BTF - BE45 - Behavioral function definition

From: Anders Nordstrom (andersn@bcarsb82)
Date: Wed Sep 03 1997 - 10:26:37 PDT

Subject: BTF - BE45 - Behavioral function definition

Behavioral Task Force - Errata Submission

Assigned Enhancement Request Number: BE45
Errata Name (Description): Behavioral function definition
Section: 11
Date Submitted: 961217
Requestor: Steve Meyer

Status: Submitted (priority not yet assigned)

Errors found in the Verilog LRM (IEEE 1364-1995).

Details:

I just ran into the following function that I think may be illegal
according to the IEEE P1364 standard but probably needs to be explicitly
allowed because break and continue require named blocks. The function below
does not work in Cver because it treats named blocks as task enables
since they can be disabled, and for nested named blocks the downward
chain disabling rule (section 11-1 Oct. 1995 draft) must be applied.

The problem is probably one of documentation clarity. I read section
11 page 1 (but not the remaining examples) as allowing disable of any
named block, therefore named blocks are illegal inside function definitions,
but my interpretation must be wrong since the function is part of the DA
solutions benchmarks and works on other simulators.

Also, I am not sure what the meaning of parallel blocks (fork-join)
is inside functions.

The LRM grammar, in my view correctly, does not define timing behavior
semantics inside functions (see A.2 and A.6). It allows any statement.

I assume it is illegal from outside a function to disable a named block
inside a function because that implies timing control that is explicitly
dis-allowed inside functions. Such disable is, I think, unlikely except for
debuggers that allow statement by statement tracing inside functions.
However, I read section 5.4.2 sentence 2 as allowing functions to be
interrupted in the middle (say) to execute other events because there is
no non timing control statement group atomicity condition.

I hope behavior within functions and for named blocks will be defined more
exactly in the new version of the LRM.

/Steve 

---

// function from DA Solutions public domain bench marks das_cpu example
function [0:31]	sign_extend;

input	[0:15]	D2;
reg	[0:31]	result;

begin
   begin
	begin : sextend
        integer i;
        begin
	for (i = 0; i < 16; i = i+1)
	begin
	result[i] = D2[0];
	result[i+16] = D2[i];
	end
	end
   end
   sign_extend = result;
end
endfunction
endmodule

// here is a test top level module 
module top;
reg [0:15] D, X;
 
initial
 begin
  D = 6;
  X = sign_extend(D);
  $display("sign extend of %b is %b", D, X);
 end
-- 
Steve Meyer				Phone: (415) 296-7017
Pragmatic C Software Corp.		Fax:   (415) 781-1116
220 Montgomery St., Suite 925		email: sjmeyer@crl.com
San Francisco, CA 94104

<p>Hi, Steve.

I'll add my two cents worth.

Let me preface my remarks by saying that I have only the April 1995 draft
since I apparently have not been judged eligible to get 1364 mailings on a 
regular basis.

I don't agree with your interpretations.

I don't think that named blocks should be considered task enables.
My P1364 draft seems to distinguish between named blocks and tasks.
1364 does not imply, in my opinion, that because named blocks can be disabled,
then they are like task enables.  You don't explain why you think that
one necessarily follows from the other.

In certain respects, they are similar, but different in others.

I don't see a reason that named blocks should be illegal in functions.

I have in the past argued that functions MUST be executed atomically
in order to guarantee coherent execution.

Regards,
******************************************************************************
Shalom Bresticker				email: shalom@msil.sps.mot.com
Motorola Semiconductor Israel, Ltd.		Tel #: +972 9 9522268
P.O.B. 2208, Herzlia 46120, ISRAEL		Fax #: +972 9 9522444
******************************************************************************

+-----------------------------------------------------------------------+
|	Anders Nordstrom						|
|	Senior ASIC Designer						|
|									|
|	Northern Telecom Ltd.		Email: andersn@nortel.ca	|
|	P.O. Box 3511 Station C		Phone: 613-763-9186		|
|	Ottawa, Ontario K1Y 4H7		Fax: 613-763-2626		|
+-----------------------------------------------------------------------+

This archive was generated by hypermail 2.1.4 : Mon Jul 08 2002 - 12:54:41 PDT and
sponsored by Boyd Technology, Inc.