Subject: RE: [sv-ec] RE: [sv-bc] Minor changes for final LRM
From: David W. Smith (david.smith@synopsys.com)
Date: Thu May 15 2003 - 08:04:59 PDT
Jay,
Thanks for the clarification and the willingness to pursue the issue. Based
on the change in semantic I have removed LRM-359 from the list of changes.
Regards
David
David W. Smith
Synopsys Scientist
Synopsys, Inc.
Synopsys Technology Park
2025 NW Cornelius Pass Road
Hillsboro, OR 97124
Voice: 503.547.6467
Main: 503.547.6000
FAX: 503.547.6906
Email: david.smith@synopsys.com
<http://www.synopsys.com/> http://www.synopsys.com
-----Original Message-----
From: owner-sv-ec@eda.org [mailto:owner-sv-ec@eda.org] On Behalf Of Jay
Lawrence
Sent: Wednesday, May 14, 2003 4:21 AM
To: Arturo Salz; David W. Smith; sv-ac@eda.org; sv-bc@eda.org; sv-cc;
sv-ec@eda.org
Subject: RE: [sv-ec] RE: [sv-bc] Minor changes for final LRM
Because of the fact that the object can't be used prior to being declared in
the example you give below, this change is just not needed.
You also need to consider scopes that are not sequential like modules. Are
we now required to initialize objects in them in the order of declaration as
well?
This revision is for obvious things like typos, BNF mismatches etc. I
believe this change adds a new semantic requirement.
Jay
===================================
Jay Lawrence
Senior Architect
Functional Verification
Cadence Design Systems, Inc.
(978) 262-6294
lawrence@cadence.com
===================================
-----Original Message-----
From: Arturo Salz [mailto:Arturo.Salz@synopsys.com]
Sent: Tuesday, May 13, 2003 9:16 PM
To: Jay Lawrence; David W. Smith; sv-ac@eda.org; sv-bc@eda.org; sv-cc;
sv-ec@eda.org
Subject: Re: [sv-ec] RE: [sv-bc] Minor changes for final LRM
Jay,
That sentence only applies to initial values within a scope, and it simply
extends to static initialization the same semantics as automatic variables.
What you say is true: initial values are allowed to contain hierarchical
names, so a complete ordering of elaboration and assignments of initial
values across scopes is needed, but that process is not necessary for local
variables.
Furthermore, That sentence allows code like the following:
task x;
int a = b + 1; // b = ?
int b = 5;
...
endtask
to be treated as an error, since b is used before being defined. The only
exception are hierarchical expressions.
However, the declaration:
task x;
int b = 5;
int a = b + 1; // a = 6
...
endtask
will yield the expected results whether the task is static or automatic. The
intent was not to add anything new but to clarify that static and automatic
declarations have the same semantics.
Arturo
----- Original Message -----
From: Jay <mailto:lawrence@cadence.com> Lawrence
To: David W. Smith <mailto:david.smith@synopsys.COM> ; sv-ac@eda.org ;
sv-bc@eda.org ; sv-cc <mailto:sv-cc@eda.org> ; sv-ec@eda.org
Sent: Tuesday, May 13, 2003 5:00 PM
Subject: [sv-ec] RE: [sv-bc] Minor changes for final LRM
David,
I have reviewed the LRM changes here and have 1 question (which may become
an objection) and 1 definite objection.
First the question, did the sv-ac review and approve the addition of this
appendix G as part of change LRM-356?
I'm aware that they had been discussing John's definition here, but it is a
major addition to the LRM that I saw no vote on it for inclusion, there has
certainly been no debate on the reflector for the last 2 weeks (unless I've
been removed from sv-ac). Erich Marschner is our usual sv-ac representative
but he is traveling in Japan and somewhat out of reach. Unless there was a
specific vote to include this major addition, I would ask that it be
removed.
Secondly, I object to the change of functionality in LRM-359. The addition
of the clause "Within a scope, initial values are applied in the order of
declaration". Is not a clarification, it adds a new requirement on
implementation and is not complete. Initial values are allowed to contain
hierarchical names, therefore specifying the ordering within a scope is not
sufficient. A complete ordering of elaboration and assignment of initial
values across all scopes (including parameters, localparams and defparams)
would need to be done to make this initialization deterministic. This
addition is a hack that adds no determinism.
Jay
===================================
Jay Lawrence
Senior Architect
Functional Verification
Cadence Design Systems, Inc.
(978) 262-6294
lawrence@cadence.com
===================================
-----Original Message-----
From: David W. Smith [mailto:david.smith@synopsys.com]
Sent: Tuesday, May 13, 2003 7:33 PM
To: sv-ac@eda.org; sv-bc@eda.org; 'sv-cc'; sv-ec@eda.org
Subject: [sv-bc] Minor changes for final LRM
Greetings,
I requested from the chairs and SV champions any changes they felt were
required to the Draft 6 LRM before it is released. These changes are meant
to be corrections to what was provided or correct missing items.
I have just posted to the SV-EC web site
(http://www.eda.org/sv-ec/Draft_6_Review/LRM_Issues.html) the current set of
changes. These changes (and any others that are deemed by the Chairs to be
appropriate) will be added to the final draft of the LRM.
Regards
David
David W. Smith
Synopsys Scientist
Synopsys, Inc.
Synopsys Technology Park
2025 NW Cornelius Pass Road
Hillsboro, OR 97124
Voice: 503.547.6467
Main: 503.547.6000
FAX: 503.547.6906
Email: david.smith@synopsys.com
<http://www.synopsys.com/> http://www.synopsys.com
This archive was generated by hypermail 2b28 : Thu May 15 2003 - 08:11:21 PDT