SV-ACMinutes2011_03_22 < P1800

P1800 Web>SystemVerilogAssertionCommittee>SVACMeetingMinutes>SV-ACMinutes2011_03_22 (2011-03-25, ErikSeligman) (raw view)EditAttach
Minutes of SV-AC Meeting

Date: 2011-03-22

Time: 16:00 UTC (9:00 PDT)

Duration: 1.5 hours

Dial-in information:

--------------------

Meeting ID:            38198

Phone Number(s):

1-888-813-5316         Toll Free within North America

Live Meeting: [[https://webjoin.intel.com/?passcode=6362971]]

Agenda:

-------

- Reminder of IEEE patent policy.

See: http://standards.ieee.org/board/pat/pat-slideset.ppt

- Minutes approval

- Email ballot results

2476: Need clarification about system functions $onehot, etc

2804: Need to clarify rule (b) in 16.15.6 to allow inferred clock when

expression appears in procedural assertion.

Both issues passed: 11y/0n/0a. There were friendly amendments.

- New issues

- Issue resolution/discussion

3377: Fix method names in C.2.3 ended sequence method

- Enhancement progress update

3295: need a way to control only asserts/covers/assume directives

3191: Allow sequence methods with sequence expressions

3069: Relax rules for $global_clock resolution

3213: Update definition of sampled value

3195: Local Variables Flow Out Issue in and/or/intersect/implies

Attendance Record:

------------------

         Legend:

                 x = attended

                 - = missed

                 r = represented

                 . = not yet a member

                 v = valid voter (2 out of last 3 or 3/4 overall)

                 n = not a valid voter

                 t = chair eligible to vote only to make or break a tie

Attendance re-initialized on 2010-07-06:

  v[-xxx-xxx...........................] Ashok Bhatt (Cadence)

  v[xxxxxxxx-xxx-xxxxxxxxx-x-xxxxx--xxx] Laurence Bisht (Intel)

  v[xxxxxxx-xxxxxxxxxxxx-xxxxxxxxxxxxx-] Eduard Cerny (Synopsys)

  n[---xx---xxx--x-xxxxxxx-xxxxx-xxxxxx] Ben Cohen

  n[----------------xx-x-xxx-x--xxxxxxx] Surrendra Dudani (Synopsys)

  v[xx-x-x----x-x-x--xx---xxxx---x-xxxx] Dana Fisman (Synopsys)

  n[--------------xxxxx-xxxx-x-xxxxxxxx] John Havlicek (Freescale)

  v[x-xxxxxxxxxxxxxxxx-xxx-xxxxxxxxxxxx] Tapan Kapoor (Cadence)

  v[x-x-x..............................] Jacob Katz (Intel)

  t[xxxxxxxxxxxxxxxxxx--xxxxxxxxxxxxxxx] Dmitry Korchemny (Intel ¿ Chair)

  v[xx-xxxxxxxxxxxxxxx--xxxxxx-xxxxxxxx] Scott Little (Freescale)

  v[xxxxxxxxxxxxxxxxx-xxxxxxxxx-xxxxxxx] Manisha Kulshrestha (Mentor

Graphics)

  v[xxxxxxxxxxxxxxx-xxxxxxxxxxxxxxxxxxx] Anupam Prabhakar (Mentor Graphics)

  v[x-xxx-xxx--x-xx-xxx-xx--xxxxxxx-xxx] Erik Seligman (Intel)

  v[xx-x-xxx-xxxx-xxxx--xxxxxx-xxxxxxx.] Samik Sengupta (Synopsys)

  v[xxxxxxxxxxxxxxxxx-xxxxxxxxxxxxx-xxx] Tom Thatcher (Oracle ¿ Co-Chair)

  n[x---xx-------x.....................] Srini Venkataramanan (CVC Pvt

Ltd)

    |- attendance on 2011-03-22

  |--- voting eligibility on 2011-03-22

Minutes

-------

- Reminder of IEEE patent policy.

