Minutes of Weekly Meeting, 2010-06-28
Meeting called to order: 10:37 AM EDT
1. Roll Call
Carl Walker
Eric Cormack
Adam Ley
Heiko Ehrenberg
Patrick Au
Ian McIntosh
Brian Erickson
Brad Van Treuren
Tim Pender
Peter Horwood (10:38)
2. Review and approve previous minutes:
6/21/2010 minutes:
- No corrections noted.
- Motion - Brad, Second - Patrick, no objections or abstentions.
3. Review old 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?
- All to consider what data items are missing from Data Elements diagram
- All: do we feel SJTAG is requiring a new test language to obtain the
information needed for diagnostics or is STAPL/SVF sufficient?
see also Gunnar's presentation, in particular the new information he'd be
looking for in a test language
(http://files.sjtag.org/Ericsson-Nov2006/STAPL-Ideas.pdf)
- Ian/Brad: Draft "straw man" Volume 4 for review - Ongoing
- All: Review "Role of Languages" in White Paper Volume 4 - Ongoing
- All: Review 'straw man' virtual systems and notes on forums:
http://forums.sjtag.org/viewtopic.php?f=29&t=109. - Ongoing
- All: Add to, or comment on, the bullet point list of architecture drivers. -
Ongoing.
- All: Provide forum comment on the graphics used during the meeting; suggest
"building blocks" that may be used in future:
http://forums.sjtag.org/viewtopic.php?f=29&p=257#p257 - Ongoing.
- ALL: review / comment in preparation for upcoming meetings. - Ongoing
- Tim: Draft matrix of SJTAG features against evolving solution options. -
Ongoing.
4. Discussion Topics
- White Paper Volume 3 Review - What do we need for a generic view of
gateways/linkers/controllers.
- [Ian] With where we are in the slides, we need to think about how we get to
a generic representation of gateways, linkers, etc.; what are the things we
feel we need to know about them to describe them?
- [Ian] I'm not sure how best to handle this - would bringing up the existing
slides help at all?
- [Brad] That may give us somewhere to start.
- {Shared SJTAG Volume 3 Consolidated V6.ppt}
- [Brad] In my view, we can really decompose this into two areas, Control and
Configuration, but I don't know if that's the right strategy.
- [Eric] I think you're definitely on the right path with that.
- [Brad] What I'm seeing on P1687 is what they're calling the SIB, where you
have one bit that you turn on or off to enable or disable the path. How do
we model something like that?
- [Eric] I think Tim talked about that recently?
- [Ian] Maybe that was Michele?
- [Tim] I was talking about selecting between multiple hosts.
- [Brad] Yeah, that's a whole different thing: First we have to describe how
we manage the secondary side, then we can get into managing the primary
side.
- [Brad] Do we think it's possible to generically describe these devices?
- [Patrick] It's be difficult.
- [Brad] I'm just looking for a gut feeling at this stage. We know that the
tools already support these things, just not in a consistent way.
- [Tim] What we've been presenting here is more of a detailed view. You need
drivers to select the path; there's no one method because you have various
things: ScanBridge, ASP, etc., each with different protocols. You'd need to
have abstraction and deal with it at a high level.
- [Brad] Your use of the word 'various' is interesting because I was doing
some Microsoft COM programming recently and they use 'variants' to deal with
differing data types, and I wondered if something similar could apply here.
However, I think you get into just the issues you mentioned. The variant
deals with storage space and the data type, but not the protocols.
- [Brad] I think this is coming down to the protocols behind the control
block. Ultimately, there's a GPIO driving the lines with some means to
control that GPIO.
- [Brad] Sometime when Michele is on the call, maybe we can discuss a language
we presented to IJTAG, NSDL, which allows us to separate structure from
function. Maybe we have to deal with this from a software perspective and
define interface that have to be available to be SJTAG compliant. Maybe the
detail of the implementation is outside of SJTAG, coming from the device
provider.
- [Peter] All the tools handle these differently. As a vendor, we'd want to
supply a single document that all tools use.
- [Ian] I'd think that should be the aim.
- [Brad] Yes, and even for some of our own gateway solutions, where we have
speciality features, we'd want to do that. Trying to find available ways to
control things is why I2C based solutions have evolved.
- [Brad] I tend to separate function from behaviour: A bridge adds itself to
the chain but an ASP doesn't, but both connect a chain to the board edge.
- [Ian] I think that this area of gateways, selectors, etc., is key to SJTAG.
We've got to get this right, because the whole point of SJTAG is managing
chains on several boards. We have to be sure we've covered all the avenues,
so I think we're going to have to spend some time on this.
