Minutes of Weekly Meeting, 2013-04-01

Meeting called to order: 11:06 AM EDT

1. Roll Call

Ian McIntosh
Brad Van Treuren
Carl Walker
Heiko Ehrenberg
Adam Ley
Brian Erickson
Harrison Miles (joined 11:07)
Eric Cormack (joined 11:12)
Tim Pender (joined 11:14)

Patrick Au
Peter Horwood

2. Review and approve previous minutes:


  • No corrections noted.
  • Carl moved to approves, seconded by Heiko, no objections or abstentions.


  • No corrections noted.
  • Carl moved to approves, seconded by Adam, no objections or abstentions.

3. Review old action items

  • 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
  • Harrison will attempt to come up with a table of use cases and their associated layer and what can be done at that layer. Ongoing.

4. Reminders

  • Consider Adam's three points (from the action from the first weekly meeting) and suggest what is preventing us from answering those questions:
    • Establish consensus on goals and constraints
    • What are we trying to achieve?
    • What restrictions are we faced with?
  • Forum thread for discussion: http://forums.sjtag.org/viewtopic.php?f=3&t=172

5. Discussion Topics

  1. Follow up on Mike Westermeier's slides/Harrison's slides
    • Brad had previously indicated to Ian that there were some aspects of Mike's slides that he'd like to discuss further. One area that had direct relevance to recent discussions was the overlap between Brad's protocol stack and Mike's vision. This also had a bearing on the slides Harrison had prepared.
    • {Presentations 4-4-Westermeier-STAPL_BTW2005_mw_11_3_05.pdf (BTW 2005) and BTW2006-Van-Treuren.pdf (BTW 2006) shared http://files.sjtag.org/BTW2005/4-4-Westermeier-STAPL_BTW2005_mw_11_3_05.pdf
      http://files.sjtag.org/BTW2006/BTW2006-Van-Treuren.pdf }
    • Considering slide 5 of Mike's presentation, Brad saw this as defining the some primitives, which included some nontraditional test operations that nevertheless leveraged the capabilities of 1149.1. Harrison was of the opinion that Module Identification was the only high level item here. Brad agreed for the dependency tree perspective. Ian noted that most of the things listed were more complex than vector level primitives, while Harrison claimed that these were well established within the JTAG domain while Module Identification was a relatively new concept. Brad felt that had been around for some time too, but it was significant for SJTAG that this had more of a board level scope while the others operated more at the device level.
    • Moving on to slide 6, Brad observed that Mike's Application Layer equated to Brad's Application, the Hardware Layer to the Test Access Port, while everything else was wrapped up in the Embedded JTAG layer; Mike had in fact simplified much of the stack but this still carried the requirement to define the interfaces between blocks.
    • Harrison proposed that while JTAG may be our starting point we should remain open to the prospect of alternate interfaces, in the same way that P1687 has made that provision. Brad referred to slide 10 of his presentation and noted that this was essentially what P1687 was alluding to. Harrison added that this could influence the mid-plane or backplane architectures but didn't mean that a board couldn't have a scan chain.
    • Brad explained that this was why in TFCL he tried to abstract to a 'test step' and simply manage when it gets applied. What he and Mike had debated at BTW was whether you could do this with just STAPL or did you need more traditional software languages such as C++? Mike had argued that he could integrate all that he needed using STAPL. Harrison commented that P1687 had added things that moved in that direction, such as Access Links and callbacks. Brad suggested that the EXPORT function in STAPL was really a callback, or at least was implemented like one.
    • Harrison felt that Access Links helped to deal with the case of COTS items being mixed with in-house items.
    • Brad pointed to his slide 13 and considered that this discussion was showing that the whole 1149.1 and Set/Clear strata needed to be fleshed out more: We are not defining how the communication might differ case-to-case because there is no standardization here. Brad added that he and Michele had been working on NSDL to fill that void.
    • Referring to Mike's slide 6 again, Brad thought that Process Operation was defining the hardware interactions. Harrison noted that some clients were asking for the hardware layer to be a higher speed bus. Brad added that others wanted it all to be I2C, but harrison felt that lacked the necessary speed for things like quad core cellphones. Brad questioned if we could still be 'SJTAG' if we deviated from 1149.1, although Harrison didn't seek to exclude 1149.1, simply allow for other mechanisms, that may in fact emulate JTAG.
    • In Mike's slide 7 he set out his 'goals' for a JTAG language. Ian remarked that the first bullet was the reuse topic that we had often identified in previous discussion. However Ian wondered what the scope of 'Production' was intended to be here - Did that extend post-manufacture and into the field? Harrison saw this as a pertinent issue as many firms were starting to recognise the potential for field service to exploit JTAG. Brad noted that direct reuse may not be possible, but 'salvage reuse' was more likely. It was more expensive to develop the system level tests but it could leverage the previously developed board tests. Ian added that in his experience the board was gradually pared back or constrained until it ran in the system.
    • Brad questioned the second bullet on checking file integrity, and felt that may be more of an issue for the environment, e.g. a file being extracted from a zip is already being checked so a CRC check within STAPL is largely redundant. Harrison thought it is more a question of assurance rather than integrity. Ian added that a common QA question was how they could tell that the files in use were the intended ones, less so whether the files were 'valid' at all. Harrison felt that someone needed to take ownership and provide certification; it was hard to make it work automatically. Brad suggested that via TFCL as a single point of entry this might be simpler. A change to the TFCL program would align everything else, and with the JTAG Plug'n'Play where the tests reside on the UUT then it doesn't have to look to some other resource to find its data.

6. Key Takeaway for today's meeting

  1. The Test Access Port/Scan Interface strata needs to be expanded (e.g. to accomodate alternative interfaces).

7. Schedule next meeting

Next Meeting:
April 8.

April schedule:

8. Any other business


9. Review new action items


10. Adjourn

Eric moved to adjourn at 12:01 PM EDT, seconded by Tim.

Respectfully submitted,
Ian McIntosh