On Mon, Mar 07, 2005 at 12:55:52PM -0800, Matthew Dillon wrote:
> I'm still thinking the IF queueing API should be rewritten to accomodate
> the ethernet header issues directly, and then implement ALTQ within the
> API rather then require the IF code to know about it.
The ALTQ API has a slightly different semantic, but otherwise is not different
from the classic queue. The ethernet header is problematic just for the
upper layers. It does not effect the dequeue.
Joerg