adjustments,
improve variable names, keep HTML umlaut, keep Logout]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Certain files are only included, but are not meant to be accessed
via HTTP directly. Move those to a dedicated directory that is not
served via HTTP.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Reviewed-by: Christian Dresel <freifunk@dresel.systems>
Add support for the TP-Link Archer C50 v4, a low-cost mt7628-based
dual-band router.
Label MAC address is on ethernet and 2.4 GHz WiFi.
The v4 uses the same hardware as the v3 variant, but v4 includes
the newer split uboot.
ATTENTION:
Initial flashing of this device requires additional steps:
As all installation methods require an U-Boot to be integrated into the
image (and we do not ship one with the image) we are not able to create
an image in our build-process.
Download a TP-Link image from their Website and a FFF/OpenWRT sysupgrade
image for the device and build yourself a factory image like following:
TP-Link image: tpl.bin
OpenWRT sysupgrade image: owrt.bin
> dd if=tpl.bin of=boot.bin bs=131584 count=1
> cat owrt.bin >> boot.bin
This image can be used for Web-UI and recovery, but not TFTP.
Additional instruction can be found in the OpenWrt commit
01dcd574a248 ("ramips: add support for Archer C50 v4")
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Reviewed-by: Fabian Bläse <fabian@blaese.de>
Fixes low signal issue for 2.4 GHz for the TP-Link Archer C50 v4.
The first two bytes in the eeprom are the chip id. The working
devices have 0x7628 there, whereas the non-working devices have
0x7600 there. This chip id gets checked by the function
mt7603_check_eeprom() which leads the driver to ignore the
contents of the eeprom partition and load default values from otp.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Reviewed-by: Fabian Bläse <fabian@blaese.de>
This firewall was introduced as a countermeasure for very slow routers
directly connected to the internet without any firewall.
Our routers have got quite a bit faster since then. Also, a setup like
this is highly uncommon, especially for slower routers.
Therefore this firewall rule is removed.
Fixes: #138
Signed-off-by: Fabian Bläse <fabian@blaese.de>
Reviewed-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Reviewed-by: Robert Langhammer <rlanghammer@web.de>
[bump PKG_RELEASE]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
When using NATs and tunnels at the same time, the correct
source address has to be used so the ICMP errors is sent
through the NAT. This is necessary so the NAT can modify
the ICMP payload so it is correctly identified by the
destination host, which is required for PMTUD
Fixes: #142
Signed-off-by: Fabian Bläse <fabian@blaese.de>
Reviewed-by: Christian Dresel <fff@chrisi01.de>
Reviewed-by: Robert Langhammer <rlanghammer@web.de>
[bump PKG_RELEASE]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
For historic reasons, the wan interface is set to eth1 as default
value. When updating the config for one-port devices in
configurenetwork, the same value is put there again if the mode
is switched to something != WAN, instead of just removing the
value.
While ifname actually is a mandatory value, this has been handled
inconsistently in the past, where ethmesh ifname was deleted, and
wan ifname was just changed back to eth1, when assigning the
actually relevant eth0 to a different task.
This concept was set up with a one-port device in mind, i.e. a
device where there is no eth1. However, this very setup routine
got applied to the Nanostation M as well (which is treated as
as one-port), where we suddenly have two interfaces and the eth1
exists.
So, while the user assumes it's unconfigured, the second port
actually becomes set up as WAN if the first one is != WAN.
If connected to a second device with CLIENT (=default) to provide
PoE there, this will create a loop.
So, finally, in order to somehow fix this mess, this patch just
changes the hardcoded "eth1" to "eth2". While this is no proper
fix, it perpetuates the original idea of keeping wan set to
something, but nothing which actually exists. However, there are
no sideeffects and we keep this minimal-invasive.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Reviewed-by: Fabian Bläse <fabian@blaese.de>
This seems to be a copy/paste error, what we want here is to
delete the entry from ethmesh, as we use WAN.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Reviewed-by: Fabian Bläse <fabian@blaese.de>
Users might want to manually set up proper scheduling or qos using tc.
bmon can be used to monitor current throughput and packet rates.
Both are now included in layer3 variant.
Signed-off-by: Fabian Bläse <fabian@blaese.de>
Reviewed-by: Christian Dresel <fff@chrisi01.de>
Reviewed-by: Robert Langhammer <rlanghammer@web.de>
Flashing instructions:
The factory image needs to be uploaded via the OEM firmware GUI.
Notes:
The device is implemented as two-port in OpenWrt, i.e. it has
eth0/eth1 interfaces without switch setup. As our firmware currently
does not support that, this uses a switch setup with one port for
LAN.
