The System JTAG Working Group

Minutes of Weekly Meeting, 2010-06-07

Meeting called to order at 10:38 AM EST

1. Roll Call

Eric Cormack
Ian McIntosh
Patrick Au
Carl Walker
Adam Ley
Tim Pender
Brian Erickson
Brad Van Treuren
Heiko Ehrenberg
Peter Horwood

2. Review and approve previous minutes

05/24/2010 minutes:

Draft circulated on 25th May:
No corrections noted
Eric moved to approve, seconded by 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.
Ian: Make initial attempt at migrating some presentation material onto Wiki. - COMPLETE
Ian: Update draft Newsletter with sample primitive graphics and sample of survey responses. - COMPLETE

4. Discussion Topics

White Paper Volume 3 Review - Discussion of system description diagrams
- [Ian] I'm assuming that the slide set has not progressed any since the last meeting?
- [Brad] I have been making some revisions, but I didn't think they were stable enough to bring to the rest of the group yet. I'm mainly trying to focus on standardizing on the naming of signals going into our blocks and on creating the small graphics for then to use later. I also need to add in the suggested changes from three weeks ago.
Feedback on draft Volume 3 Wiki page
- {Shared http://wiki.sjtag.org/index.php?title=System_Architectures}
- [Ian] I've linked this from the existing Volume 3 wiki page, so that it is fairly easy to find, but doesn't disturb the existing page too much.
- [Ian] As we said we didn't want to simply regurgitate all our slides, I've tried to pick and chose a few slides that seemed to best capture the key elements and then fill the gaps using the text. The images are thumbnailed and may not always look readable, but they can be clicked on to get a larger view.
- [Ian] I trying to write up the text, I found I wanted to talk about the distinction we made between a TAP Interface and an External Path, but I didn't want to interrupt the flow of the main text. I wanted a sidebar, but the wiki didn't offer that, so I had to spend a little time creating a wiki 'template' to do that for me.
- [Ian] I've covered the basics of scan cells and chains and carried on into path selection.
- [Ian] I only went as far as to the discussion on Switcher vs. Linker and the 2-Path-Linker diagram; That seemed to be just about as far as the slides take us just now.
- [Brad] You�ll want to add a diagram for the Gateway that I�m working on.
- [Ian] Yes, but for now (until we have additional material) I think this Wiki page is pretty much as far developed as it can be. I don�t think we should add too much more from the existing slides to the Wiki;
- [Tim] Is there a printable version of this?
- [Ian] The original plan was to use the wiki as a collaborative writing tool and then port the whole lot across into Open Office Writer to produce a printed version of the White Paper, but I'm not that's really what you're asking.
- [Tim] In the Wiki tool box you can click "Printable Version".
- [Ian] In Print Preview it looks kind of OK; you lose some of the layout and the sidebars aren't really that distinct since you don't get the background.
- [Eric] If you go into the Page Setup you can select 'print background colors and images'. Then it looks OK.
- [Ian] Ah! OK, that's good.
- [Tim] Yeah, that seems OK.
- [Brad] I think you did a great job of getting the 30 slides or so down to this very crisp and concise Wiki entry, and this is the way we should go. A lot of this going to be background; we still have to show how all this relates to SJTAG.
- [Ian] Our target audience for the White Paper is probably familiar with the basics of JTAG, so that we don�t need to nor want to spend too much time on those basic details. The complete slide set could be much more useful for a tutorial-like setting, where some of the audience may not have the same background knowledge (which we can then provide with our slides).
- [Ian] Are there any other comments on what I have done so far? Is it OK to continue in this format?
- [Eric] Yes.
Recent forum activity
- {Shared 'Adaptive Clocking / JTAG Synchronization', http://forums.sjtag.org/viewtopic.php?f=3&t=119}
- [Ian] Tim made this post on devices that use TCK_RET (Return TCK). I don't think it's really a new subject but I haven't encountered any devices that use this and it isn't something we've discussed. I guessing this is something that you are running into, Tim?
- [Tim] Yeah, well this becomes an issue where you need to support multiple host controllers, where maybe you need a controller for DSP emulation as well as, say, an external test controller. The TDO coming out has to be matched with the TCK_RET for skew/delay.
- [Tim] At system level the emulator has to go through a gateway then the TCK_RET needs to also be routed in a similar way to keep the delay matched.
- [Ian] Some gateways have one or more uncommitted signals that can be used for things like this. But it becomes another factor in choosing which gateways suit your application.
- [Brad] Some gateways provide a 'Connect' signal to drive some external logic that can route additional signals. We've used that as some gateways only have unidirectional signals, which didn't really work for us in some cases.
- [Tim] Even at the board level there are problems. I have a CPLD that supports multiple hosts; in that I've included a passthrough that has no linking mux delays, and that may come into play here.
- [Brad] Tim, you may want to add a link to the Board Test Workshop paper you presented to the forum post.
- [Tim] OK, I can do that. {ACTION}
- [Tim] The code for the CPLD in that paper didn't include the passthrough I just mentioned.
- [Brad] No, but it's still a useful reference.
- [Ian] At the present, we don't seem to get into much emulation work at the system level; that's predominantly all done at lower levels of integration. But things could change.
- [Tim] In a lot of cases emulation won't work at system level anyway: The emulation pod might be expected to be withing and inch or so. There can be many parallel signals that need to be accessed and trying to get the timing right could be nearly impossible.
- [Ian] On the MPC555 we tried to extend the emulation port. We could daisy chain two or three of the flexi-cables that came with the pods to make a longer cable that worked OK, but if we tried to replicate that in our own tracking or cables it wouldn't work. There was obviously something about the construction that we weren't understanding.
- [Ian] But wasn't emulation one of the things that 1149.7 was expected to address?
- [Brad] That was our hope! You often have to reset to get the emulation to work and then you lose a lot of the state information that you wanted to examine. Many processors now let you probe to get some state of the processor even though it's not emulation as such.

