[sv-ac] comments on 1731

Hi All:

I have reviewed the proposal for 1731 and noticed three things
that I would like to see addressed before we vote.  The second
is the most important.

J.H.


2007-06-26
----------

. p. 1.  I recommend using the labels "a1", "a2" rather than "p1",
  "p2" for the assertion directives.  This change is for consistency
  with style across Clause 16.

. p. 1.  I don't think that the sentences

     Both assertions p1 and p2 read: whenever en is high and req
     rises, at the next cycle gnt must be asserted.  The difference
     between these assertions is that in p1 the rise of req is
     evaluated relative to the assertion clock - posedge clk, while in
     p2 the rising of req is evaluated relative to posedge fclk.

  adequately explain the difference.  I recommend changing them to
  something like the following (in which I also use the labels "a1",
  "a2" as suggested above):

     Both assertions a1 and a2 read: whenever en is high and req
     rises, at the next cycle gnt must be asserted.  In both
     assertions, the rise of req occurs if and only if the sampled
     value of req at the current posedge of clk is 1'b1 and the
     sampled value of req at a particular prior point is distinct from
     1'b1.  The assertions differ in the specification of the prior
     point.  In a1 the prior point is the preceding posedge of clk,
     while in a2 the prior point is the most recent prior posedge of
     fclk.

  One may want to clarify that req is a 1-bit signal or that the
  condition for rise applies only to the LSB of req.

  This positioning of the description is somewhat awkward because the
  details of the semantics of $rose have not yet been discussed.  One
  could move the entire subject starting with "When these functions
  are used" to the end of subclause 16.8.3 to address this.
  
. p. 2.  I like simplifying "w^{i+1,j},{},{} |== (!c[*0:$] ##1 c).
  Consider changing "(c[->1])" to "c[->1]" because the parentheses
  are not needed.

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