On 3 Oct 2025, at 10:32, Ratheesh Kannoth wrote:

>> From: Eelco Chaudron <[email protected]>
>> Cc: [email protected]
>> Subject: [EXTERNAL] Re: ovs CT
>
>
>> I guess it’s always good to look at the unit tests there are tons of good 
>> examples.
>> For example:
>
>> https://urldefense.proofpoint.com/v2/url?u=https-3A__github.com_openvswitch_ovs_blob_2c4596732bff2dc353341873e3e1403814726eda_tests_system-2Dtraffic.at->
>>  
>> 23L6122&d=DwIFaQ&c=nKjWec2b6R0mOyPaz7xtfQ&r=aekcsyBCH00_LewrEDcQBzsRw8KCpUR0vZb_auTHk4M&m=qZ7MmlO8C9PuSmpV4kJSJdDRAVyqro9Wm-Oh3OvMfxoP2uLExXmjvzZCuQesA1P1&s=z_jHxEc3NES2wI5XbPhqeryfW9VKXGIfigfR3fdOh8c&e=
>> //Eelco
>
> Thanks.  OVS logs were showing the issue.  CONNTRAC ZONE should be enabled in 
> kernel.
>
> One more related question. If I use below commad,  ARP packets will pushes a 
> offload rule pushed for ARP packets.
>  But there is no rule push for TCP packets.  Will ovs  userspace APPwill push 
> a offload rule when connection is committed or established ?

I’ve not been looking how conntrack was implemented using TC, but here is nice 
presentation I googled;

https://www.youtube.com/watch?v=sst35r4cRsI

> ovs-ofctl add-flow br0 "table=0,priority=10,arp,action=normal"
> ovs-ofctl add-flow br0 
> "table=0,priority=100,tcp,ct_state=-trk,action=ct(table=1)"
> ovs-ofctl add-flow br0 
> "table=1,in_port=2,tcp,ct_state=+trk+new,action=ct(commit),1"
> ovs-ofctl add-flow br0 "table=1,in_port=1,tcp,ct_state=+trk+est,action=2"
> ovs-ofctl add-flow br0 "table=1,in_port=2,tcp,ct_state=+trk+est,action=1"
> ovs-vsctl set Open_vSwitch . other_config:tc-policy=skip_sw
> ovs-vsctl set Open_vSwitch . other_config:hw-offload=true
>
> -Ratheesh

_______________________________________________
discuss mailing list
[email protected]
https://mail.openvswitch.org/mailman/listinfo/ovs-discuss

Reply via email to