IPSec item - RE: [cgl_discussion] POC 9/24 Cancelled

Venkata Jagana jagana at us.ibm.com
Thu Oct 2 12:22:13 PDT 2003

Note: Modified the subject

   >- IPSec -> This one we just recently added.  See note with Forrest and
   Mika's comments on what needs to be >done to the user space portion of

   >- Per input at the F2F:  Of the two implementations, FreeSWAN is known
   >be stable, but will not become part of the kernel for the foreseeable
   >future.  The implementation that is part of the kernel needs someone to

   >test to see how useful it is.  The tools aren't user-friendly or usable

   We have validated the IPSec v4 implementation currently in 2.6.x
   kernel around 2.5.70 or 2.5.71 time frame and found that the
   implementation is highly stable from both functional and
   stree testing standpoint.

   We have used TAHI IPv4 tests for functional validation and found that
   the results are really good. This validation is done for both transport
   and tunnel mode operations.

   Bottomline: IPsec v4 in 2.6.x kernels is as stable the other
   As far as the tools are concerned, they are in usable condition. We have
   some issues and fixed them. Ofcourse, they have been evolving since
   Not sure about user-friendliness, every one's views are different.


                      "Fleischer, Julie N"                                                                                                    
                      <julie.n.fleischer at intel.c        To:       "Shureih, Tariq" <tariq.shureih at intel.com>, <sdake at mvista.com>,             
                      om>                                <cgl_discussion at osdl.org>                                                            
                      Sent by:                          cc:                                                                                   
                      cgl_discussion-bounces at lis        Subject:  RE: [cgl_discussion] POC 9/24 Cancelled                                     
                      09/30/2003 02:37 PM                                                                                                     

> -----Original Message-----
> From: Shureih, Tariq
> Steven:
> I'd like to add the following to the next meeting's agenda:
> The definition of CGL 2.0 requirement fulfillment for openhpi.

Steve -
I can't make it to tomorrow's POC (or I may be able to make it, but late).
I had some questions/tracking items from the top-10 that I wanted to bring
to the POC, though:

(from http://www.osdl.org/docs/cgl_poc_top_10_projects.txt)
- Asynchronous Events -> We haven't discussed in a meeting since finding
that epoll+libevent may implement this requirement.  Do we think
epoll+libevent meets the full requirement (and this no longer needs to be
top-10)?  If not, which one(s) do we want to ask steering for resources
for:  AEM or epoll+libevent?

- PKI CA -> Is there additional analysis CGL needs to do to be able to get
distro input on this one?

- IPSec -> This one we just recently added.  See note with Forrest and
Mika's comments on what needs to be done to the user space portion of this.

For the others, there are just status questions:
- Persistent Device Naming: status on uSDE
- TIPC status
- message passing status

- Julie

**These views are not necessarily those of my employer.**

cgl_discussion mailing list
cgl_discussion at lists.osdl.org

More information about the cgl_discussion mailing list