SV-AC_Minutes_2011_08_30 < P1800

P1800 Web>SystemVerilogAssertionCommittee>SVACMeetingMinutes>SV-AC_Minutes_2011_08_30 (2011-09-13, ErikSeligman) (raw view)EditAttach
Minutes of SV-AC Meeting <br />Date: 2011-08-30 <br />Time: 16:00 UTC (9:00 PDT) <br />Duration: 2 hours

Dial-in information: <br />-------------------- <br />Meeting ID: 38198

Phone Number(s): <br />1-888-813-5316 Toll Free within North America

Live Meeting: <br /> [[https://webjoin.intel.com/?passcode=6023798]]

Agenda: <br />-------

- Reminder of IEEE patent policy. <br />See: http://standards.ieee.org/board/pat/pat-slideset.ppt

- Minutes approval

- Email ballot results: <br />Issue 2476 passed with friendly amendments: 8y/0n/0a. <br />Issue 2578 passed with friendly amendments: 7y/0n/0a. <br />Issue 3033 failed: 4y/1n/0a. <br />Issue 3133 passed with friendly amendments: 8y/0n/0a. <br />Issue 3206 failed: 7y/1n/0a. <br />Issue 3385 failed: 7y/1n/0a.

- New issues

- Champions' feedback <br />3069 Relax rules for $global_clock resolution

- Issue resolution/discussion <br />2093: Checker construct should permit output arguments

- Opens

Attendance Record: <br />------------------ <br />          Legend: <br />                  x = attended <br />                  - = missed <br />                  r = represented <br />                  . = not yet a member <br />                  v = valid voter (2 out of last 3 or 3/4 overall) <br />                  n = not a valid voter <br />                  t = chair eligible to vote only to make or break a tie

Attendance re-initialized on 2010-07-06:

   v[xxx-xxx--xxx---xxxx-xxx-xxx...........................] Ashok Bhatt <br />(Cadence) <br />   v[--xx--xxx-xxxxxx-xxxxxxxxxx-xxx-xxxxxxxxx-x-xxxxx--xxx] Laurence <br />Bisht (Intel) <br />   v[xxxxxxxxxxxxxxxx-xxxxxxxxx-xxxxxxxxxxxx-xxxxxxxxxxxxx-] Eduard <br />Cerny (Synopsys) <br />   v[xx-xxxxxxxxxxx--------xx---xxx--x-xxxxxxx-xxxxx-xxxxxx] Ben Cohen <br />(Accellera) <br />   n[-----------------------------------xx-x-xxx-x--xxxxxxx] Surrendra <br />Dudani (Synopsys) <br />   n[------x-xxxxxx........................................] Shaun Feng <br />(Freescale) <br />   n[--------x-x--x-x-xxxx-x-x----x-x-x--xx---xxxx---x-xxxx] Dana Fisman <br />(Synopsys) <br />   n[---------------------------------xxxxx-xxxx-x-xxxxxxxx] John <br />Havlicek (Freescale) <br />   v[xxxx---x-xxx-xx-xxxx-xxxxxxxxxxxxxxxx-xxx-xxxxxxxxxxxx] Tapan <br />Kapoor (Cadence) <br />   v[xx---x-xx-xxxxx-xxxx-x-x..............................] Jacob Katz <br />(Intel) <br />   t[xxxxxxxx--xxx-xxxxxxxxxxxxxxxxxxxxxxx--xxxxxxxxxxxxxxx] Dmitry <br />Korchemny (Intel � Chair) <br />   v[xxxxxxxxxx-xxxxx-xxxx-xxxxxxxxxxxxxxx--xxxxxx-xxxxxxxx] Scott <br />Little (Freescale) <br />   v[x-xxxxxxxxxxxxxxxx-xxxxxxxxxxxxxxxxx-xxxxxxxxx-xxxxxxx] Manisha <br />Kulshrestha (Mentor Graphics) <br />   v[xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx-xxxxxxxxxxxxxxxxxxx] Anupam <br />Prabhakar (Mentor Graphics) <br />   v[-xxxxxxxxx--xxx-xxxx-xxx-xxx--x-xx-xxx-xx--xxxxxxx-xxx] Erik <br />Seligman (Intel) <br />   v[xxxxx-xxxx-xxxxxxxxxx-x-xxx-xxxx-xxxx--xxxxxx-xxxxxxx.] Samik <br />Sengupta (Synopsys) <br />   v[xxxx-xxxxxxxxxxxxx-xxxxxxxxxxxxxxxxx-xxxxxxxxxxxxx-xxx] Tom <br />Thatcher (Oracle � Co-Chair) <br />   n[------------------xx---xx-------x.....................] Srini <br />Venkataramanan (CVC Pvt Ltd) <br />   n[--------x.............................................] Vibarajan <br />Viswanathan (Marvell) <br />     |- attendance on 2011-08-30 <br />   |--- voting eligibility on 2011-08-30

