[sv-cc] SV-CC Meeting minutes for 11/03/2004

Minutes of 11/03/2004 SV-CC Meeting.

ATTENDEES
000000000
444444444
111110000
100009999
022102211
360369250
xxxxxxxxx Charles Dawson
xxxxxxxxx Francoise Martinolle
xxxxxxxxx Doug Warmke
xxxxx-xxx Bassam Tabbara
xxxxxxxxx Andrzej Litwiniuk
xxxxxxxxx Joao Geada
xxxxx-xxx Jim Vellenga
xxxxxxxx- Ralph Duncan
xxxx--x-- Rob Slater
xxxxx-x-- Sachchidananda Patel
xxxxxxx-- Michael Rohleder
xxx-xxx-- John Stickley
xxxxxx-x- Jim Garnett
xxxxx--x- Steven Dovich
xx-x-xxxx Ghassan Khoory
--------x Swapnajit Mitra
--------x Karen Pieper
------x-- Angshuman Saha
--x------ Kevin Cameron
x-------- Amit Kohli

1. Reviewed Patent information.

   - Charles Dawson read the patent information.

2. Reviewed minutes of the 10/26/04 Meeting.

   - Ghassan reported that he was at the last meeting and the minutes
     did not reflect that.
   - ?/?. Minutes accepted as amended.

3. Liaisons

   - Francoise reported that the SV-BC has accepted a 1364 change
     that will result in some expressions not having a value. This will
     require a chance to vpi_get_value(). Francoise took an action
     to find the Item number.
   - No other meetings were reported on.

4. New business

   - There were no further objections to Items 044, 078, 156, and 198,
     so they are now considered PASSED.
   - Chas brought up that PTF 342 has had a proposal for several months
     now. JimV and Stu took actions to make sure we do not delete
     parts which are important. Since Stu is not attending our meeting,
     it seems unlikely he will be able to do this. JimV was unsure he
     would be able to do this. Chas took an action to re-verify that
     the cross references in the remainder of the specification have been
     properly fixed. Steve suggested that the sections which are deleted
     have a reference to the prior version of the spec. Chas to add this
     to the proposal.
   - No new business was brought up.

5. Reviewed of items with proposals.

   - Item 205: Binary compatibility for packed arrays as fields and as elements of unpacked arrays

     This item was discussed further. In particular, Francoise wanted
     answers to the questions she posed via email. ?/?.

     We voted on this as follows:

       For: Francoise, Doug, Bassam, Ralph, JimG, Rob, Sachi, Michael, John
       Against: Andrzej, Sachi, Ghassan
       Abstain: JimV, Joao, Steven

     Proposal PASSED.

   - Item 123: Clarify meaning of "member typespec" in VPI

     JimV commented that there was a problem he encountered while we talked about this
     last week. He fixed it shortly after the meeting. He also realized that Item 59
     (as well as the previously agreed upon 121, and 122) was a duplicate. Any
     objections to marking 59 as a duplicate? No.
     ?/? PASSED (unanimous).

   - Item 050: Change DPI svLogicVec32 representation to match PLI/VPI aval/bval representation

     Doug explained that the issue was that DPI values were not compatible with
     VPI values. Andrzej thought that this could have been corrected a few years
     ago, but Synopsys has customers now. Chas pointed out that there are many
     more VPI applications out there, and to make the two interfaces usable together
     the application developer would have to do data marshaling. Joao thought
     that the main point behind DPI was performance, that DPI gives back pointers
     to the simulator's version of the data while VPI makes a copy. Joao argued
     that VCS would have to make a copy if this passes for DPI. ?/?

     We voted on this as follows:

       For: Francoise, Doug, Bassam, Ralph, JimG, JimV, Steven, Michael, John
       Against: Andrzej (Strongly NO. Votum separatum!), Joao, Sachi, Ghassan
       Abstain: Rob

     Proposal PASSED.