Signed-off-by: Fabian Eppig <fabian@eppig.de>
[add commit message, apply alphetic sorting in bsp, remove config
changes apart from adding CONFIG_TARGET_DEVICE]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
What was intended as grouping of logic operators actually invoked
a subshell. Remove the subshell by using a better choice of operators.
Found by shellcheck.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Reviewed-by: Fabian Bläse <fabian@blaese.de>
This removes two useless cat as found by shellcheck:
sumnew=$(cat "$hoodfiletmp" 2>/dev/null | sha256sum | cut -f1 -d " ")
^-- SC2002: Useless cat. Consider 'cmd < file | ..' or 'cmd file | ..' instead.
While not functionally relevant in our case, note that
cat <non-existant-file> | sha256sum
actually returns a hash code, while
sha256sum <non-existant-file>
does not return anything on stdout.
Since we check the existance of $hoodfiletmp before calling sha256sum,
though, we always have a hash value for at least one file, so two empty
checksum won't happen at that point.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Reviewed-by: Fabian Bläse <fabian@blaese.de>
Until now, IPv6 connectivity was only ensured by some custom sysctls.
OpenWrt has a proper way of enabling IPv6 client (SLAAC, as well as DHCPv6)
for an interface. Switching to OpenWrt's way of configuring client addresses
for an interface might also make configuration more reliable, as the appropriate
sysctls are now set by netifd. Especially OnePort and TwoPort devices will
benefit from this change, as IPv6 auto configuration does not have to be manually
enabled and disabled for a physical interface, but rather is set as an option for
our logical wan interface.
At the same time this change enables DHCPv6 client support for WAN.
Signed-off-by: Fabian Bläse <fabian@blaese.de>
Reviewed-by: Robert Langhammer <rlanghammer@web.de>
Reviewed-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
[bump PKG_RELEASE]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
The hidden AP creation is moved to the end of
configurehood, so it is executed right after hoodfile
changes are processed.
When keeping the long sleep before trying to gather hoodfiles
via wireless or ethernet, this should decrease the delay after
hood changes to a minimum, as mesh nodes don't have to wait
until configurehood on VPN nodes is executed a second time.
Because hoodfiles gathered via wireless or ethernet are not
copied to hoodfilewww (which is used to deliver the active
hoodfile via ethernet or wireless), only authoritative hoodfiles
(keyxchange and gateway) trigger the creation of a hidden AP.
Signed-off-by: Fabian Bläse <fabian@blaese.de>
Reviewed-by: Robert Langhammer <rlanghammer@web.de>
Acked-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
[bump PKG_RELEASE]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Instead of seperately checking for various conditions,
which don't actually guarantee that the hoodfile can be
fetched in a certain way (e.g. internet is available but
keyxchange is down), the already built in return value of
the hoodfile gathering functions is utilized.
This change slightly changes the behaviour of nodes in
certian edge cases:
- If no hoodfile could be fetched from keyxchange, the
next delivery method (getGatewayHoodfile) is used
- If the gateway is unable to deliver a hoodfile, nodes
now behave like the gateway is unreachable, instead of
continuing to use old hoodfiles
These behaviour changes should be an improvement over the
previous behaviour:
- VPN nodes don't disconnect and break the whole network
if the keyxchange is unreachable, but instead try to
fetch the hoodfile from the gateway
- Instead of checking for batman gateway announcements,
which are completely unrelated to hoodfile delivery using
fe80::1, the actual status of the hoodfile download is
utilized. This has two effects:
- hoodfile delivery using fe80::1 works even if batmans
gateway selection isn't used at all
- if the batman gateway selection is active, but fe80::1
hoodfile delivery is broken in the hood, the nodes disconnect
from the hood and try to gather their hoodfile from nerby
nodes. Previously they continued to use the old hoodfile.
This should make misconfigured gateways more apparent.
Signed-off-by: Fabian Bläse <fabian@blaese.de>
Reviewed-by: Robert Langhammer <rlanghammer@web.de>
Reviewed-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
This adds new strings to the support list for the TP-Link CPE210 v3
that are supposed to work with the existing setup.
Without it, the factory image won't be accepted by the vendor UI on
these newer revisions.
This has been merged upstream in commit 4a2380a1e778.
Backport to 19.07 is planned already, so the patch can be removed
again when we move to the next point release.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Acked-by: Fabian Bläse <fabian@blaese.de>
This adds a few improvements and fixes for vxlan support.
The following two patches are already backported to openwrt-1907 and
can be dropped after bumping to the next point release:
0011-vxlan-fix-udp-checksum-control.patch
0012-vxlan-bump-and-change-to-PKG_RELEASE.patch
The other two patches won't be backported and have to be kept until
we move up to 20.xx:
0013-vxlan-remove-mandatory-peeraddr.patch
0014-vxlan-add-capability-for-multiple-fdb-entries.patch
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Reviewed-by: Johannes Kimmel <fff@bareminimum.eu>
Tested-by: Johannes Kimmel <fff@bareminimum.eu>
The factory image can either be flashed via the vendor WebUI or
the bootloader using nmrpflash.
