Thread (7 messages) 7 messages, 2 authors, 10d ago

Re: [PATCH net-next v2 1/2] amt: mark relay data as a UDP tunnel packet before sending it

flat view

From: Omar Ramadan <hidden>
Date: 2026-10-04 14:01:37
Also in: linux-kselftest, lkml

On Sat, Oct 3, 2026, Sashiko wrote:
Should this have a Fixes: tag?
No, on purpose. Eric asked for this to go to net-next, since the bug
needs a non-default setting:

  "Note that tx checksum offload is off by default on amt, so the bug
  would not trigger by default."
  "A better fix would be (for net-next) to call
  udp_tunnel_handle_offloads() in amt_send_multicast_data() like other
  UDP tunnels do"

https://lore.kernel.org/all/CAL4Wiiqw_7Mxa5W1EQgw2=KpJNNxG=cBLXsHvpio29GvD1UKZg@mail.gmail.com/ (local)
This is a pre-existing issue, but it is closely related to this patch.
I have not reproduced it; so far this comes from reading the code. It
would need rx-gro-list on the relay's ingress device, and tx
checksumming plus tx-scatter-gather-fraglist on the amt device (all
three off by default), and a GRO aggregate that passes ipmr's DF/MTU
check. It predates this patch, and the cover letter already lists it as
not touched and untested. I will send a separate fix for it once I have
reproduced it.
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help