Hi Champions, Attached are the minutes from today's meeting. There are action items listed for sv-sc sv-cc Neil -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean. Champions August 07, 2008 Conference call Thursday 8-10am PDT Attendees: ---------- 1. * Stu Sutherland 2. - Surrendra Dudani 3. * Brad Pierce 4. * Francoise Martinolle 5. * Shalom Bresticker 6. * John Havlicek 7. * Dave Rich 8. * Neil Korpusik 1. Review IEEE patent policy ref: http://standards.ieee.org/board/pat/pat-slideset.ppt Move: Shalom, John - assume the patent policy was read Passed unanimously 2. List of Mantis items for review 2.1 1728 SV-SC Introduce "let"statement - fixed - Unanimously passed by voice vote 12/04/07. - Sent back to SV-AC by Champions Nov 8th - Was approved by Champions - Sent to sv-sc by the Working Group March 27 - Sent back to sv-sc by Champions July 24 (mostly editorial changes) - SV-SC voted at 2008-07-29 mtg to approve Dmitry's updated proposal, which addresses the issues posed by the champions. Abstain: Tom, Steve, Manisha Passes: 8y, 0n, 3a - The sv-sc made another small update, based on input from Shalom. The wording of the BNF footnote at the end of the proposal for a constant let is a little awkward: 45) In a constant_let_expression, all arguments shall be constant_expressions and its right hand side shall be constant expression itself provided that its formal arguments are treated as constant_primary there. I think "be constant expression" should be "be a constant_expression". - SV-SC voted to approve latest edit at 2008-08-05 mtg. 9y/0n/0a. John - his issues were addressed in the updated version. Move: Brad - approve the proposal for Mantis 1728 Second: John Passed Unanimously 2.2 2327 SV-AC 2173 adds property case, need to add vacuity definition and multi clocking behavior in it - fixed - approved by voice vote 2008-06-30: 5y/0n/0a - sent back to the technical committee by the Champions in the July 10th, conference call. - Passed by email vote 5y/0n/0a called 2008-07-23 John - in "evaluation attempts" the word attempts was missing in several places. This was corrected. Move: John - approve the proposal for Mantis 2327 Second: Shalom Passed Unanimously 2.3 1900 SV-SC Add new 'checker' construct to SVA - fixed - 2008-01-15: Passed by e-mail vote, 7y/0n/3a. - The proposal failed in the Champions email vote which ended on Feb 4th, 2008. Failed with 2 no-votes - was moved to the sv-sc - the Champions feedback and a lot of sv-sc issues were addressed - Passed voice vote at SV-SC meeting, 2008-07-22, conditional on some specific minor edits (implemented in version posted by Erik Seligman on 2008-07-23). 9y/0n/3a Neil and Francoise abstained due to not enough time to review. Steven Sharp abstained because he believes it is under-researched and too many changes have been made too quickly, though he has not actually spotted any specific problems. Note: There are two proposal to review (part1 and part2). Part1 the 7/18 version, part2 the 7/23 version Dmitry updated part2 on 8/04, but it was not approved by the sv-sc. Email from John before the meeting: I noticed what I think is a small mistake. There is a switch based on whether or not 1900 passes. But in the "else" part (where 1900 does not pass) the text about the variables excluded from using Preponed values includes the checker variables. Shalom - All the issues flagged in his email about part 1 are now addressed - Tom created a new version with those changes, but it was not yet uploaded to Mantis. - His issues with part 2, will take time to go through John - sent email this morning: Part 2, p. 6. In the example, is v1 static? If so, then I think that the declaration assignment is performed only once, so the behavior of check_loop is not described correctly. This could be fixed by making v1 automatic or by changing logic [7:0] v1=0; to logic [7:0] v1; v1 = 0; John - see text on p7, the assign will only occur once. Shalom - see a mantis 1556, which is related to this situation. AI/SV-SC - make the change to the example shown above. Part 2, p. 13. It is stated that the assume set of F1 excludes F1.B1.u2 because the only formal argument of the child checker bar is bound to actual r that is an inactive free variable. Is it true that F1.B1.u2 is in the assume set of F1.B1? I think that the rules say that it is. John - described his concern in the meeting. None of the other Champions were prepared to answer this question. AI/SV-SC - We left this question for the committee to answer. John wants to make sure his understanding is correct. Part 2, p. 14. Miscounting. x |=> ##5 1'b0 fails 6 cycles after x == 1. John - the example at top of page 14. It will fail 5 cycles later. - |=> includes a one clock cycle delay. AI/SV-SC - change "five clock cycles" to "six clock cycles". Part 2, p. 14. The exception to Preponed sampling for constants and automatics is too tight. Local variables are excluded, as are the active free checker variables. In the last paragraph of 17.6.3 "Concurrent assertions