Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Send ICMP too bigs directly #2223

Merged
merged 10 commits into from
Mar 2, 2020
Merged

Conversation

fasaxc
Copy link
Member

@fasaxc fasaxc commented Mar 2, 2020

Description

  • Send ICMP too big messages back out the same interface for workloads.
  • Allow GSO packets through.

Includes #2222 plus markups.

Todos

  • Unit tests (full coverage)
  • Integration tests (delete as appropriate) In plan/Not needed/Done
  • Documentation
  • Backport
  • Release note

Release Note

None required

tomastigera and others added 10 commits February 27, 2020 15:57
no need to check mtu if we do dsr, so check dsr first
On the return path into a nodeport tunnel, we can rely on the fact that
we encaping the return traffic in the host namespace, large packets are
GSO (because of veth) and the host iface will segment the encaped
packets correctly based on the host iface MTU.

base csum_offset on iphdr offset and ihl, it makes the offset correct
with IP options.
Althoug may not be 100% precise, gso_segs > 1 should be a good sign
that we are handling GSO packet and thus we do not need to care about
the mtu in tcp case.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants