ISAC: Minutes from meeting on 20 March 2008

These are also on the website.
Chuck Swart


-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.


Minutes of ISAC meeting held via telecom on 20 March 2008

Present: Larry Soule, Chuck Swart, Ajay Verikat

Absent: Peter Ashenden, Jim Lewis, Lance Thompson

Next meeting: demand driven. The ISAC will meet as needed to
support transition of the LRM from D4.0 to IEEE ballot. 
If nothing else comes up, the ISAC will next meet to review LRM
changes.

TOPIC: IEEE reformatting. The IEEE wants us to reorganize the LRM to
meet their current standards. This will make it very difficult to
trace historic developments, which will make life harder for the ISAC
and for much individual use. However, it is generally agreed that even
if the VASG argued against the reorganization, eventually the IEEE
will pressure the VASG to reorganize the document. So it was agreed
not to resist the reformatting. This is subject to adequate funding
for the work, a topic outside the scope of the ISAC.

TOPIC: Bugzilla 192 Subtype problems for out-mode signal parameters,
and 193 Sensitivity of concurrent procedure call with out-mode signal
parameter

The proposal to disallow reading of OUT mode parameters is accepted as
the best solution. This restriction is backward compatible and could
later be relaxed if a better solution can be found.

ACTION: Chuck to update Bugzilla issues.

TOPIC: IR 2126 Concatenation ambiguity 

The IR was approved, subject to a few  minor editorial changes.

ACTION: Chuck to forward to VASG (after incorporating small changes.)

TOPIC: IR 2123 Process resumption and callbacks

The general solution, that a process resumes whenever there is an
event on a signal in the wait sensitivity, is accepted. However,
several ISAC members are concerned that the callback triggers are
optional. For example, optimized code could determine that the wait
expression evaluates to FALSE, and bypass the VHPI callbacks. This
gives a degree of indeterminacy and potential non-portability to the
language.  The proposed interpretations affects only callbacks related
to VHPI, so introduces no new indeterminism outside of VHPI
applications.  The VHPI working group has considered this issue, and
their decision, after careful deliberation, to allow this behavior,
should be respected.

It was also observed that the described situation is a relatively
uncommon corner case, and is unlikely to cause significant problems to
most actual VHPI users.  Chuck will add a note to the analysis
pointing out these observations.  The issue is ISAC-Approved.

ACTION: Chuck to send note to ISAC/VHPI for quick review, then to
forward the IR to VASG for vote.

TOPIC: IR 2124 Ordering of process execution and callbacks.

The proposed wording relaxes the stated requirements on ordering
of callbacks related to processes to only require that, for
each given process the related vhpiCbSuspend callback associated with
that process occurs after that process suspends and that the
related vhpiCbResume callback is executed before the process resumes.
No other assumptions are made relative to other processes.


One ISAC member observed that the relaxed ordering prevented a VHPI
user from determining process ordering and that at least one existing
debugger application exploited this capability. The VHPI working group
has considered this issue, and their decision, after careful
deliberation, to relax the ordering should be respected. It was also
observed that the described situation is a relatively uncommon corner
case, and is unlikely to cause significant problems to most actual
VHPI users.  Chuck will add a note to the analysis pointing out this
observation.  The issue is ISAC-Approved.

ACTION: Chuck to send note to ISAC/VHPI for quick review, then to
forward the IR to VASG for vote.