Minutes of Weekly Meeting, 2007-11-26
Brad Van Treuren
Meeting was called to order at 8:09am EST
Review of meeting minutes for 11/19/2007;
Approved as is (moved by Ian, second by Peter)
Changes to the minutes are:
- Correct spelling of Ian McIntosh's name
Discussion of open action items:
- Adam proposed we cover the following at the next meeting:
- Establish consensus on goals and constraints
- What are we trying to achieve?
- What restrictions are we faced with?
- still ongoing
- Establish whether TRST needs to be addressed as requirements in the ATCA specification if it is not going to be managed globally (All)
- still to be resolved
- Provide feedback of more use cases not yet identified to Brad (All)
- awaiting feedback ...
- Review tables (Goals vs. use case matrix) on slides 38-41; (All)
- awaiting feedback ...
- Ian to look into general observer solution for visibility of our work to outsiders. - done
Carl Walker is not on the call; Email reflector overview is deferred until Carl returns.
- visibility levels of SJTAG forum and ATCA forum:
- Ian: right now SJTAG is set OPEN, ATCA is PRIVATE see also Ian's email with discussion of options;
- OPEN: Public access for discussion
- CLOSED: Public access for observation
- PRIVATE: Not open to the public
- Brad: do we want to publish minutes or only highlights?
- Adam: once (if) we become an IEEE working group, meeting minutes will be required; don't see a reason why not to publish minutes now, too, on the forum website;
- Al: How did we correspond during the MicroTCA meetings when fleshing out the JTAG support?
- Brad and Adam: It was more ad hoc email messages passed around the committee. Nothing formal.
- Adam: PICMG was a closed group. Not recommended to follow this process. Typically, working groups have a web site where public reviews of the information is displayed.
- Ian: you could add minutes to a private webpage, too, instead of the public webpage;
- Adam: for open process (allowing people to follow what we are doing), meeting minutes are needed;
- Brad: just want to avoid issues P1687 ran into (having been cited as being too closed)
- Brad: where to put documents (static website)?
- Ian: we already have a static webpage we can use [http://www.mcintoshuk.plus.com/sjtag/]; docs can be made available there;
- Ian: Concern over abridged minutes which may lose the thread of thought that went behind the reasoning of the views.
- Brad: We seem to have consensus to keep SJTAG Forums OPEN and the "ATCA" Forum CLOSED
Discussion: Concern over how to implement 4 pin JTAG interface if hosted from ATCA HUB boards which must also be tested by external tester.
- Adam: we need a connector (interface) for the external tester.
Interface must be on HUB board to facilitate connection and arbitration to the system.
- Brad: This may create arbitration issues if 2 HUBs are installed in the system.
- Adam: access needed from the outside for any external tester, backplane may not be the best spot;
- Adam: The real problem is the arbitration of which ATCA HUB is going to be master.
- External interface is really an arbitration on the HUB as to which interface is the master to the backplane - external vs. local.
- There could be some sort of priority system or monitor TCK activity.
- Brad: in MicroTCA there are two access points;
- Adam: external tester plugs into Master ATCA Hub board; either external tester or Master ATCA Hub controls the chain; real problem is the switching on the backplane;
- Ian: can we use slot ID's or something like that for determining priority;
- Al: There are IDs but have to ask others about this for ATCA.
- Peter: Multi-Drop will still be a problem with regard to multiple HUBs in a system
- Brad: the inactive, not the active, processor should run the tests on the offline/stand-by boards/modules in redundant systems (for periodic tests and maintenance); GoAhead commercial software performs this kind of function already in Linux.
Embedded Boundary Scan controller must not be active on backplane by default, to allow external tester to drive the chain (external tester could then be the primary access point by default);
- Brad: How do other sectors of industry outside of telecommunications manage test priorities. I remember my father talking about the coupling of autopilots that would verify their results with each other and then the offending unit would drop off-line if it disagrees with the other two. (Brad's father was an airline pilot)
- Ian: All backplanes are "custom" and that's what we want to move away from.
- Ian: Most of the avionic instrumentation either have redundant voting systems as you described or they have a method of degrading gracefully before total failure (e.g. RADAR) to limp along as best as possible.
- Brad: Arbitration for HUBs is covered by the system software architecture
- Adam: That falls apart with regards to external testing (Brad: This is why the default case is embedded TAP is disabled)
- Peter: test vector generation needs to avoid bus contention on backplane signals;
- Al: switch boards (hub boards) are likely to be similar (a few different types sold on the market) rather than people to custom design switch boards; However, we should cover all cases since it is in the ATCA spec that way.
Dec 3, 2007; 8am EST
New action items:
- suggestions for static website (structure, contents) - ALL
- send Ian links to IEEE websites for ideas - Brad
- suggest new meeting time (send times to Brad) - ALL
Any other business:
Meeting adjourned at 9:08am
P.S. Thanks to Heiko, Peter and Adam for providing their notes to aid in capturing the details.