have invariant scheduling semantics - whether present in checker code or design code. These assertions always sample the values available while processing the Preponed region, except for constants or automatics as described in Reference to 2398, <add the following text here> and they are always evaluated when processing the Observed region." John - last sentence, add the following text "local variables, and free checker variables" AI/SV-SC - add this text in the spot indicated above to the last para of 17.6.3 (first confirm that this text is precise) Shalom went over this remaining issues: (He skipped the trivial stuff in the conference call) Shalom - p2 next to the last dash list, mentions automatic variables. Can a checker contain a loop? In particular, a SystemVerilog style loop, where the loop index variable has automatic lifetime? John - might be illegal if don't reference the loop variable. Stu - a repeat loop might be better. - it would be confusing if can't use the declaration in the loop. John - thinks it is coming from context of the procedural context within a checker. Don't want that part of the checker to get values from the local context...? AI/SV-SC - this point should be made more clear. Shalom - p4, bottom, "A checker may be instantiated in one of the following, where a concurrent assertion may appear" - From "A checker may be instantiated in one of the following..." To "A checker may be instantiated wherever a concurrent assertion may appear (see 16.15)." - cross ref also seems to be wrong (16.14) should be 16.15. - Don't need the list on p4. Need to add generate block to the list in 16.15? AI/sv-sc - reword this and check the cross ref. , also update the list in 16.15 - part 1 adds an item to the list in 16.15. a generate block is not in part 1. (Missing?) - this may cause a chain reaction of changes (is it in 1728?) 1728 had the effect of propagating it to many places. AI/sv-sc - The two should be consistent. Shalom - this proposal does change it since it adds a checker to the list - the first item in the list, not worded the same, the wording in part2 should be exactly the same as 16.15 AI/sv-sc - The two should be consistent. Shalom - global issue throughout the proposal (ref: mantis 2398) - 2398 touches a bunch of sections. sv-sc should figure out the right cross references Stu - that is a definite issue, he would need to be guessing. Shalom - there are about a dozen places. AI/sv-sc - check all of the cross-references (taking 2398 into account) Shalom - p5, bottom, last sentence, "Procedural assertion statements" "Procedural assertion statements, assertion statements within a procedural block of a checker, shall be treated just like other procedural assertion statements as described in" Not worded well, is the part after the comma suppose to be a definition of a procedural assertion statement? - 2398 makes a distinction between static and concurrent. can there be an immediate assertion in a checker? Can immediate assertions be in a checker? AI/sv-sc - reword this sentence at the bottom of p5. There is also a question about the ability to have an immediate assertion in a checker. Shalom - p6, example, header of module m, "int [20:0] foo" is illegal. FM - bit [20:0] might be what they intended. Sahlom - there was also a problem in 2398. (const apostrophe) - const ` has the tick going the wrong way on page 6 of part2. AI/sv-sc - fix the illegal declaration. - const ` is wrong. Shalom - p6, "foo[const`(v1)]" is selecting a single bit. was the intent to use 8 bits? - maybe foo should be an array of bytes Stu - request to sc - change names foo, bar to something meaningfull AI/sv-sc - update the example so eliminate the inconsistency in number of bits - change foo, bar to some other name, which is meaningful Shalom - 17.5, p8, "event control statement", an event control is not a statement AI/sv-sc - correct this Shalom - "only monitored on the first clock tick" was deleted by 2398. John - thinks that statement is wrong. AI/sv-sc - reconcile 1900 part2 with the changes in 2398 Shalom - p9, near top, "A free variable may assume any value at every point in time. Formal analysis tools shall take into account all possible values of the free checker variables imposed by the assumptions and assignments (see 17.6.1). Simulators shall assign random values to the free variables as explained in 17.6.2" - Not sure that this means, "every point in time". At every time step, can have a different value? - was the intent meant to be - every tick of the global clock? John - thinks it means for each time step AI/sv-sc - check to see if this is correct, as written. Shalom - p9, after example, 2nd bullet, "If there is a reset, it becomes low at the next tick of the clock." - it wasn't quite clear what was being stated AI/sv-sc - check to see if this is worded properly. <this is how far Shalom got in his review of part 2> Move: Shalom - send the proposal for mantis 1900 part2, back to the sv-sc Second: Stu Passed Unanimously 2.4 2110 SV-SC Allow checkers in procedural for loops - duplicate - SV-SC voted to close in email vote ending 2008-07-21. Feature is now incorporated into 1900. 