See: http://standards.ieee.org/board/pat/pat-slideset.ppt

Participants were reminded of the IEEE patent policy.

- Minutes approval

Erik: Move to approve minutes

Scott:     Second

     Vote results: 10y, 0n, 0a

Dmitry:    Srini will attend as an Accellera representative

     Still trying to get approval for Ben

- Email ballot results

2476: Need clarification about system functions $onehot, etc

Erik: Question about Ed's friendly amendment

Ed:  You need an extra set of curly braces in the text.

(Manisha joined)

Tom: Cross-out text was missing in the proposal.

Dmitry:    Will call for another e-mail ballot to accept these changes.

2804: Need to clarify rule (b) in 16.15.6 to allow inferred clock when

expression appears in procedural assertion.

Erik: Ben had commented that there was an argument mis-match in an example,

    but it looks correct.

Erik Move to accept amendments to 2804

Scott Second

     Vote results 11y, 0n, 0a

- Issue resolution/discussion

3377: Fix method names in C.2.3 ended sequence method

Erik: Move to approve proposal

Lawrence: Second

     Vote results: 11y, 0n, 0a

- Enhancement progress update

3295: need a way to control only asserts/covers/assume directives

Manisha: Modified BNF to keep only one production

Dmitry:    Terminal productions should be in red.

Manisha:   Added lock and unlock types.

     Added editorial changes.

Scott:     Should you discuss control types in numerical order?

Manisha:   Made lock and unlock types 0 and 1 so that they wouldn't

     be in the middle of the list if we add more control types.

Lawrence: Clarify that even though an assertion may have been locked twice,

    a single unlock will unlock it.

Scott:     I think it's clear as is that a second lock for an assertion should

     not affect anything.

Dmitry:    In examples, should typeset keywords in bold.

Srini:     User will get confused with all these numbers

     Could we allow named argument connection

     It's not currently allowed for PLI calls.

     We would have to consult SV-BC for this.

Manisha:   Asked about defining pre-defined names or enum's

     But whatever we add might conflict with existing code

     Thought about adding this to standard package, but this had the

     same problem. Importing the standard package might cause conflicts

     with existing code.

     The User is always free to define their own ENUMs

Dmitry:    We should de-couple this issue from the definition of the function

     itself. We could file an enhancement with the SV-BC.

Srini:     We should make this easier for user to use.

Anupam:    You should use `defines in these examples to make them clearer

Dmitry:    Should use let instead

Tom: Should the new $assertControl have control types to cover the

     assertions in 20.12.

Manisha:   What about severity? Nobody commented about it.

     If we don't add it now, we might need to add another control task

     later.

Dmitry: Add an extra argument now for future use?

Manisha:   Since we don't know how severity might be defined in the

     future, it might be difficult to know how to define this

     placeholder argument.

Dmitry:    Then probably we'll just have to define a new function when the

     notion of severity is defined.

3191: Allow sequence methods with sequence expressions

Jacob:     Has uploaded a new proposal

Anupam:    One correction still needs to be made

Dmitry:    When Jacob uploads the new proposal, will call for a vote.

Continuous assignments in checkers

Dmitry:    Perhaps we could restrict free variables so that free variables

may only appear on RHS of an assignment to another free variable.

These assignments would occur in the Observed region.

     bit a

     rand bit v

     rand bit w

     assign a = v; // Illegal

     assign w = v; // Observed region

     assert property (w == v) would then pass

Tom: Why would this be needed? You could use assume property instead to

     get the same effect.

     assume property (w == v) would ensure that v and w are equal.

Dmitry:    Less efficient

Tom: Two possible approaches:

     1. Prohibit free variables from appearing on RHS of continuous

        assignments. This would allow for future definitions of

        semantics without causing backward-incompatibility.

     2. Allow free variable on RHS of a continuous assignment.

        But it should be clear that the simulation semantics are followed.

        And you won't get expected results if you combine different

        variable types in novel ways.

Meeting adjourned.
Topic revision: r1 - 2011-03-25 - 21:12:34 - ErikSeligman
P1800