Re: Q: Analysis port binding

Philipp, David,

I agree with Philipp on all counts.

unbind() exists and does what you want, but is not particularly
well-documented. I will try to improve the description.

bind() is virtual. We should indeed make unbind() virtual too.

John A

From:
"Philipp A. Hartmann" <philipp.hartmann@offis.de>
To:
David C Black <dcblack@xtreme-eda.com>
Cc:
"systemc-p1666-technical@eda.org WG" <systemc-p1666-technical@eda.org>
Date:
25/10/2010 20:46
Subject:
Re: Q: Analysis port binding
Sent by:
owner-systemc-p1666-technical@eda.org

David, all,

if I understand the analysis ports correctly, it's intended to use
bind()/unbind() dynamically during simulation. The mere existence of
unbind() already suggests that, right?

Moreover, since they are not derived from
sc_port<tlm_analysis_if<T>,...>, I think they can also be
created/destroyed while the simulation is running.

I agree, that both features should probably be documented in the
standard, since the name ..._port might imply the opposite.

Did we agree to make tlm_analysis_port<T>::bind() virtual during the
bind/operator() discussion? In that case, I would propose to make
unbind() virtual as well.

Thanks,
   Philipp

On 23/10/10 16:53, David C Black wrote:
> Is it legal to bind()/unbind() analysis ports during simulation? It
seems to me this would be a useful thing, but it is not specifically
discussed in the standard.
>
> Use case:
> Turn on analysis only during certain sections of
simulation.
> Turn off to reduce overhead
>
> The current TLM permits this behavior, but I don't find it documented
anywhere.

-- 
Philipp A. Hartmann
Hardware/Software Design Methodology Group
OFFIS Institute for Information Technology
R&D Division Transportation · FuE-Bereich Verkehr
Escherweg 2 · 26121 Oldenburg · Germany
Phone/Fax: +49-441-9722-420/282 · PGP: 0x9161A5C0 · http://www.offis.de/
-- 
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.