RE: TLM-2.0 compliance

All,

See clause 15.2.3 (versus 15.2.1 and 15.2.1). I want to say that a model
that uses a user-defined protocol type but does not use
tlm_generic_payload is neither TLM-2.0 base-protocol-compliant nor TLM-2.0
custom-protocol-compliant.

Similarly, a model that uses the tlm_generic_payload but does not use the
standard sockets (or derivatives) would not be compliant in either sense.

In other words, legal SystemC models that use isolated features from the
TLM-2.0 class library would be neither TLM-2.0 base-protocol-compliant nor
TLM-2.0 custom-protocol-compliant. To be compliant in either sense a model
must use all of the main features of the interoperability layer, not just
pick-and-choose.

Does that work for everybody?

Thanks

John A

From:
Stuart Swan <stuart@cadence.com>
To:
"john.aynsley@doulos.com" <john.aynsley@doulos.com>, "Jeremiassen, Tor"
<tor@ti.com>
Cc:
"systemc-p1666-technical@eda.org" <systemc-p1666-technical@eda.org>
Date:
05/10/2010 21:36
Subject:
RE: TLM-2.0 compliance

John-
 
I fully agree with all of your comments below, in particular that these
terms need to be included in the LRM.
 
Thanks
Stuart
 
From: owner-systemc-p1666-technical@eda.org [
mailto:owner-systemc-p1666-technical@eda.org] On Behalf Of
john.aynsley@doulos.com
Sent: Tuesday, October 05, 2010 3:36 AM
To: Jeremiassen, Tor
Cc: systemc-p1666-technical@eda.org
Subject: RE: TLM-2.0 compliance
 
Tor, All,

"Implementation" and "application" are defined in 1666, where they refer
to "the SystemC class library". Hence this clause (3.1.2) does indeed need
to be tweaked. Now that SystemC and TLM-2.0 are both part of P1666, I
guess the obvious step would be to expand the definitions of
"implementation" and "application" to simple roll in TLM-2.0 with SystemC
such that they become parts of a single whole.

Does anyone wish to argue that we should distinguish between a SystemC
implementation and a TLM-2.0 implementation? I guess that the pragmatic
move would be to say that the unqualified terms cover both, and that
qualification can be used where necessary (i.e. "SystemC implementation",
"TLM-2.0 implementation"). The only place I think this impacts is the
version number macros/API - a current debate.

I fully agree that sockets of types derived from the standard sockets are
allowable. I was just being lazy with my wording. That was what I meant to
imply by prefixing the definitions with "informal" ;-)

John A

From:
"Jeremiassen, Tor" <tor@ti.com>
To:
"john.aynsley@doulos.com" <john.aynsley@doulos.com>,
"systemc-p1666-technical@eda.org" <systemc-p1666-technical@eda.org>
Date:
05/10/2010 11:23
Subject:
RE: TLM-2.0 compliance
 

I think first of all we need to make sure that we use the same term
(implementation vs application) in all the definitions and ensure that it
is properly defined.
  
Second, the ?custom protocol compliant? needs refinement as it is
perfectly allowable to derive new sockets from the standard socket
classes, and the current wording seems to exclude that possibility. In
fact, derived socket types should be allowable for all of these compliance
categories.
  
Tor
  

---
Tor Jeremiassen, Ph.D.
Simulation and Modeling CTO
SDO Foundational Tools
Texas Instruments                    Ph:    281 274 3483
P.O. Box 1443, MS 730                Fax:   281 274 2703
Houston, TX 77251-1443               Email: tor@ti.com 
 
From: owner-systemc-p1666-technical@eda.org [
mailto:owner-systemc-p1666-technical@eda.org] On Behalf Of 
john.aynsley@doulos.com
Sent: Tuesday, October 05, 2010 3:35 PM
To: systemc-p1666-technical@eda.org
Subject: TLM-2.0 compliance 
  
All 
Back along we started a discussion on whether P1666 should make any 
statements concerning TLM-2.0 compliance. The protogonists were Stuart, 
Hiroshi Imai, and myself. We reached the conclusion that we wanted to 
define three terms explicitly in the 1666 LRM: 
Informally, 
"A TLM-2.0 compliant implementation" = An implementation that implements 
everything in the 1666 LRM including the TLM-2.0 interoperability layer 
and the TLM-2.0 utilities 
"TLM-2.0 base protocol compliant" = An application that obeys all the 
rules of the base protocol as spelled out in the 1666 LRM 
"TLM-2.0 custom protocol compliant" = An application that uses the 
standard initiator and target sockets specialized with a user-defined 
protocol traits class, but is not obliged to obey any of the base protocol 
rules (though recommended to follow the rules of the base protocol as far 
as possible) 
Does this group wish to continue this discussion and add such term to the 
LRM? 
Thanks, 
John A 
-- 
This message has been scanned for viruses and 
dangerous content by MailScanner, and is 
believed to be clean. 
-- 
This message has been scanned for viruses and 
dangerous content by MailScanner, and is 
believed to be clean. 
-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.