Minutes of 10/13/2004 SV-CC Meeting.
ATTENDEES
000000
444444
110000
009999
102211
369250
x-xxxx Ghassan Khoory
xxxxxx Francoise Martinolle
xxxxxx Doug Warmke
xx-xxx Bassam Tabbara
-xxxxx Andrzej Litwiniuk
xxxxxx Joao Geada
xxxxxx Charles Dawson
xx-xxx Jim Vellenga
xxxxx- Ralph Duncan
x--x-- Rob Slater
xx-x-- Sachchidananda Patel
xxxx-- Michael Rohleder
-xxx-- John Stickley
xxx-x- Jim Garnett
xx--x- Steven Dovich
-----x Swapnajit Mitra
-----x Karen Pieper
---x-- Angshuman Saha
1. Reviewed Patent information.
- Charles Dawson read the patent information.
2. Reviewed minutes of the 10/06/04 Meeting.
- Steve/Ralph. ACCEPTED.
3. Liaisons
- Chas to report on the IEEE 1800 meeting that occurred on 10/11/2004.
- Chas reported that:
- The PTF and SV-CC have been merged together.
We will add an additional 1/2 hour onto the beginning of this
meeting and will cancel the PTF meeting beginning next Wednesday.
All PTF issues will need to be addressed here.
- Karen's proposed process for dealing with Items was approved.
We will begin following the process immediately. All proposals
will need to be in the approved form before we can vote on them.
- We need to have everything wrapped up by early December.
At that point, the editors will finalize the initial draft of
P1800.
- The data types sub-group of the BTF was moved under the SV-BC
committee and permitted to continue their efforts on data types
for nets.
- The encryption sub-group was moved out from under the BTF to be
a separate entity under 1800.
- Chas requested clarification of whether the 3 major issues
remaining on the PTF agenda can be addressed given the current
PAR. The 1800 working group concluded that they were and gave
the SV-CC/PTF committee permission to continue to pursue them.
- There were other interesting issues discussed, and everyone
should check out the minutes.
4. New business
- Ralph's question on embedded packed types.
Before charging off on Item 205, Ralph wanted clarification on what
the intent of packed types and embedded packed types was, and how that
would help the users. Andrzej commented that any type would have the
exact same representation, regardless of where it lives. This is
a different story for sub-elements of packed typed objects. Ralph
asked if the intent was to allow any address of a packed type, then
the utilities that work on actual representations should work on all
of them. Andrzej said this was in a sense an extension of C, and
that it is pretty natural (even if not explicit in that standard).
Francoise agreed with Ralph's premise. Andrzej said this is important
for structs passed by reference, regardless of DPI. Without this,
you would need a function to access a struct passed by reference's
value.
The group intends any packed type to be delivered to C in a form
that is consistent. Andrzej took an action item, to make sure the
LRM contains a statement to the effect that for the C layer of DPI,
representations of a type are always the same regardless of where
it is (packed struct, member of array, ...etc.).
5. Review of items with proposals
- Item 205: Clarify size relationship of packed type canonical/vendor formats.
Ralph requested a postponement until next week to write
a new proposal based on discussion of embedded packed types.
- Item 047: Fix 3 erroneous cross-references.
Joao/Bassam. PASSED (unanimous).
- Item 046:Fix 1 typo and 1 grammar error.
Joao/Steve. PASSED (unanimous).
- Item 160: Multiple packed dimensions (agree on interpretation), fix cross-ref & text.
Ralph could not find a definition for linearization. Still don't know.
Joao should know this week.
- Item 049: Decide on semantics: require import fn return value data type, add sentence.
Francoise sent email to SV-BC last week, but no feedback yet.
6. Old Business
- Item 80: vpiTypedef property issues
Discussed. Francoise has made a new proposal. Decided to
remove the typespec property and kept the iteration. JimV
noted the proposal needed a semi and a comma. What about
NULL vs null? Joao replied that it should be using NULL in
this case. Francoise will fix up proposal, and we'll vote on
it next week.
- Everyone to send the set of items to which they have already
made proposals.
DONE.
- Chas to get the database updated to reflect the previous meetings.
Chas has made significant progress on this. One un-answered issue
is that the approved procedures for dealing with changes does not
adequately address the diagrams. Chas to bring up issue with Karen.
- Francoise to ask Peter Ashenden what was done to improve
printing from Rational Rose.
Steve recalled a conversation with Peter on this subject, and
believes that Peter has found a reasonable way of producing
a format that the IEEE will find acceptable.
Meeting ended at 1:05pm (EDT)
7. Action items
- Everyone to send the set of items that they will be working on.
- Chas to get the database updated to reflect the previous meetings.
- Francoise to ask Peter Ashenden what was done to improve
printing from Rational Rose.
- Francoise to inquire about the feasibility of third parties
shipping the UML for the diagrams.
- JimV to resubmit a proposal for Item 123.
- Joao/Francoise to file SV-BC item asking to define linearization.
- Francoise to check with SV-BC on default return type of functions.
- Chas to ask Karen about updating the diagrams (does not fit well
with approved process).
- Andrzej to make sure the LRM says that for the C layer of DPI,
representations of a type are always the same regardless of where
it is (packed struct, member of array, ...etc.).
8. Items for consideration at the next meeting (they already have
proposals):
- Item 080: vpiTypedef property issues
-- Charles Dawson Senior Engineering Manager NC-Verilog Team Cadence Design Systems, Inc. 270 Billerica Road Chelmsford, MA 01824 (978) 262 - 6273 chas@cadence.comReceived on Thu Oct 14 09:52:35 2004
This archive was generated by hypermail 2.1.8 : Thu Oct 14 2004 - 09:52:57 PDT