SV-EC Committee Meeting Monday June 26 2006 11:00am - 1:00pm PST [Minutes distributed for approval at next (sv-ec) committee meeting] (10211210202 ) Day (99333708256 ) (10000000000 ) Month (21123345566 ) (00000000000 ) Year (56666666666 ) --------- Attendees ---------- (AAAAAAAAAAA ) Arturo Salz (-AAA-A-AAAA ) Cliff Cummings (A-AAAAAAAAA ) Dave Rich (AAAAAAAAAAA ) Francoise Martinolle (AAAAAAAAAAA ) Mehdi Mohtashemi (AAAAAAAAAAA ) Neil Korpusik (AAAAAAaAAAA ) Ray Ryan (-AAAAA-AAAA ) Gordon Vreugdenhil (----AAAAAaA ) Steven Sharp (-A--AA----A ) Phil Moorby (---AAA---A- ) Stu Sutherland (--AAAaaA--- ) Surrendra Dudani (AA--------- ) Doug Warmke (-----a----- ) Don Mills (Aa--------- ) Chris Spear ^^------------ Considered ONE MEETING. ** Minutes taken by Neil Korpusik and Mehdi Mohtashemi ////////////////// June 26, 2006 ///////////////////////// Logistics note: You can use *6 for mute/unmute key sequence Agenda: 1. IEEE patent policy ref: http://standards.ieee.org/board/pat/pat-slideset.ppt Move: Cliff- Assume that the patent policy was read Second: Gord Abstain: none Opposed: none passed 2. Review Meeting minutes, of June 5th, 2006 [not available, will be posted on the sv-ec site] 3. Action items review. [Set for next time] 4. Mantis items that have a proposal Randomization/Constraint related --------------------------------- 889 (Randomization)clarifications for resizing dynamic arrays (Doug) Ray: update the 889. Significant for dynamic arrays or handles, randomize does not allocate any objects. NOTE: Some people still can't see any of the blue or red or strikeouts. Gord - is there a statement about NULL handles? - bullet 6 - take out the () (parenthetical comment) Arturo - has an action item to create a new mantis item for null handles. - rand_mode of items - if integers, don't retain rand_mode state didn't have a real preference on this. - set rand_mode to active for all of the elements that get added, those that are retained will keep their current rand_mode values. Ray - "In resizing a dynamic array by randomize or new the rand_mode of each existing element is preserved and the rand_mode for each new element is active." Francoise - how do users know which were added versus which were retained? It could be difficult to debug. Dave - "new"ing a whole new array? Ray - no. retain elements from original array. for handles, you retain the original handle values. Dave - what about pass by reference? Adding elements, ref to old elements are no longer valid. Ray - 5.6.1 allows the previous values to be retained by passing an initializer to new(). Motion: Ray -- approve the proposal for 889 Second: Arturo Oppose: none Abstain: none Passed Unanimously AI/Ray - take out "and shall treat it as a state variable". last paragraph. AI/Mehdi - will update mantis with a new .pdf with the latest fix by Ray. (so that the red/blue shows up). 1382 03-13-06 rand_mode settings after a shallow copy or assignment (Neil/Ray) Ray: shallow-copy, embedded cvrgrp handle is set to null. some name changes in the example. last one some discussion about variables of base class. bullet 2 - null for embedded covergroup name changes were made in the example for clarity last paragraph was added embedded coverage groups are treated differently. Neil: One typo in the last paragraph: in the lasts.. should be in the last. add it to part 2, the note refers to instances: all the handles for global cvrgrp gets copied, and embedded cvrgrp is set to null. Francoise: do we need begin/end for function. Cliff: no we do not need it. Francoise: questions for bullett 2. contradictory: Ray: 2 distincts cvrgrp, normal cvrgrp and embedded cvrgrp (inside class). if you declare a cvrgrp in a package, and then. Gord: the embedded cvrgrp have to have a tighter coupling. Ray: shallow copying does not have any initial construction Gord: if shallow copy, then it is probabley better not use embedded cvrgrp. NOTE: Typos: - baseA - should be in blue - "In the lasts..." - in classB -- A --> baseA - bullet 2 - "internal states uses" --> "internal states used" AI/Ray: update the mantis item proposal to d. DONE. Motion: Ray approve 1382 errata_1382_D.htm Second: Gord Oppose: none Abstain: none Passed Unanimously Miscellaneous --------------------------------- 978 5.15 array methods, minor errata (Shalom/Neil) [see 1270] Neil: updated the mantis item today. Francoise: how does xor work. Dave: the order of operations maybe the problem. Does not say that it is applied to each element, before the reduction Steven: do a reduction items with this propoerty. Arturo: In contrast to array locator method. Dave: Expression in the with clause, apply to every element. Steven: to each element before it is reduced. Dave: the current definition is not what is implemented in Vera. Is this considered a function call, then would it be a state variable. Arturo: whether it imposes partitiong on the constraints. Dave: it does not participate in it? Arturo: it should not partition it. Ray: randomize defines behaviour of function call. Arturo: there is semantics for function call in randomization. Ray: if it is function call then it introduces partitioning. Gord: for reduction and ordering the with is value operator, and for locator it is ordering. like to see locator method, array ordering and reduction in a separate area. Arturo: should write a parapragraph about it. Ray: the 'item' where is that described. Arturo: described above, does not need to be declared. implicit declaraion. 5.15.1 Ray: item is decribed in locator section, AI: Neil update the proposal - 5.15.1 add the new paragraph here, or even earlier. AI: Dave: mantis item for partitiong constraints for array locator methods. 