Bad udp/tcp cksum logged in all queries

Unbound
1

hi

could i please have some feedback about what possible reasons would
cause this ?

for example:

1.2.3.4.47507 > tinnie.arin.net.domain: Flags [S], cksum 0x0f97
(incorrect -> 0xa32b), seq 2709415216, win 29200, options [mss
1460,sackOK,TS val 60120911 ecr 0,nop,wscale 7], length 0

1.2.3.4.51245 > u.arin.net.domain: [bad udp cksum 0xec18 -> 0xaa99!]
12116% [1au] DS? 31.192.iN-aDdr.aRpA. ar: . OPT UDPsize=4096 OK (48)
18:21:34.244863 IP (tos 0x0, ttl 52, id 11573, offset 0, flags [none],
proto UDP (17), length 531)

i never noticed this but i don't have any issues at all receiving
replies with udp/tcp checksums ok

the question remains, although these are logged, is it a problem ?

unbound 1.4.22 from debian jessie amd-64 repos

any other relevant settings required & i can revert

2

under linux/unix platform, youcan turn off the checksum for packet：
the command is /sbin/ethtool -K eth0 rx off tx off.

------------------ Original ------------------

3

Hi Shmick,

could i please have some feedback about what possible reasons
would cause this ? (...) i never noticed this but i don't have any
issues at all receiving replies with udp/tcp checksums ok (...) the
question remains, although these are logged, is it a problem ?

The wireshark FAQ has a nice section[1] about the effects of bypassing
your network stack with capture tools. I suspect that is very relevant.

Regards,
Yuri

[1] https://www.wireshark.org/faq.html#q11.1

4

thanks,
sorry for that - off-topic in the end
i think i'll leave as is with offload to hardware NIC

Yuri Schaeffer wrote: