I followed the openwrt wiki for Asus TUF AX6000 (with some AI help) in order to use WED (wifi) and bridger for the dumb access point.
To check whether it works I ran iperf3 and crosschecked with cat /sys/kernel/debug/ppe0/bind
but the result was always nothing.
In dmesg it shows [ 11.852720] bridger[1438]: memfd_create() called without MFD_EXEC or MFD_NOEXEC_SEAL set
indicating that something is not loaded correctly (acc. to AI).
Here some additional info:
opkg list-installed | grep bpf
kmod-sched-bpf - 6.6.86-r1
libbpf1 - 1.5.0-r1
cat /sys/module/mt7915e/parameters/wed_enable
Y
brctl show
bridge name bridge id STP enabled interfaces
br-lan 7fff.a0ad9f48412c no phy1-ap0
lan4
lan2
lan5
lan3
lan1
phy0-ap0
lsmod | grep mt79
cfg80211 299008 4 mt7915e,mt76_connac_lib,mt76,mac80211
mac80211 577536 3 mt7915e,mt76_connac_lib,mt76
mt76 73728 2 mt7915e,mt76_connac_lib
mt76_connac_lib 45056 1 mt7915e
mt7915e 135168 0
lsmod | grep ppe and lsmod | grep bridger show nothing
cat /etc/openwrt_release
DISTRIB_ID='OpenWrt'
DISTRIB_RELEASE='24.10.1'
DISTRIB_REVISION='r28597-0425664679'
DISTRIB_TARGET='mediatek/filogic'
DISTRIB_ARCH='aarch64_cortex-a53'
DISTRIB_DESCRIPTION='OpenWrt 24.10.1 r28597-0425664679'
DISTRIB_TAINTS=''
Btw, the night before (20-Jun-2025, CET) I couldn't create an image, because something went wrong with the imagebuilder, indicating several corrupt packages, but this morning I could create an image including bridger without issues.
Key Questions
Is is worth to turn on WED and use bridger for a dumb access point ?
If yes, what should I do to make it run ?
Off-topic or linked to it ?: Why is my wifi 5 GHz so slow (up 560 MBit/s, down only 180 MBits/s, my other AP Netgear WAX206 with the identical config is much faster with 80 MHz), see config (dawn does not influence the speed btw):
If your client is apple too many 11R )ptions make its location service nuts.
1 yes
2 install bridger package on a standrd image
3 because you have 2 160MHz options, one is half dfs, half neighbours, other is all dfs. Try 80MHz instead, as per your own writing.
Please connect to your OpenWrt device using ssh and copy the output of the following commands and post it here using the "Preformatted text </> " button:
Remember to redact passwords, MAC addresses and any public IP addresses you may have:
These are my imagebuilder parameters for dumb access points I use, except the line with ethtool-full bridger, only for Asus TUF AX6000 (I know you didn't ask for, maybe it helps):
Can't do today, wife is watching TV ... will check late late later
In CH only channels 100-128 seem allowed, lower channels 36-48 don't work in openwrt (iwinfo says 6 GHz and NOHT, if I try channel 42 with HE160). See here:
Not related to the issue at hand, but you're going to end up locked out when you run a sysupgrade at some point in the future. The firewall service will become re-enabled when you sysupgrade, and input is set to REJECT for your defaults in there. That will lock you out entirely.
Change the input rule to ACCEPT. And you can leave the firewall service enabled.
Is it possible that WED and bridger are correctly running if I also get this:
dmesg | grep -i "wed"
[ 9.913090] mt798x-wmac 18000000.wifi: attaching wed device 0 version 2
[ 9.972192] platform 15010000.wed: MTK WED WO Firmware Version: DEV_000000, Build Time: 20221012175005
[ 9.981521] platform 15010000.wed: MTK WED WO Chip ID 00 Region 3
service bridger status
running
dmesg | grep -i "offload" -> I get nothing as output, also with iperf3 -c (is this ok?)
I tested locally now with 160 MHz (116 center channel) and 80 MHz (60 center channel) with iperf3 to a local PC (two 2.5 GbE switches in between, one Mikrotik, one Ubiquiti) from my Macbook Air 15 M4 ("brewed" iperf3).
Results (iper3 -c "ip-address" -P 16):
160 MHz (-R, download): 1.67 GBit/s
160 MHz (upload): 1.48 Gbit/s
80 MHz (-R, download): 969 Mbit/s
80 MHz (upload): 900 MBit/s
To be noted, at 160 MHz, a 2400 MBit/s link is established to the airbook
at 80 MHz, a 1200 MBit/s link is established.
Now the maybe irritating iperf3 tests to eg speedtest.iway.ch:
160 MHz (-R, download): 550 MBit/s
160 MHz (upload): 230 MBit/s
80 MHz (-R, download): 550 Mbit/s
80 MHz (upload): 480 MBit/s
(80 MHz, with speedtest.net - if I choose the right server, I get up to 800 MBit/s download and > 700 MBit/s upload)
"Analysis"
-> AI tells, that with 160 MHz there are more "hickups" with iperf3 over internet due to higher latencies compared to local network, which may explain the dramatic degradation of internet speed.
-> However, when I check interference (e.g. wifi analysis or wifi explorer), at 160 MHz there is almost no congestion, there is a 80 MHz band with -92 dBm overlapping with my 160 MHz from a neighbour. Not sure about radar signal, if this is shown in openwrt.
Fyi, I have 1 GBit/s symmetric internet, FTTH, I limited with SQM to 900'000 kbit/s download and 850'000 kbit/s upload for wan and lan (lan values are "flipped"), so a quick test on the PC (two switches to wan router, just fyi)
with SQM I get iperf3 download 851 MBit/s / upload 760 MBit/s
without SQM iper3 (above not tested with this settings!): download 945 MBit/s, upload 860 MBit/s
Do I really have an issue with WED / bridger or are these results expected ?
Maybe other users have similar observations, and it may help to understand (?).