   - Item 199: More detail needed for DPI treatment of dynamic array arguments

     Discussed. ?/? PASSED (unanimous)

   - Item 200: Clarifications needed in DPI Annex E.6, "Data Types"

     We determined that Item 205 needs to be amended because the section number
     conflicts with the number proposed here. This amendment was accepted (Ralph
     to implement). ?/? PASSED (unanimous)

Meeting ended at 1:05pm (EDT)

6. Old Business

   -

7. Action items

   SV-CC action items:
   - Chas to assign remaining Items to those without open ones now.
   - Chas to get the database updated to reflect the previous meetings.
   - Francoise to ask Peter Ashenden what was done to improve
     printing from Rational Rose.
   - Francoise to inquire about the feasibility of third parties
     shipping the UML for the diagrams.
   - JimV to resubmit a proposal for Item 123.
   - Joao/Francoise to file SV-BC item asking to define linearization.
   - Francoise to check with SV-BC on default return type of functions.
   - Chas to ask Karen about updating the diagrams (does not fit well
     with approved process).
   - Andrzej to make sure the LRM says that for the C layer of DPI,
     representations of a type are always the same regardless of where
     it is (packed struct, member of array, ...etc.).
   - Francoise and Bassam to reconcile sections 28 and 31.

   PTF action items:
   - Steve to compare BNF with the access available
     for attributes to see if they match
   - Francoise to remove "+" from tags in UML diagrams and
     add vpi prefix where appropriate.
   - Francoise to send out HTML for 1364-2001 diagrams, using
     something other than JPG for importing diagrams into frame.
   - Stu to write proposal for PTF 368.
   - Francoise to write proposals for PTF 373, 374, and 396.
   - Steve to write proposals for PTF 311, and 495.
   - Sachi to write proposals for PTF 307, 312, and 313.
   - All to review Generates proposal from ETF committee.
   - Francoise, et all to review BTF generates proposal
     for the upcoming vote, with particular emphasis on
     how we will address generates in VPI.
   - Stu to enter new PTF item for save/restart/reset issue.
   - JimG to write proposals for PTF 517, 533, and 534.
   - Chas to write proposal for PTF 296.
   - Stu to write an addition to the proposal for PTF 342.
     This will cover that PLI 1.0 was deprecated in section 20
     and include some of the stuff currently in section 21
     (like the descriptions for the checktf and calltf).
   - Francoise to lookup wording for PTF 524 in VHPI.
   - JimV to try to rework proposal for PTF 530 to address other
     issues we found in 26.6.17.
   - Francoise will open a new PTF issue to look for situations like 25.6.15,
     where multiple methods are used access the same object enclosure
   - Chas to reword proposal for PTF 525.
   - Draft a straw man proposal using a clean slate with no concern for
     existing PLI/VPI on the best way to represent all Verilog and
     SystemVerilog kinds and types. This straw man will then be used as a
     basis for discussing backward compatibility with the existing reg, net,
     variables, functions, and parameter diagrams. It may be decided that
     full backward compatibility is not possible, or is not the best approach
     moving forward.
   - Sachi will file a PTF item for the clarification of what can be done
     at ROsync time and putting values in future times.
   - Francoise to file a PTF item that asks to specify the order that iteration
     occur in, when the order is important.
   - Steve to add ETF item for Annex C to remove the Informative label, but
     still allow the contents to be optional.
   - Chas to re-verify the cross references are properly updated for PTF 342.
   - Chas to add to proposal for PTF 342 a reference to the prior version of the
     specification

8. Items for consideration at the next meeting (they already have proposals):

   - Item 201: More details needed on DPI string argument handling
   - Item 277: Spelling error in Table 31-4 "Instance" not "Instances"
   - Item 060: bad cross-reference in notes 3 and 4 of diagram 31.10
   - Item 274: Small 2-state type svBitVec32 provokes C coding difficulties
   - Item 050: Change DPI svLogicVec32 representation to match PLI/VPI aval/bval representation
   - Item 156: Jeita 31: In Index, issue with blocking and DPI imports
   - Item 199: More detail needed for DPI treatment of dynamic array arguments
   - Item 200: Clarifications needed in DPI Annex E.6, "Data Types"
   - Item 201: More details needed on DPI string argument handling
   - Item 274: Small 2-state type svBitVec32 provokes C coding difficulties

--
Charles Dawson
Senior Engineering Manager
NC-Verilog Team
Cadence Design Systems, Inc.
270 Billerica Road
Chelmsford, MA  01824
(978) 262 - 6273
chas@cadence.com