-
Notifications
You must be signed in to change notification settings - Fork 2.7k
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
Support L2-less devices with fast redirect #23935
Support L2-less devices with fast redirect #23935
Conversation
This comment was marked as outdated.
This comment was marked as outdated.
8e388eb
to
708f18c
Compare
This comment was marked as resolved.
This comment was marked as resolved.
fc65f1f
to
4e377c3
Compare
c449da4
to
5291f99
Compare
This comment was marked as outdated.
This comment was marked as outdated.
5291f99
to
886455c
Compare
This comment was marked as resolved.
This comment was marked as resolved.
886455c
to
71097fc
Compare
This comment was marked as resolved.
This comment was marked as resolved.
This comment was marked as outdated.
This comment was marked as outdated.
71097fc
to
2ccd928
Compare
3d37e2e
to
820e50d
Compare
Now it results with even more failures.... |
Originally the host routing mode would be set lagacy if wireguard is enabled or any l2-less device exists. This commit lifts the limitation. Fixes: cilium#15075 Signed-off-by: Zhichuan Liang <gray.liang@isovalent.com>
20b053f
to
1bcd981
Compare
For this failure, I can make it pass locally with 2 changes:
So I presume the e2e test failures I hit before were both flakes, although the policy one is maybe not recorded before. I can fix these "delete CRDs without waiting" race condition in e2e test in a separate PR later. |
/test |
🎉 would be great to have a follow-on issue for this aspect! |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Just one tiny fix-up needed. CI is all green already ✔️
This commit adds l2 header before handle_ipv4 calls ipv4_local_delivery, in order to enable fast-redirect from l2-less dev (e.g. cilium_wg0) to internal veth. Fixes: cilium#15075 Signed-off-by: Zhichuan Liang <gray.liang@isovalent.com>
This commits add unittest case for L3 skb fast redirecting to L2 device. Signed-off-by: Zhichuan Liang <gray.liang@isovalent.com>
The limitation for wireguard was originally recorded as not compatiable with host-routing, and it's no more the case once pr cilium#23935 is merged. Signed-off-by: Zhichuan Liang <gray.liang@isovalent.com>
1bcd981
to
09e8dc4
Compare
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Looks great, thank you @jschwinger233 ! 🚀
/test Job 'Cilium-PR-K8s-1.16-kernel-4.19' failed: Click to show.Test Name
Failure Output
If it is a flake and a GitHub issue doesn't already exist to track it, comment |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Nice work!
* Therefore, we created a stud, mock_hanle_policy, to simply check if the | ||
* our skb reaches there. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Not worth a respin, but if you need to update your PR for another reason:
* Therefore, we created a stud, mock_hanle_policy, to simply check if the | |
* our skb reaches there. | |
* Therefore, we created a stud, mock_hanle_policy, to simply check if | |
* our skb reaches there. |
There's another typo in the description of your first commit: lagacy
.
The failing update test is known flake #24379:
|
Marking this for v1.13 backporting as another PR relies on some these changes. |
This PR adds L2 header for skb on L2-less devices before redirecting to internal veth, and enables bpf-based host routing work with wireguard as well as L2-less interfaces.
Fixes: #15075
Signed-off-by: Zhichuan Liang gray.liang@isovalent.com