[Bug 272944] Vnet performance issues

https://bugs.freebsd.org/bugzilla/show_bug.cgi?id=272944

--- Comment #9 from Zhenlei Huang <[email protected]> ---
(In reply to Mark Linimon from comment #7)
> ^Triage: is this still a problem on supported versions of FreeBSD?

The data path of if_epair(4) does not change much, so I believe the issue still
persists.
I found that on CURRENT while I was working on lagg(4) race issues.

@Mark

I think `options RSS` shall apply to real network NICs, but not to if_epair(4).
A per-cpu taskqueue should fulfill and scale well on machines with many cores.

-- 
You are receiving this mail because:
You are the assignee for the bug.