Thanks for this and your reply. I’m an old ‘newbie’!
Last time I created an OS from scratch was in 1975…
Had to take the media from London to Warsaw (only airline networks in those days), a 14" 7.25MB replaceable disk that just fitted into an overhead locker! With a letter from CoCom permitting me to export a forbidden material to a Warsaw Pact country… I just love GitHub!
Regards
can you please summarize your current state in the other thread so we can clean up this thread? the ONT discussion has nothing to do with WED/Beamforming
What else should I check? Should the regular hardware offloading be also enabled for idk the WAN interface ? Should I configure flowtables? Maybe I am missing a kernel module that is responsible for that?
Please note that I am not using OpenWRT but just a regular linux.
I though openwrt have backports so this setting should be also in mainline linux…
@dangowrt can you say since which kernel version wed should work and if the way is same as in openwrt? I guess it also needs some dts properties in r3 dts…
I have answered you because you have pointed me,but answer was for @ghostbuster91
In mainline Linux WED is fully supported since v6.2.
You will need to set the module paramter wed_enable=Y for the mt7915e module.
As offloading depends on classifying packets into flows you do always depend on nftables. In order to use WED in a dump/bridged AP which doesn’t actually do any routing/NAT what-so-ever we got a eBPF hack in OpenWrt called bridger. So if you want to use WED on an OS other than OpenWrt, you will need to manually create the needed nftables rules and if you want to use WED on non-routing APs you will also need to package and build bridger.
Are these the same as for hw-nat with flowtable and offload flag…just adding the wifi interfaces? If not is it somewhere documented?
table ip filter {
flowtable f {
hook ingress priority filter + 1
devices = { lan3, lan0, wan }
flags offload;
}
chain forward {
type filter hook forward priority filter; policy accept;
ip protocol { tcp, udp } flow add @f
}
Btw. Needs offload done before or after other rules are applied? E.g. forward-chain with drop policy (and/or last rule reject) allowing some specific flows
The packets follow the normal forwarding rules, until they are established. Then offload kicks in and forwarding rules are totally bypassed for the established connection.