RE: P1666 Ballot resolution discussion - Part 4

Fine with me. Deleted.

John A.

From:
"Michael (Mac) McNamara" <mcnamara@cadence.com>
To:
Stuart Swan <stuart@cadence.com>, Jerome CORNET <jerome.cornet@st.com>,
"john.aynsley@doulos.com" <john.aynsley@doulos.com>,
"systemc-p1666-technical@eda.org" <systemc-p1666-technical@eda.org>
Date:
27/04/2011 20:49
Subject:
RE: P1666 Ballot resolution discussion - Part 4

I would lean to option 1 ? delete the sentence.
 
 

 
Michael McNamara | Vice President and General Manager, System Level
Design
M: 1.408.348.7025 W: www.cadence.com E: mcnamara@cadence.com
 
 

 
 
From: owner-systemc-p1666-technical@eda.org [
mailto:owner-systemc-p1666-technical@eda.org] On Behalf Of Stuart Swan
Sent: Wednesday, April 27, 2011 11:49 AM
To: Jerome CORNET; john.aynsley@doulos.com;
systemc-p1666-technical@eda.org
Subject: RE: P1666 Ballot resolution discussion - Part 4
 
John, Jerome, All-
 
I don?t have a strong opinion on the ?Pointers of references to shared
memory should not be used as a backdoor mechanism?
sentence in 17.1.1. I see three possibilities:
 
1) Delete the sentence. If it is deleted, the rest of the paragraph
seems to provide enough guidance on how the TLM1 interfaces are intended
to be used.
2) Leave the sentence as-is. It is not too far off the mark, though
arguably a bit too concise..
3) Try to add more precision to the sentence, something along the
lines of: If pointers or references are used in TLM1 transactions, it is
strongly recommended that the application adhere to a protocol for usage
of the pointers and references that is fully aligned with the message
passing semantics of the associated TLM1 interface calls.
 
I really do not have a strong opinion between the three options.
 
Thanks
Stuart
 
From: Jerome CORNET [mailto:jerome.cornet@st.com]
Sent: Wednesday, April 27, 2011 9:13 AM
To: john.aynsley@doulos.com; systemc-p1666-technical@eda.org
Cc: Stuart Swan
Subject: RE: P1666 Ballot resolution discussion - Part 4
 
John, all
 
comments below.
 
From: john.aynsley@doulos.com [mailto:john.aynsley@doulos.com]
Sent: Wednesday, April 27, 2011 3:23 PM
To: Jerome CORNET; systemc-p1666-technical@eda.org
Subject: Re: P1666 Ballot resolution discussion - Part 4
 
Jerome, All,

Finally, Jerome from ST has raised 3 issues as follows. Comments please.

Issue #1. "The bibliography in "Reference" section is missing several
references that were present in the latest draft document reviewed by the
P1666 working group. Synchronize this section with the latest approved
version discussed by the working group."

RESOLUTION:

The bibliography in question is the following:

The following books may provide useful background information:
Transaction-Level Modeling with SystemC, TLM Concepts and Applications for
Embedded Systems, edited
by Frank Ghenassia, published by Springer 2005, ISBN 10 0 387-26232-6(HB),
ISBN 13 978-0-387-26232-
1(HB)
Integrated System-Level Modeling of Network-on-Chip enabled
Multi-Processor Platforms, by Tim Kogel,
Rainer Leupers, and Heinrich Meyr, published by Springer 2006, ISBN 10
1-4020-4825-4(HB), ISBN 13
978-1-4020-4825-4(HB)
ESL Design and Verification, by Brian Bailey, Grant Martin and Andrew
Piziali, published by Morgan
Kaufmann/Elsevier 2007, ISBN 10 0 12 373551-3, ISBN 13 978 0 12 373551-5

This bibliography came from the OSCI TLM-2.0 LRM, not from the previous
version of the 1666 SystemC standard, which had no bibliography. Under
IEEE rules, it is not permissible to include non-normative references in
the main body of the standard, which fact resulted in the draft LRM being
initially rejected by the IEEE. Hence the bibliography was removed. (It
would be possible to add back this bibliography as an Annex at the end of
the LRM, but I do not propose to do this unless there is a consensus to do
so.)
[JC] Just as a curiosity, what is the definition of ?non-normative?? What
makes the current reference to the ?Requirements Specification for TLM 2?
?normative??

Issue #3. Subclause 17.1.1 ' "Pointers of references to shared memory
should not be used as a backdoor mechanism". Re-reading and reviewing this
section, I find this comment part misleading: existing TLM-1 protocols do
make use of pointers to shared memory sometimes, which this comment seems
to exclude. All in one, this part seems like an unnecessary restriction.
Maybe it could be rephrased as "being careful with pointers or references
to shared memory" or removed altogether (I don't think it reaches the
original objectives for introducing this comment; tbd)'

RESOLUTION

This issue was previously debated in the Working Group and the conclusions
agreed (with Jerome) and written up the LRM. Hence my default position is
that I do not propose to make any changes. However, I would invite Jerome
to seek support for his view on this reflector over the next 10 days.
[JC] This issue was briefly discussed in the Working Group, but its
misleading character was uncovered afterwards. I think it is Stuart that
wanted to introduce that
specific comment on backdoor mechanisms, which I now read as unnecessary
restrictive. Stuart, could you comment?
Thanks,
 
Jerome
 

-- 
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.