7y/0n/0ea/6ia (ea = explicit abstain, ia = implied abstain by not sending email) Move: Brad - approve the resolution of duplicate for Mantis 2110 Second: Dave Passed Unanimously 2.5 2411 SV-SC Allow using triggered method in sequences - duplicate - SV-SC voted to close as duplicate in email vote ending 2008-07-21. This is exactly the same proposal as 2415, accidentally entered twice. 7y/0n/0ea/6ia (ea = explicit abstain, ia = implied abstain by not sending email) Move: Brad - approve the resolution of duplicate for Mantis 2411 Second: Dave Passed Unanimously 2.6 2089 SV-SC Allow checker construct (0001900) to include final blocks with immediate assertions - duplicate - SV-SC voted to close as duplicate, in email vote ending 2008-07-21, since feature is now included in 1900. 7y/0n/0ea/6ia (ea = explicit abstain, ia = implied abstain by not sending email) Move: Brad - approve the resolution of duplicate for Mantis 2089 Second: Dave Passed Unanimously 2.7 1809 SV-BC forward references into $unit package - fixed - On March 25, 2008 the SV-BC unanimously approved the attached proposal. - The proposal was unanimously approved by the Champions in the April 24th, 2008 conference call. - The proposal was unanimously approved by the Working Group in the conference call of May 01, 2008. - Editor Note: I spent considerable time trying to add this proposal into Draft 6, and finally had to back everything back out. There were too many problems with the proposal to implement in its current form. The proposal needs to be rewritten to: 1) Use proper IEEE wording (no first person "we" or third person "you"). 2) Use paragraph styles consistent with the rest of the LRM 3) Correct grammar, spelling, and punctuation errors. - On July 21, 2008, the SV-BC approved the updated proposal (1809-v9.htm) to address this issue. The vote was not unanimous. Shalom Bresticker abstained because he did not have enough time to review the proposal. Shalom - Dave made more than just cosmetic changes to the proposal Brad - FM was also ok with it. Shalom - Gord was ok with it as well. Stu - looked at it last week. Move: Dave - approve the proposal for Mantis 1809 Second: Brad Passed Unanimously 2.8 2428 SV-EC There is a typo in the example of section 18.6 (page 468) i[1] is incorrect, it should be i[0] - fixed - Approved on July 21 2008, minor editorial changes Move: Brad - approve the proposal for Mantis 2428 Second: Stu Passed Unanimously 2.9 2448 SV-BC BNF for packed arrays of enums - fixed - On July 21, 2008 the SV-BC unanimously approved the attached proposal to resolve this issue. Brad - this mantis was created due to followup from Champions Move: Brad - approve the proposal for Mantis 2448 Second: Shalom Passed Unanimously 2.10 1835 SV-CC JEITA: APIs Give a overarching name for APIs - fixed - The SV-CC PASSED this option on 07/30/2008. (unanimous) - The latest proposal is a frame file The .pdf file is definitive, and is the one that the SV-CC voted on and approved. [Re: Jim Vellenga] Move: Stu - approve the proposal for Mantis 1835 Second: Shalom Passed Unanimously 2.11 2226 SV-CC VPI Handle behavior for HDL objects of dynamic lifetime is not defined - fixed - The proposal was sent back to the SV-CC by the Champions in the April 10th, 2008 conference call. - The SV-CC reviewed the latest proposal and PASSED it on 07/30/2008 (unanimous). - John Sheilds has now placed the mantis item into the resolved state. (aug 5th) <note: there are 5 parts, totaling ~80 pages> Stu, Shalom, Dave, Brad - haven't reviewed it Stu - expects to partially rely on cc anyway, will review at least some of it. Shalom - there is an incorrect cross reference. p806 36.16, detail 26 AI/Neil - Call for an email vote that ends next Wednesday (Aug 13th) Move: Stu - call for an abbreviated email vote for the proposal for Mantis 2226 Second: Brad Passed Unanimously 2.12 529 SV-CC 27.21 & 27.22 Cannot get a handle to a class instance - duplicate - On 07/30/2008 the SV-CC voted to declare this a DUPLICATE of Item 0528. Move: Stu - approve the resolution of duplicate for Mantis 529 Second: Brad Passed Unanimously 2.13 742 SV-CC Need VPI info on parameters that are types - fixed - The proposal was sent back to the technical committee by the Champions in the July 10th, 2008 conference call. The vote was unanimous. - The SV-CC PASSED this proposal on 07/30/2008 (unanimous). Shalom - p8, new detail 3), should first occurrence of vpiExpr be for a "value parameter" - Not a disaster, but not that clear Friendly amendments: FM - detail 3, "of a parameter" --> "of a value parameter" - detail 5 is not quite right, should only be for value parameter. "if a parameter" ---> "if a value parameter" - detail 4 overridden value parameter AI/sv-cc - update the proposal with these friendly amendments Move: Stu - approve the proposal for Mantis 742 with friendly amendments Second: Shalom Passed Unanimously 2.14 2447 SV-CC VPI data model needs to include enum vars in packed arrays (#1230 changes) pending SV-BC change (#2374). - fixed - The SV-CC PASSED this proposal on 07/30/2008 (unanimous). Move: Shalom - approve the proposal for Mantis 2447 Second: Francoise Passed Unanimously 2.15 2396 SV-SC Add edge identifier edge for events - fixed - Passed by voice vote in SV-SC meeting, 2008-07-22. 