960 one dimension (assigned) (Brad) [also 1447 sv-bc] Sub-clause 5.8 From: "If one dimension of a formal..." To: "If a dimension of a formal..." No discussion was required on this proposal. Move: Dave - Approve the proposal for mantis 960 Second: Arturo Abstain: none Opposed: none passed unanimously 1330 Missing BNF for parameterized virtual interfaces (Dave) [1324 is duplicate -- can be closed if above approved] Cliff - we see green underlined text for the new text Gord - leery of having a parameter list in the middle of a dotted name. - needs to check if they know how to parse this correctly. Dave - parameter list is only an issue for a virtual interface - similar to a specialization of a class Arturo - classes don't have modports so they don't have this complication. - parameterized interfaces are allowed, we want to show the bnf using this for virtual interfaces. Gord - usually simple parameterizations are allowed without the parens. Arturo - parens are optional? - in the older BNF (Steven) only allowed for simple parameters Move the parameter assignment after interface_identifier. There is a trailing .modport_identifier To make it consistent with module or class parameters. Neil - missing example (see AI section of the previous minutes) Gord - library names not really part of name space interfaces and library management issue interface i <---- which one? A particular kind of interface and a particular specialization. Must be the same interface with these parameter value assignments. Arturo - how could it be different? Gord - configurations could possibly effect it. it becomes tough to talk about it Dave - there could be different configs one library - definition another library - instance Gord - the "type" of the interface There is also a module-like behavior of interfaces Arturo - "... or any port connections.." Dave - anything that requires a connection is creating a scope. A specialization of an instance of an interface. An interface definition could contain an interface port. Gord - boils down to knowing enough about virtual interface to be able to use it. A virtual interface shall not itself contain generic interface ports (this is required, not sure if sufficient) Arturo - Virtual generic interface isn't valid Don't want to specialize the interface based on it ports, since you can't reason about it. Dave - you can pull in a type from an interface. you need to know the type of the interface could get part of specialization through a interface port connection don't want a port passed by ref (separate issue - arturo) Gord - don't allow interface ports in virtual interfaces? Arturo - not sure we want to go that far. Doesn't yet have a particular situation in mind. Gord - a virtual interface can't be generic (from Arturo - there is specific text already on this). AI/Gord - new language proposal to handle the port connection issues. AI/Dave - examples via email on the issue with ports in interfaces 1424 process::resume() should be a function (not a task) (Neil) Cliff - same issue in annex C.5 - same changes required there Move: Dave - to approve 1424 with friendly amendment for c.5 Second: Arturo Abstain: none Opposed: none passed unanimously AI/Neil - update the proposal and update with the same changes for c.5 5. Next meetings. July 10th 11-1 July 31st 11-1 Possible: August 14th, 28th Gord - will be out most of July Send input to Mehdi on items that could go out for an email vote. ========== ADDITION TO the minutes, not part of regular ========= ========== meeting discussions, informational purposes ========= ========== Action Items updated based on June 5th ========= Action items: June 26, 2006 ------------------------- AI: 889 Ray take out "and shall treat it as a state variable". last paragraph. AI: 889 Mehdi will update mantis with a new .pdf with the latest fix by Ray. (so that the red/blue shows up). AI: 978 Neil update the proposal - 5.15.1 add the new paragraph here, or even earlier. AI: 978 Dave: mantis item for partitiong constraints for array locator methods. ---> new item 1517 (Enhancement) AI: 1330 Gord - new language proposal to handle the port connection issues. AI: 1330 Dave - examples via email on the issue with ports in interfaces AI: 1424 Neil - update the proposal and update with the same changes for c.5 Action items: June 5th, 2006 ------------------------- AI: 1313 Gord - make an update to item 3. to use "intersection" terminology. AI: 889 Arturo - create a mantis item describing how to handle the null handles for class, to be included in 13.3 Next to last sentence can go away -- blue in proposal. AI: Ray - redo 889 use a different process for updating this mantis item, so that we don't continue to have problems seeing the new text and the strike-outs. rewording: in general dynamic array does a resize with initialization of elements. same effect as x=new (random size ) AI: 1382 Ray - update the proposal with our feedback. add text afterward for the discussion text AI: 978 Neil - update 5.15.3, one sentence - "Array reduction methods can be applied to any unpacked arrays of integral types. Action items: May 22, 2006 ------------------------- AI: 889 Ray - redo 889 proposal Action items: May 8, 2006 ------------------------- AI: 976 Dave - put a proper proposal in place AI: 960 Mehdi - discuss this with Matt to resolve the overlap. Action items: April 10, 2006 ----------------------------- AI: 1330 Dave, add the example list to do. AI: 261/262 combined Steven put together some examples of overflow -- Arturo unsigned/signed examples. Action items: March 13, 2006 ---------------------------- AI: new item - Gordon to create a new mantis item for randomize() method call item (Ray, Gordon have discussed - still working on it internally) ==================================================================================