5. Schedule next meeting

June 14, 2010

Schedule for remainder of June 2010:
June 21, 2010
June 28, 2010

6. Any other business

The motion on signal naming conventions from the meeting of May 3rd remains tabled.

[Ian] One other thing that I wanted to mention was that submissions for Poster Session at ITC close on June 20th. I wanted ask if we felt that what we were doing now on Volume 3 could reasonably make a poster?
[Brad] I don't know if we are going to be far enough along. I have a concern about who would be presenting, as I won't be able to attend.
[Eric] Do we know who will be present?
[Ian] I can't make any firm statement right now, but I'm fairly confident that if we got an acceptance that I would be able to attend.
[Heiko] Since it is in Austin, I will be there.
[Patrick] We ought to do something; we weren't there last year. The slides seem pretty good to me.
[Adam] I would expect to be able to attend.
[Brad] There's nothing that's new; it's more a discussion of what's presently available. We first need to understand the problem domain before saying how we propose to standardize. I don't know if we'll have the maturity in the discussion in the next month and a half.
[Ian] OK, but like Patrick, I'm quite concerned that if we run through a second year with no ITC representation then we'll drop off people's radar.
[Brad] Based on my experience, I don't think it'd be approved on what we have so far.
[Heiko] We could try to set up a Fringe Meeting, and if anyone turns up we can present the slides then.
[Brad] I'm just thinking of past Fringe Meetings - there was only about two nonmembers that turned up in each case.
[Heiko] I think that's pretty much the average for Fringe Meetings just now.
[Adam] It's a matter of promotion. I think the number of people who go seeking the Fringe Meetings now are almost negligible. You can ask for RSVPs.
[Ian] We could send out a mailing to try to gauge interest.
[Brad] What about Volume 2? What is there that's new there?
[Ian] That's the Use Cases.
[Brad] Yes, we presented something on those before.
[Adam] One thing I'd say, is that I think there should be something 'new', but not necessarily 'novel'. The organizers are always looking for items on standards in work, so I think they get viewed quite favourably.
[Adam] It is only the one page abstract that is required by June 20th. The presentation itself can pretty much be done on show day. However, the one page abstract is what gets included in the Proceedings.
[Brad] Will we have enough of Volume 3 accomplished by November?
[Patrick] That's four months away?
[Brad] I agree that having a date cast in stone may be something to help focus our minds, but we'd need to have a good handle on how want to manage the selected architectures.
[Patrick] Can we discuss this next week?
[Ian] I think we need to. Maybe exchange some emails between now and then.
[Brad] Or have a forum discussion.
[Ian] Yes, I'll set up a starter post for this. {ACTION}

7. Review new action items

Tim to add links to his BTW2003 paper in the 'Adaptive Clocking / JTAG Synchronization' forum post (completed during meeting)
Ian to create a discussion topic for participation at ITC 2010

8. Adjourn

Peter moved to adjourn at 11:44 AM EST, seconded by Tim.

Respectfully submitted,
Ian McIntosh