11y/0n/1a Abstains: Lisa - just not sure, seems like there are other ways to do it. - The proposal was sent back to the SV-SC by the Champions in the July 31st, 2008 conference call. - SV-SC approved latest edit in 2008-08-05 mtg. 9y/0n/0a. Shalom - one change in bug note was not made - the SDF change is ok AI/sv-sc Need to double-check all of the editorial changes: Shalom - Last paragraph in top section of p5. The last blue sentence before 29.4.3 should use "named sequences" since that is in the current draft. Move: Shalom - approve the proposal for Mantis 2396 , with Editorial amendments that are listed in the bug note, the change to last sentence before 29.4.3 also needs to be updated. Second: Dave Abstain: Brad - doesn't think it is necessary, thinks it is confusing. 1. Some people will think that @ some_name is level sensitive instead of edge sensitive 2. Doesn't like that all the editorial issues were ignored. Passed Unanimously 2.16 2088 SV-SC Allow Checker construct (0001900) to include covergroups - fixed - Passed by voice vote at SV-SC meeting 2008-07-22, conditional on Tom reviewing p.3 example with two identical bins after return from vacation: did he intend this legal but odd case to illustrate something subtle, or should we replace it with a more typical case of disjoint bins? 8y/0n/4a Abstains: - Gordon based on email vote, likely to be substantive user-based issues resulting in implementation divergence. Risk factors are too high. Addresses current user needs, but will have future changes. - Steven based on same issues as Gord and hasn't reviewed the proposal. - Mirek: No expertise in cover groups. - Manisha: Agrees with Gordon - The proposal was sent back to the SV-SC by the Champions in the July 31st, 2008 conference call. - SV-SC approved latest edit in 2008-08-05 mtg. 9y/0n/0a. <was not discussed - it is dependent on 1900, which we didn't finish> 2.17 2434 SV-SC Changes to seq/prop defaults and typing of actuals - fixed - Passed in voice vote at SV-SC meeting, 2008-07-22. 9y/1n/0a Mark Hartoog votes no - not needed, creates backwards compatibility issues. Clarification - not 100% sure about the note at the end, but the whole basic idea of casting these expressions to their self-determined type has the potential to create very subtle backwards compatibility issues per the face-to-face. - The proposal was sent back to the SV-SC by the Champions in the July 31st, 2008 conference call. - Gord was on vacation and was not able to address this Champions issues. Mike Burns will work on a new proposal. - The svsc has requested input from the Champions on any additional issues that they might see with the current proposal. - Mike posted a proposal Aug 6th. John - annex F - rewriting alg. which didn't align to rewording of words for exclusion cases. - discussed part of it with Gord before he left for vacation. new wording is then fine, but annex f still needs to be aligned with the new wording. - flattened property and flattened sequence. - also an impact with 1549 - local variable formal arg with default expressions (?) text said it was illegal, but there was an example using it. Stu - p4 example, b_d - need to show where declared? - assuming that whatever is using this sequence has a b_d? John - no, it resolves in the declaration of the sequence. Dave - p4, add a "logic b_d, d_d;" to the example. AI/Dave - he updated the proposal in the meeting. John - we were trying to disturb existing text as little as possible. Stu - the comment about b_d is still confusing John - what is it called in a task, function? Dave - it is a default actual argument John - what is on lhs is the formal, the right side is the default actual arg if none is passed in. Stu - he is now ok with it. Shalom - "default actual arg" is only used in assertions Move: Stu - approve the proposal for Mantis 2434 with the friendly amendment Second: Dave Passed Unanimously 2.18 Mantis 2398 - John - p8, example, the declaration assignment needs to be added back in. - otherwise the first time through is not right for other times other than time 0. - need both the declaration assignment and not allowing the posedge clk at time 0. AI/sv-sc - make this change to the example John - another issue - p2, in the else part, shouldn't be talking about free checker variables. - John already sent email on this, but it was not fixed yet. AI/sv-sv - make this change 3. Next Meeting: Aug 21 // assuming that the Working Group will approve it. NeilReceived on Thu Aug 7 18:26:30 2008
This archive was generated by hypermail 2.1.8 : Thu Aug 07 2008 - 18:26:55 PDT