Signed-off-by: Fabian Bläse <fabian@blaese.de>
Reviewed-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
mtr can be a very helpful tool when debugging unstable
networks. The tool is able to list packet loss to all
routers to a destination in a nice console-based interface.
Signed-off-by: Fabian Bläse <fabian@blaese.de>
Reviewed-by: Tim Niemeyer <tim@tn-x.org>
Reviewed-by: Robert Langhammer <rlanghammer@web.de>
For the layer-3 firmware, we currently do not support a hood file
for automatically creating a mesh. However, the link for the
hood file is still created in fff-web.
Move this setup to fff-hoods, which is specific to the node firmware.
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Reviewed-by: Fabian Bläse <fabian@blaese.de>
OpenWrt only builds and installs a packet for each architecture.
If a package is already fully built and installed for a architecture,
it is not rebuilt.
Because we have two different BSPs building two different targets (BOARDs)
using the same architecture (ath79 and ar71xx, mips), the fff-network package
is not reinstalled when switching between those.
However, we have defined an install step, which copies the necessary network
files seperated by board. But because the package is not rebuilt when switching
targets, the wrong network files might be present in the package.
To resolve this issue, the network files are now seperated by ARCH instead of
the target (i.e. the BOARD variable).
Signed-off-by: Fabian Bläse <fabian@blaese.de>
Acked-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Acked-by: Robert Langhammer <rlanghammer@web.de>
[bump PKG_RELEASE]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
This patch moves device support for ar71xx devices that are supported
in ath79. Building them with ath79 will be the new default.
The only devices remaining in ar71xx will be the following, as backporting
them to openwrt-19.07 is too complicated:
- cpe210-v1
- cpe510-v1
Accordingly, no tiny devices are left in ar71xx and we can drop the relevant
patches, and build ar71xx as generic again.
For the tl-wr741nd-v2, in ath79 the tplink_tl-wr741-v1 image is used.
The move from ar71xx to ath79 requires some adjustments on the way:
- The board names and image names on ath79 contain the vendor name,
where the former have it separated by a comma (tplink,cpe210-v2)
and the latter use an underscore (tplink_cpe210-v2). It is
safe to assume that this is the only difference between board and
image names.
Consequently, the ath79 devices will use their full board name also
in our firmware. A lot of renames in fff-boardname can be dropped.
The rename for fff-sysupgrade is already present in fff-upgrade.sh
While at it, fix that for the WDR4900 v1 as well.
- Due to a different switch driver, the startup of ethernet devices
is altered, which leads to eth0 and eth1 being swapped for some ath79
devices compared to ar71xx. This has been adjusted for SWITCHDEV/WANDEV
and MAC address setup.
- Since we have direct support for the AC Mesh now, use the proper
name instead of the AC Lite image. For Ubiquiti, different device
variants are now available as separate images.
- Remove left-over power-m-xw entry in cpuport file
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
Reviewed-by: Fabian Bläse <fabian@blaese.de>
Tested-by: Fabian Bläse <fabian@blaese.de>
This patch is already included in the openwrt-19.07 branch,
so our backport can be removed when upgrading to or past
OpenWrt 19.07.4.
Signed-off-by: Fabian Bläse <fabian@blaese.de>
Reviewed-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
[adjust commit title]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>
The babel interface type 'tunnel' has some disadvantageous properties for
our network.
First, babel tries to evaluate the tunnel performance using the rtt. However,
this makes the network quite unstable, as rtt might fluctuate a lot, especially
on less reliable connections (e.g. LTE). Instead of fully falling back to an alternate
route, this rtt evaluation leads to a lot of flapping routes. Additionally, rtt
evaluation changes the metric of routes quite often, which leads to many unnessessary
babel messages in our network.
Also, babeld disables split-horizon processing on 'tunnel' interfaces per default.
However, split-horizon processing can be done in our point-to-point tunnel setup without
any issues and has the advantage of significantly reduced babel messages on a link with
many uplink routes.
Therefore, wireguard babel peers now use the interface type 'wired'.
Signed-off-by: Fabian Bläse <fabian@blaese.de>
Reviewed-by: Johannes Kimmel <fff@bareminimum.eu>
Reviewed-by: Robert Langhammer <rlanghammer@web.de>
[bump PKG_RELEASE, adjust commit title prefix]
Signed-off-by: Adrian Schmutzler <freifunk@adrianschmutzler.de>