Minutes <br />------- <br />- Reminder of IEEE patent policy. <br />See: http://standards.ieee.org/board/pat/pat-slideset.ppt

- Minutes approval <br />Scott: Move to approve minutes <br />Ben: Second <br />        Vote results: 10y, 0n, 0a

- Email ballot results: <br />Issue 2476 passed with friendly amendments: 8y/0n/0a. <br />Scott generated a new PDF with no change tracking. <br />Scott: Move to approve proposal for 2476 <br />Tom: Second <br />        Vote results 10y, 0n, 0a

Issue 2578 passed with friendly amendments: 7y/0n/0a. <br />Ben: Added the "and" <br />Tom: Move to approve proposal for 2578 <br />Scott: Second <br />        Vote results: 10y, 0n, 0a

Issue 3113 passed with friendly amendments: 8y/0n/0a. <br />Ben: Suggested moving comment lines to a separate line so they don't <br />        wrap around. <br />Dmitry: Moved comments to line before the related line. <br />Tom Move to approve proposal for 3113 <br />Scott: Second <br />        Vote results: 10y, 0n, 0a

Issue 3206 failed: 7y/1n/0a. <br />Scott: Erik had changed "at the instant" to "at the time" Wants to <br />know reason for <br />        the change. Will get feedback from Erik.

Issue 3385 failed: 7y/1n/0a. <br />Scott: Felt that some wording was too loose. Will get feedback from Erik.

Issue 3033 failed: 4y/1n/0a. <br />Anupam: There's nothing in the proposal which prevents checker from <br />assigning <br />        to checker arguments. Since checker arguments are substituted, an <br />        assignment to a checker formal argument is an assignment to the actual <br />        argument. This could be non-intuitive to have a checker modify the <br />        signal connected to it. <br />Manisha: Her comments were addressed in updated proposal

Dmitry: Will update proposal and call for e-mail ballot.

- Champions' feedback <br />3069 Relax rules for $global_clock resolution

Dmitry: Got no feedback from SV-BC. Will wait one more week. <br />Scott: SV-EC and SV-BC met yesterday: SV-EC did not discuss this topic.

- Issue resolution/discussion <br />2093: Checker construct should permit output arguments

Jakob: Reviewed proposal <br />Anupam: Proposal has two semantics for connecting output port: one by <br />substitutions, <br />             one by continuous assignment.

Anupam: Substitutions only works if type of output is the same as actual. <br />        If not same, would need a continuous assignment. <br />        Need a restriction that type of argument must match? <br />Jakob: In rest of language, there's no such restriction. <br />Dmitry: In properties, an output argument is allowed for a local variable. <br />        Tried to match the semantics of the property output argument. <br />Jakob: For untyped output, the type of the formal should be inferred <br />from the actual <br />Manisha: How about a net connected to checker output. Continuous <br />assignment? <br />Scott: Checker ports: would they follow net port connection rules or <br />var port? <br />        Modules: ports are net ports by default. have to declare a var port. <br />        Do checkers have var ports by default? <br />Manisha What is difference between var ports and net ports <br />Scott: See 23.3.3 <br />Jakob: In either case, it says there is a cont assign between formal and <br />actual. <br />Scott: Rules in 23.3.2 would apply for var ports <br />Anupam: Does it say in the proposal that checker output ports are <br />variable ports or <br />        wire ports. <br />Dmitry: Since wires are not allowed in checkers, checker ports would be <br />var ports. <br />Scott: Don't see advantage for substitution semantics for output ports. <br />        Current rules for output ports are defined. <br />Anupam: Fine with either semantic, as long as there is only one. <br />Jakob: With cont assign semantic, will need specific statement saying <br />formal port <br />        type is inferred from actual (for untyped formal). <br />Jakob; One option is to require output ports to be typed. <br />        Why would we need untyped checker outputs? <br />Ed: This would be the same as properties: Local variable outputs in <br />properties <br />        must have a type.

Dmitry will create modified proposal. Anupam, Jakob, and Tom will review.

if-else property. <br />Anupam: Semicolon at the end is not required. Kind of strange. <br />        It's legal either with or without an ending semicolon. <br />        It's the only property expression that doesn't require a semicolon at <br />        the end. <br />        Couldn't come up with a way to specify it that was backward compatible. <br />        Is anyone <br />Tom: Does it cause any compiler issue?

        If nobody cares, we will drop the issue.

Meeting adjourned.
Topic revision: r1 - 2011-09-13 - 19:50:39 - ErikSeligman
P1800