eigrp has a coding pattern that will generate a crash

[donaldsharp]

Commit 3fa39a35ed -> identified in ospf6d a pattern where the daemon could crash. eigrp copied the exact same pattern and a event_execute can be issued where there is an existing write scheduled and the queue can be cleared. This needs to be addressed

[diivious]

I looked at the code and not sure it has the same issue. From what iI see, you main addition was to add an event_cancel when ospfd is terminated.

EIGRP already has this code in place though the call path: eigrpd_instance_destroy eigrp_finish_final EVENT_OFF eigrp_cancel

[mjstapp]

I think the problem was around eigrp_write(): it asserts if it doesn't find any work to do, and it's run both as an async, scheduled task and synchronously via event_execute(). I've got an open PR that will allow event_execute to cancel a pending task for the same callback, for example, which would be one way to address this sort of situation.

[diivious]

can you share a link to the PR please?

[mjstapp]

can you share a link to the PR please?

the PR was merged just yesterday, in fact, so the change is now in master

[github-actions[bot]]

This issue is stale because it has been open 180 days with no activity. Comment or remove the autoclose label in order to avoid having this issue closed.

[frrbot[bot]]

This issue will be automatically closed in the specified period unless there is further activity.