Subject: Verilog++ 16th Committee Meeting 28 Jan 2002
From: Tom Fitzpatrick (fitz@co-design.com)
Date: Mon Jan 28 2002 - 10:30:44 PST
Verilog++ 16th Committee Meeting
January 28th, 2002
Vassilios leading. TomF taking minutes. Action Items highlighted.
Next Conference Call: 2/25/02
9am-11am PST
(405)244-5555 x3715
(aaaaa-aaaaaaaaaa) Vassilios Gerousis *
(raaa-aaarar-aaaa) Dave Kelf
(------aaa--aaaa=) John Sanguinetti
(aaaaaaa--a-aa---) Dennis Brophy *
(aaaaaaaaaaaaaaaa) Stu Sutherland *
(-a--------aaaaa=) David Knapp *
(--aaaaaraaar-aaa) Tom Fitzpatrick
(a-arraa-aaaaaaa=) Phil Moorby
(aa-aaaa-aaaaaaaa) Anders Nordstrom *
(aaaaa--a-aaaaaa=) Cliff Cummings *
(aaaaaraaa-aaaaaa) Simon Davidmann *
(aaaaaaaaaaaa====) Peter Flake *
(-aaaa-aaaaa-aaaa) Stefen Boyd *
(aaaaaaaaaaaaaaaa) David Smith *
(aa-aaa--a-a--aa=) Mike McNamara *
(aaaaaaaaa=======) Kevin Cameron *
(aa-a-aa-a=======) Andy Tsay *
(a-aa============) Alec Stanculescu
(--a=============) Adam Krolnik
(aa==============) Paul Graham *
(-a==============) David Seieert *
Minutes of last meeting were quickly reviewed and accepted.
Stu still needs to update the BNF.
$root discussion held off until Stefen is available.
State Machines
Cliff made his case against state machine syntax.
Co-Design: The State Machine syntax is aimed at a higher level of
abstraction than the low-level implementation issues raised. Need to come up
with a way of specifying the state encoding.
Cliff has not heard back from Kurt Baty for an example of hierarchical FSMs.
Stu believes that state machines are a lower priority which can be postponed
to the next revision.
Anders pointed out that Cliff's examples were not strictly apples-to-apples.
Cliff claimed that this was because it was difficult to replicate common
coding styles using the SystemVerilog state machine construct.
Anders and Cliff feel that the transition-based syntax has some advantages.
Cliff would like Co-Design to respond directly to the examples he sent out.
If Co-Design cannot respond by next meeting, state machines may get
postponed to the next revision.
Email reflector can be found at www.eda.org/vlog-pp/hm
Interfaces: Cliff sent out some examples. Peter will reply to Cliff's
examples and this response will be sent to the whole group.
Cliff suggested using the ".*" notation for connecting interfaces as well as
modules. More useful at the "architectural" level at the top level. Best
used when port names and net names match. Anders pointed out that ".*" loses
the documentation value of seeing the port connections.
Cliff points out that iHDL (Intel) has the ".*" capability.
Review of other Open Issues:
What's left to do on Interfaces?
Draft4/BNF updates (stu)
Cliff owes list of suggestions for deprecation
Recommendation back from Assertion Committee
State Machines
".*" (for modules, but has impact on interfaces)
Combining reg/wire type - Cliff to provide better feedback.
Inferred declarations
Interfaces:
Cliff suggested changes to clarify examples in the document (sent from
Vassilios on 12/17/01 -
http://www.eda.org/vlog-pp/hm/att-0212/01-InterfaceSection_Proposals.pdf)
These suggestions will be considered separate from the ".*" issue.
Changes are cosmetic, not technical. Cliff will make sure that suggested
changes will compile.
Stu will incorporate wording changes into Draft4, unless a specific
objection is raised by 2/1/02.
Stu: Draft4 will be the HDLCon draft. Will be available on 2/15/02 with
updated BNF based on 1364-2001 BNF. Will be reviewed at 2/25/02 meeting.
If Assertions committee does not approve assertions in time for the final
draft of SystemVerilog to be approved by Accellera board, assertions will be
removed from this version and added to the next revision.
Cliff to submit deprecated item list by 2/15/02 for review at 2/25 meeting.
Deprecated items are suggestions that these will not be supported in a
future version. Recommendation that tools should warn users if such items
are encountered.
Stu to send Cliff deprecation wording from HTML spec.
Peter/TomF to review Cliff's proposals and respond by 2/22.
State Machines
.*
Cliff to send out combined reg/wire type and inferred declaration proposals
by 2/25.
Simon suggested Vassilios and Dennis need to determine voting rules
Cliff proposed that anyone who attended 3 of last 4 or 75% of meetings
overall be eligible to vote.
Simon seconded.
Approved unanimously.
Voting prodedure ammended to allow proxy representation.
Proposed - Mike McNamara
Seconded - Cliff
Approved unanimously
This archive was generated by hypermail 2b28 : Mon Jan 28 2002 - 10:30:40 PST