RE: [sv-ac] RE: sv-ac 1531

Hi John,

Yes, that is a possibility, to declare the variable outside the
property. Perhaps we need both, though, because if declared on the
outside, you cannot create a property-based library. 
Yes, initialization to all vars.
ed


> -----Original Message-----
> From: John Havlicek [mailto:john.havlicek@freescale.com] 
> Sent: Friday, July 14, 2006 9:47 AM
> To: Eduard.Cerny@synopsys.COM
> Cc: dbustan@freescale.com; Eduard.Cerny@synopsys.COM; 
> sv-ac@eda-stds.org
> Subject: Re: [sv-ac] RE: sv-ac 1531
> 
> Hi Ed:
> 
> Another alternative is to have the user declare the variable
> outside of the sequence or property and then allow the local 
> variable assignment syntax to be used for assigning to the 
> external variable.
> 
> This way of doing things allows one property to move information 
> to another property, not just one attempt of a property moving
> information to another attempt of the same property.
> 
> I think this way of doing things also makes it less important to
> account for the update to the external variable in the formal
> semantics of the assertions.
> 
> I think that in any case, multiple updates to the static variable
> in the same timestep will execute in non-deterministic order.
> 
> I agree with Doron that the initialization should be added for
> all local variables.  This enhancement was requested a long time
> ago.
> 
> J.H.
> 
> > X-Authentication-Warning: server.eda-stds.org: majordom set 
> sender to owner-sv-ac@eda-stds.org using -f
> > X-MimeOLE: Produced By Microsoft Exchange V6.5.7226.0
> > Content-class: urn:content-classes:message
> > Date: Fri, 14 Jul 2006 06:17:47 -0700
> > Thread-Topic: sv-ac 1531
> > Thread-Index: Acamy4G8DmiIMBYVRdOgnDTXESgyvAAfCTlw
> > From: "Eduard Cerny" <Eduard.Cerny@synopsys.com>
> > Cc: <sv-ac@eda-stds.org>
> > X-OriginalArrivalTime: 14 Jul 2006 13:17:50.0201 (UTC) 
> FILETIME=[E7E9BE90:01C6A747]
> > X-Virus-Status: Clean
> > X-MIME-Autoconverted: from quoted-printable to 8bit by 
> server.eda-stds.org id k6EDHqDu029077
> > Sender: owner-sv-ac@eda-stds.org
> > 
> > Doron,
> > 
> > I would be for leaving the semantics as loose as possible, in other
> > words, if the user writes to a global variable from two parallel
> > processes, there is nothing in the language to stop 
> him/her. Why do we
> > have to policy it in the assertios? As I said earlier, if I 
> use tasks, I
> > can do this today, but it is cumbersome. 
> > 
> > ed
> >  
> > 
> > > -----Original Message-----
> > > From: Doron Bustan [mailto:dbustan@freescale.com] 
> > > Sent: Thursday, July 13, 2006 6:27 PM
> > > To: Eduard Cerny
> > > Cc: sv-ac@eda-stds.org
> > > Subject: sv-ac 1531
> > > 
> > > Ed,
> > > 
> > > we need a  definition of a semantic for static local variables.
> > > In particular it make sense to define different flow rules
> > > for static local variables.
> > > 
> > > for example
> > > 
> > > property p1;
> > > logic [3:0] v;
> > > 
> > > ((a, v = 1) or (b, v=2)) ##1 (v == 1);
> > > endproperty
> > > 
> > > is legal, but
> > > 
> > > property p2;
> > > static logic [3:0] v;
> > > 
> > > ((a, v = 1) or (b, v=2)) ##1 (v == 1);
> > > endproperty
> > > 
> > > probably should not be legal, because the value of v flowing 
> > > out of ((a, 
> > > v = 1) or (b, v=2)) is undefined.
> > > 
> > > BTW, I think that initialization should be in a separate proposal.
> > > 
> > > Doron
> > > 
> > 
>