[sv-cc] Re: Version 2 of DPI LRM


Subject: [sv-cc] Re: Version 2 of DPI LRM
From: Stickley, John (john_stickley@mentorg.com)
Date: Mon Mar 10 2003 - 13:55:52 PST


Doug,

Excellent job on the SV_layer document. It is shaping
up quite nicely.

Here's my minor feedback.

More to follow on other docs.

------------------------------------------------------
Section 1.3.1

Remove lingering reference to non-blocking. Andrzej had already
done this but one fell through the cracks.

Second sentence should be reworded as,

"External functions shall complete their execution ..."

------------------------------------------------------
Section 1.4 - Examples of extern declarations

5th example at the top of page 5. You have two examples
on one line. Insert a newline between the newQueue()
and newElem() declarations.

------------------------------------------------------
Section 1.4.3

Andrzej had changed the function return types to allow
64 bit packed arrays at my request. However it is likely
that we only want to do this if we change the basic
data type to represent packed array logic from
svBitVec32 to svBitVec64 (and svLogicVec64 for 4-state).
I think we should seriously consider doing this in the
C layer. But, if we don't then we should go back to
allowing only 32 bit packed arrays as function return
types for consistency.

------------------------------------------------------
Section 1.7

2nd from last sentence on page 9:

"cname is optional here, it defaults to fname (for example,
for functions defined in the $root scope)."

I'm not sure what is in parentheses has anything to do with cname
defaulting to fname. Why is it in here ? I suggest we remove it.

------------------------------------------------------
Section 1.7 Examples at end of section.

I suggest making one of these examples use an aliased cname.

I would also show module scope context here to illustrate what
is being said in the comments. May I suggest
something like this:

export “DPI” global_func; // export decl. and global_func defined in $root
void global_func(args ...) begin
    ...
end

module bar();
     export “DPI” Cfoo=foo; // export decl. and foo defined in same scope
     void foo(args ...) begin
         ....
     end
endmodule

Warmke, Doug wrote:
> Team,
>
> I can't stand waiting around for the SV-CC reflector to
> finally send you my LRM edits. So here they are, addressed
> individually to the committee members who have been actively
> involved in the debates (I calculated this list by looking at
> emails, and looking at who voted on Joao's proposal.)
> Hopefully I didn't forget anyone, and hopefully I can get
> these files out on the main reflector soon.
>
> Swapnajit,
>
> Can you please find out the limit for sending around files?
> For big files, is there any kind of ftp site or alternative
> to email available?
>
> Andrzej,
>
> Could you please confirm that you received the Frame sources
> for these documents? (They might make more sense after you
> read my introductory email which was swallowed by the reflector
> along with the .pdf attachments...)
>
> Thanks and regards,
> Doug
>
>
>

-- johnS
                                                            __
                        ______ | \
______________________/ \__ / \
                                 \ H Dome ___/ |
John Stickley E | a __ ___/ / \____
Principal Engineer l | l | \ /
Verification Solutions Group | f | \/ ____
Mentor Graphics Corp. - MED C \ -- / /
17 E. Cedar Place a \ __/ / /
Ramsey, NJ 07446 p | / ___/
                                  | / /
mailto:John_Stickley@mentor.com \ /
Phone: (201)818-2585 \ /
                                    ---------



This archive was generated by hypermail 2b28 : Mon Mar 10 2003 - 14:01:41 PST