Minutes of 05/10/2006 SV-CC Meeting. ATTENDEES 0000000000000 6666666665555 0000000001111 5443332212110 1212101010302 0629515157096 xxxxxxxxxxxxx Charles Dawson x-xxxxxxx-xxx Ankur Kuchlous xxxxxxxxxxxxx Ralph Duncan xxxxxxxxxxxxx Jim Vellenga -xxxxxxxxxxxx Steven Dovich -xx-xxxxxx-xx Andrzej Litwiniuk xxxxxxxxxx-xx Abigail Moorhouse xxxxxxxx-xxxx Michael Rohleder xxxxxxxxx---- Ghassan Khoory xxxxx-------- Chuck Berking xxxxx-------- Bassam Tabbara xxxx--xxx---x Francoise Martinolle xx----------- Amit Kohli -----------x- Nasim Hussain ------------x Sachchidananda Patel -x--x-------- Stu Sutherland 1. Reviewed Patent information. - Chas reviewed the patent information. 2. Reviewed minutes of the 04/26/2006 Meeting. Ralph/Francoise. ACCEPTED 3. Liaisons - Chas to report on P1800 meeting. The meeting was suspended because the PAR had some issues. They are also working on funding for an editor. - Francoise to report on PASSED SV-BC items - Ralph brought up Item 1447 Trying to solve issues with sized and unsized packed dimensions. Should think of them as an array of pointers. Only allocate one dimension at a time. Shalom and Doug are discussing still. Ralph and Doug are writing more material. - No one reported on other meetings 4. New business - ITC issue Ralph has discussed this issue with John Stickley. Proposal made by Ralph was generally accepted. The motivation for the change is to clarify the correct/expected behavior of DPI in the context of SystemVerilog and C. A few suggestions from Amit, Francoise, Michael, and Jim were made which Ralph will try to incorporate into the proposal. The discussions were as follows: a. The current proposal mentions the behavior of invoking DPI scope utilities from non context import functions/tasks as undefined. The proposal still does not mention the behavior as undefined on invoking export functions/tasks from non context import tasks/functions. This should be mentioned explicitly in the proposal. Ralph agreed to this. b. A DPI call chain has been defined as an un-interrupted sequence of invocations of functions starting from a call to DPI-C import function declared with the context keyword. Interpretation: Any other mechanism which does not follow a continuous execution of functions will have an undefined behavior. The vendors have flexibility to extend the invocation of export functions. Suggestion: The dependence on the keyword DPI-C should be removed in the text. The requirement here is for the keyword context, which is independent of the keyword DPI/DPI-C. Ralph agreed to this. c. There was a general discussion around the compatibility of DPI and DPI-C import functions, when one called another without going through the SV design. This was to be determined unrelated to the Mantis item being discussed. The committee will resume discussion on Mantis item 1456 in the next meeting. - Francoise brought up Assignment patterns issues. Current access is very inefficient. Bassam was concerned about proposal to remove vpiStructPattern and trying to flatten. Francoise thinks it does not represent anything. Was intended to represent patterns that exist in case pattern matching but those can be represented as expressions. Bassam was concerned about having two patterns with the same name packed union with two structs with same members. Can you still tell them apart? Francoise: have to use a tag. Francoise will take a look and send back a response. Francoise will put together a proposal. Can also concern packed vectors. Chas will add to new business for next time. 5. Reviewed of items with proposals. - Item 1431 Type 'chandle' was overlooked in sv_vpi_user.h Francoise/Bassam. PASSED (unanimous) 6. Review SV-CC items with proposals (Straw poll only): - Item 1456 Clarify, circumscribe restrictions on use of DPI context utilities Not ready to vote on this one. Let Ralph and Francoise and Amit finish discussing issues with context. Motion to adjourn. 1:03pm. JimV/Abi. PASSED (unanimous) 7. Old Business - JimV to look into other things which can have size changes as discussed in 0985. One thing he found was vpiStringVar. Hasn't finished writeup. 8. Action items - JimV to look into other things which can have size changes as discussed in 0985. - Chas to send everyone the frame docs for the diagrams. - Chuck to write a document that describes the possible solutions for the compatibility issues, and they're pros and cons. - Francoise and Bassam to continue work on assignment patterns. 9. Items for consideration at the next meeting (they already have proposals): - None 10. Next meeting The next SV-CC meeting will be on 05/24/2006. The last P1800 meeting will continue on 05/23/2006. -- 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 Fri May 19 12:50:59 2006
This archive was generated by hypermail 2.1.8 : Fri May 19 2006 - 12:51:22 PDT