bad udp cksum messages in tcpdump for wg0
Hi
My DNS server is available via a wireguard interface. Because of many requests I did a tcpdump on the traffic and found messages like.
14:05:34.881307 IP (tos 0x0, ttl 63, id 33826, offset 0, flags [DF], proto UDP (17), length 54) 10.99.97.15.53628 > 10.99.97.17.domain: [bad udp cksum 0xd719 -> 0x6360!] 8446+ A? postgres. (26)
I could turn that messages off with turning off check validation on the interface
ethtool -K wg0 tx off rx off
Actual changes:
rx-checksumming: off
tx-checksumming: off
tx-checksum-ip-generic: off
The tcpdump traffic looks now like this
14:08:36.494987 IP (tos 0x0, ttl 63, id 61627, offset 0, flags [DF], proto UDP (17), length 54) 10.99.97.15.40185 > 10.99.97.17.domain: [udp sum ok] 1324+ A? postgres. (26)
I want to know if the behavior described above with the checksum errors is to be expected? Or is it necessary to turn those checks off on all the interfaces?
BR
ybaumy
_______________________________________________
WireGuard mailing list
[email protected]
https://lists.zx2c4.com/mailman/listinfo/wireguard
Hi,
Quoted 5 lines
ethtool -K wg0 tx off rx off
I want to know if the behavior described above with the checksum errors is to be expected? Or is it necessary to turn those checks off on all the interfaces?
Yeah. I need that too, on one of my routers. Apparently some checksum offloading hardware gets confused by wireguard.
Interestingly, on that machine fixing this problem requires turning off checksumming on the wg interface _or_ on the hardware interface, either way is sufficient.
Signature · 7 lines
-- Matthias Urlichs _______________________________________________ WireGuard mailing list [email protected] https://lists.zx2c4.com/mailman/listinfo/wireguard