On Jun 19, 2006, at 11:55 AM, Venkat Yekkirala wrote:



The key concept here is the flow. The flow attributes (derived from the
socket
in the locally generated case, derived from the packet in the forwarding
case, etc.)
determine the xfrm policy to use on the outbound. On the inbound,
it's again the flow attributes (derived from the packet) that need to
satisfy
a xfrm policy rule (the first matching one) completely

REGARDLESS OF WHETHER THE PACKET IS HEADED FOR A LOCAL SOCKET OR IS BEING
ROUTED OUT (FORWARDING CASE).

OK.

What you are saying makes sense. I will take a last look at the code tomorrow, and give OK (assuming optimism).

Regards,
Trent.
----------------------------------------------
Trent Jaeger, Associate Professor
Pennsylvania State University, CSE Dept
346A IST Bldg, University Park, PA 16802
Email: [EMAIL PROTECTED]
Ph: (814) 865-1042, Fax: (814) 865-3176




--
redhat-lspp mailing list
[email protected]
https://www.redhat.com/mailman/listinfo/redhat-lspp

Reply via email to