Attendees: Mark Hartoog Cliff Cummings Stu Sutherland Neil Korpusik Brad Pierce Shalom Bresticker Karen Pieper Charles Dawson Tom Fitzpatrick Yonghao Chen 0. process Shalom had a concern that our items may slip through the cracks and not make it to the champions or P1800 committee. In response we agreed to review the status of all of the items on the spreadsheet. Karen - None were marked as resolved before the 4/19 champions meeting. Needs to be in the mantis database marked as resolved. Need to pull from btf database to the mantis database for the champions review and for stu's editing. May 2nd, next champions meeting. May 5th, next P1800 meeting. May 9th, champions again We can make changes beyond the ballot feedback if absolutely necessary. Johny wants us to avoid doing this. These changes then become subject to objections in the reballoting process. Steven - will be unavailable next week (at CTC). 1. spreadsheet 576 - "resolved" - annex H - encryption - vlrm status (Karen sets this) 577 - "resolved" - replacing text These two Encryption issues will be discussed at next Champions' meeting 687 - btf 287 - item 8 - assigned to stu - 26.4 `compatibility - stu has enough information to implement it. There is some leeway for the editor - how ensure done correctly? Stu will add a formal proposal to the mantis database for this. 680 - etf 350 - 9,10 - config issue - configs outside of source files - etf 99 - - etf 582 - semicolon stuff also - [Shalom] - Brad - there are probably a dozen items related to configs in the etf database. ---> left to later in this meeting - and then resumed - Shalom - don't see consensus for Steven Sharp's proposal. Cliff - will vote no and will submit a new proposal within 2 days. - the proposal will have multiple options to choose from. - there may be no consensus - backward compatibility issues. - Cliff assumes that he will be able to get better consensus. Mark - The committee chose config vs. configuration so that it would look different from a VHDL configuration. Some tools could add an option to continue to support backward compatibility. Steven - thought it would be best to not to try to push something through due to a current lack of consensus. 704 - etf 613 - approved at last meeting - stu abstained Tom has a mantis problem to create a new mantis item for it. Karen - created a mantis item for it. 688 - etf 618 - $timeskew - assigned to shalom - was split into 2 pieces 1. ballot comment issue - superset of 618 $timeskew $skew, $fullskew - minor corrections of these as well Marked "resolved" in Mantis - will appear on next Champions' spreadsheet 2. separate issues - with $fullskew - in Boyd database 659 Fitz created Mantis item 705 for this issue. See discussion below AI/Tom - update mantis 688 with the results of 04/07 meeting. DONE - open a new mantis item for etf 659 [= Mantis 705] DONE, , includes intermediate proposal. 651 - 18 - 1364 passed it - assigned to vlrm - minor change to example 653 - 19 - cut and paste error 654 - 20 - assigned to vlrm 655 - 21 - $nochange note - passed on 4/7 unan. - missed champion meeting - to be addressed at next champions meeting - currently marked as new 656 - 22 - fro syntax issue - approved ? 657 - stu1 - stu needs to provide a proposal; resetall not allowed in a module 658 - stu2 - tom wrote a proposal for this one - notes issue AI/tom, cliff - need to make sure consistent with p1800 266 - was erroneously marked as resolved. The committee needs to vote on this before it becomes resolved. AI/all - review this proposal - for champions by next Wednesday, May 4th. 659 - stu3 - simple issue - continuation of a table 660 - - 667 - - execution at time 0 - wasn't in the spreadsheet (assigned to 1364 from sv-bc) Shalom - there seems to be agreement on the following points - always blocks cause an event at time 0. - continuous assignments execute unconditionally at time 0. issues: - should primitives or UDPs be mentioned - missing detail - all inputs to primitives except 1 bit inputs form an implicit continuous assignments. 1-bit inputs are covered by port collapsing. Dave Rich has written up a proposal. Steven doesn't like part of it. Steve is still mentioning primitives. All of this was done over email. Steven - the proposal should only mention continuous assignments and then add details on implicit cont assigns. There would also need to be details on primitive port connections and port collapsing. For switch primitives strength preservation is needed. Cliff - can we say that it is an error to attach a vector to a primitive port? Karen - or-reducing the vector sounds like a strange approach. Steven - this situation is completely undocumented. Brad - thinks of this as a wired-or. Tom - seems to be that there is a 100% chance of inconsistent current implementations. Cliff - would be happy to define this. Shalom - the ballot issue does discuss primitives. Steven - need to eval at time 0 if connected to constants or 2-state variables Steven - section on continuous assignments should be updated. 5.6.1 - could be added here. Based on further discussion it sounded like both Steven and Shalom were in agreement on how to proceed with this item (They were in "violent agreement"). Tom - can we say that a primitive port acts similar to a port connection? AI/Steven - write up a proposal - by this Friday (4/29). --- etf 659 - $fullskew Shalom - not too many people have been participating What to do with remain_active_flag in $fullskew. May also affect $timeskew. $fullskew appears to be fairly redundant. May need to go back to the original proposal from '97-'98 time frame. This would change the current description. Shalom - would like to get input from the entities about their concerns. Steve Wadsworth at AMI was the chair of the ATF (ASIC task force). steve.wadswort@poci.amif.com Dave Roberts at Cadence was involved in the past. Yonghao - the flag seems to be redundant. Shalom agreed. You appear to get the same behavior when the flag is either 0 or 1. Tom - leave $timeskew alone? An intermediate proposal exists. All issues mentioned are not yet resolved. Fitz suggested that Shalom/Yonghao continue with proposal, but treat fullskew as completely separate from timeskew wrt/ this flag. AI/Cliff,Yonghao - contact original ATF members to understand original intent of this functionality. AI/Shalom and Yonghao - try to finish up the proposal. May 6th (next Friday) Open items: ---------- - 705 - etf 659 - $fullskew - 680 - config - 667 - time 0 - 658 - notes For Monday: 1. config issues 2. time 0 proposal 3. notes proposal - vote by Wed May 4th.