[cgl_discussion] PoC concall minutes, 3/27
Rod Van Meter
Rod.VanMeter at nokia.com
Thu Mar 27 09:01:05 PST 2003
OSDL: Randy, Andy, Mark, John
Intel: Julie, Rusty
Nokia: Johannes, rdv
Rusty: we need roadmap/milestone for each project.
Porting of TAHI to Linux -- Intel engineer has contacted TAHI folks
about getting his Linux port incorporated. Port is finished, and
license is basically BSD.
OpenHPI -- license issue is taken care of, initial design discussions
now on OpenHPI mailing list (openhpi.sf.net). Intel/OR, IBM/NY,
Shanghai involved. Working on plugin mechanisms. Couple of weeks
until we have a roadmap. Have a validation suite engineer in China.
SAF requires each vendor to have their own validation suite? Some of
the proprietary vendors might be interested in an open test suite.
Mika made a roadmap for TIPC. Plan is code to be ready by end of 3Q.
By end of year, should be starting to push into mainline. Mika
working on low-level interface. 2.5 port is 7x slower than Ericsson's
2.4 in transaction latency; not sure yet why. Maybe a timer problem?
Milestones are on OSDL cluster SF site.
Rusty has a handful of low-level tests available on his web page.
Do we need a mailing list for TIPC? Email traffic/latency problems.
Mika's against taking TIPC to LKML in its current state.
v6: Venkat says:
IPv6 anycast functionality (almost 1300 lines of new code), submitted
IBM has been accepted by Dave Miller and will be part of 2.5.x
IPSec6 in 2.5.65 kernel has been tested for functionality
by TAHI tests and found several issues. We'll post the results soon
IGMPv3/IPv4 (SSM support for host) final patch (almost 1500 lines of
new code) has been submitted by IBM for review by Dave Miller. Hope
hat it will be checked in soon into 2.5.x tree
MLDv2/IPv6 development (similar to IGMPv3) is going smooth and
the final patch will be submitted by end of April
DHCPv6 ver 0.4 is scheduled to be released by end of this month.
DHCPv6 will be part of CGL compliant Service Pack from UL since
is a CGL v1.1 reqt. It will be based upon ver 0.3 of DHCPv6
Randy can probably fill-in on IPv6 MIBs status.
Randy says Venkat has asked him to broaden this into a MIB audit, more
or less. TCPv4 needs some trivial work. Still lots of v6 stats to be
added, some gated on some USAGI code which is now getting in.
Overall, slower than he'd like.
Nothing to report on HDI.
Nothing new on boot cycle detection. Will ask Ville to send some info
Frederic: what is next step with AEM? Sent FAQ to compare w/ select()
and epoll() last week.
Randy thinks it's a good idea, but too intrusive into the scheduler.
There may be other ways to do this. Dave Olien may have some
suggestions on how to do this, but he's on vacation next week. Will
have him here for meeting next week.
Might want to compare this to what SAF specifies.
What I Did With My Spring Vacation:
More information about the cgl_discussion