- [Brad] Should we set up a Tiger Team to try to identify behaviours of these
devices? Key aspects that tools have to deal with?
- [Ian] Any thoughts on that idea?
- [Eric] I'd have difficulty in committing.
- [Brad] Is anyone unfamiliar with the subject and would like some more
explanation?
- {Silence}
- [Ian] OK, maybe we'd be better to just take a couple of weeks out of our
meeting schedule to work through this as a team. Maybe we can each learn
something along the way.
- [Brad] That approach seems to have worked in the past.
- [Brad] I really appreciate having Peter's input here today.
- [Peter] I guess I'm biased and feel that our devices do it right, but I'm
familiar with the modes of operation of all the other devices. One of the
questions is how far do you go? Do we go the the level of these lines are
toggled or do we leave it all at a higher level, allowing a tool to
implement value add to gain commercial advantage?
- [Brad] I think we should look at it at a higher scope; allow tools to
differentiate themselves.
- [Peter] Then you have the issue of feature set, our parts are quite feature
rich. How do you get a common description that covers all the features? We
could describe an API of the core features, but how do you address the
things that outside the core?
- [Brad] I had in mind a Venn diagram that we could use to capture the feature
set we need for SJTAG.
- [Peter] OK, the minimum feature set that's needed to be SJTAG compatible:
It'd need to be addressable, then we'd quickly get to the question of
whether TRST needs to be managed locally, and that goes right back to one
of our earliest actions.
- [Ian] Yes, one that's still there today!
- [Brad] That's why I think we need the Venn diagram, it may give us that
'Aha!' moment.
- [Ian] OK, we'll set aside the next couple of meeting for this, maybe through
July.
- [Brad] Maybe we can all email Ian with suggested features?
- [Tim] Email or on the forums?
- [Ian] Forums maybe works better {ACTION}
- [Brad] Let's keep to the gateway interface for now, not linkers or
controllers.
- [Tim] What devices use the sync bit?
- [Brad] Well the Lattice Scan Linker does, the LASP, ScanBridge, and Peter,
your device do too I think?
- [Peter] Yes, they do. Pretty much all of these devices do, in fact.
- [Brad] Tim, I guess you're trying to avoid sync bits?
- [Tim] Yeah, we're using vendor tools a lot, so we want to get rid of sync
bits.
- Sample Webpages to Present Survey Results
- {Shared http://www.sjtag.org/survey2009/survey1.html}
- [Ian] I've built these pages to give a preview of the layout I'm proposing
for presenting our survey findings. Is there anything that comes out?
- [Brad] In question 1.1 - 1.3 can we give a breakdown of percentage in
Mil/Aero, etc?
- [Ian] I think these were really just name, company and email address, I
don't think we collected industry sector information this time.
- [Brad] When I did the 2006 results, I struggled to get a good layout. This
looks better.
- [Ian] OK, but you were probably constrained by the PowerPoint layout. I can
get a lot more on the page, and it'll adjust if the page gets shrunk, to
some extent at least.
- [Eric] In Chrome or FireFox, the layout goes narrow and has the picture
above the text, but in Internet Explorer the pictures are at the right.
- [Ian] That sounds a bit like the FireFox float drop problem again. I kind
of got out of the habit of checking pages in FireFox. I guess I need to look
into this - it can probably be fixed by adjusting the page CSS.
- [Eric] I get a Not Found if I go 'Next' then 'Next'?
- [Ian] Yes, there's no Page 3 yet.
- [Ian] I'd like to get everyone to help with preparing some of the text to
go with each question. I know we're all busy, but there's no desperate rush.
- [Patrick] What is the timescale? My workload gets a bit lighter in
September.
- [Ian] As I say there's no desperate rush. Maybe we can use the forums again,
post some draft text against the existing survey thread. That way everyone
can see which ones have been done. {ACTION}
5. Schedule next meeting
Schedule for July 2010:
No meeting July 5, 2010
July 12, 2010
July 19, 2010 - Brad will miss
July 26, 2010 - Brad and Brian will miss
6. Any other business
The motion on signal naming conventions from the meeting of May 3rd remains
tabled.
This meeting marks the end of the 2nd quarter. There are no changes to Core
Group Membership, based on attendance during the period.
7. Review new action items
- All: Post suggestions for key SJTAG gateway features on the forum (Ian will
create topic)
- All: Post suggested draft texts for survey comments in existing 2009 Survey
thread (http://forums.sjtag.org/viewtopic.php?f=32&t=83)
8. Adjourn
Eric moved to adjourn at 11:32 AM EST, seconded by Peter.
Respectfully submitted,
Ian McIntosh