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

Reset the route flag before moving the rule #472

Merged
merged 1 commit into from
Apr 8, 2020

Conversation

ahenan
Copy link

@ahenan ahenan commented Apr 6, 2020

No description provided.

@ahenan ahenan force-pushed the master branch 4 times, most recently from c9cdb32 to 2b11447 Compare April 6, 2020 00:49
@ahenan
Copy link
Author

ahenan commented Apr 6, 2020

Sometimes when the plugin is chained with other plugins e.g. sriov, the existing dynamically created routing rule will have the flags field set and that will cause the route replace to fail. Resetting the flag should sort out the mysterious "invalid argument" error reported from the kernel.

@squeed
Copy link
Member

squeed commented Apr 6, 2020

Paging @plwhite, the author of this plugin. Does this change make sense to you?

@plwhite
Copy link
Contributor

plwhite commented Apr 6, 2020

Can I just check I understand @ahenan ?

Suppose you have two devices, A and B, and you are going to move B's routes to the new table. All the routes in table B that have been dynamically changed have to be moved to the new table, but have to lose their dynamic flag and become static because as a user you cannot add dynamic routes. Presumably they can still be overridden by further dynamic changes. Is that right? And does that specifically happen with SR-IOV devices?

I'm comfortable with the change, but I think this is subtle enough that a short comment above the change saying why it is necessary would make it clearer for future users.

@ahenan
Copy link
Author

ahenan commented Apr 6, 2020

@plwhite I think we are on the same page, so you are right.

I cannot really say it happens only with sriov but since that is the case, I guess we may see it with some other configurations/setups.
Agreed, it is subtle enough to be worth a brief explanation, I have added few lines regarding this.

@ahenan ahenan force-pushed the master branch 4 times, most recently from 77f5ac6 to 7762809 Compare April 6, 2020 14:44
Signed-off-by: ahenan <ahenan00@gmail.com>
@plwhite
Copy link
Contributor

plwhite commented Apr 6, 2020

Thanks - I am comfortable with that change. It seems completely safe, and I agree is necessary in at least some cases.

@plwhite
Copy link
Contributor

plwhite commented Apr 6, 2020

I don't have permission to merge, so will wait for a maintainer to step in.

@ahenan
Copy link
Author

ahenan commented Apr 6, 2020

No worries, thanks.

@ahenan
Copy link
Author

ahenan commented Apr 8, 2020

@squeed any plans to close this PR?

@squeed
Copy link
Member

squeed commented Apr 8, 2020

lgtm

1 similar comment
@dcbw
Copy link
Member

dcbw commented Apr 8, 2020

lgtm

@squeed squeed merged commit ed16760 into containernetworking:master Apr 8, 2020
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

Successfully merging this pull request may close these issues.

4 participants