Stuart,
I ran this file through a testcase and I've found
a number of problems.  Some of them are easy to fix;
others may require discussion in committee.
Please stay tuned while I prepare a report on what I found.
Thanks and regards,
Doug 
> -----Original Message-----
> From: owner-sv-cc@eda.org [mailto:owner-sv-cc@eda.org] On 
> Behalf Of Stuart Sutherland
> Sent: Thursday, January 13, 2005 2:19 AM
> To: sv-cc@eda.org
> Subject: [sv-cc] modified sv_vpi_user.h to test
> 
> 
> 
> I have attached a modified sv_vpi_user.h file with all of the 
> changes from
> approved changes added in.  Please verify that the modifications are
> correct, and that the file compiles OK (if that is possible 
> for anyone to
> test).
> 
> I have attached both the .h file for testing, and a PDF of 
> the LRM, which
> shows the deletions and additions to the erratum #158 version 
> of the file.
> 
> As you probably know, erratum #158 replaced the entire file, 
> and several
> other errata modified the file (62, 123, 213, 256, 257, 295, 
> 333).  Most of
> the changes were specified on the pre-#158 version of the 
> file, but were not
> a problem to add to the #158 version.  However, I do have 
> questions on two
> of the changes.
> 
> Change #62 specified a location to add a constant that no 
> longer exists.  I
> added the new constant in what seemed a logical location, and 
> changed the
> constant value to fit that new location.  Please verify that 
> what I did was
> correct.  The new constant is:
> 
>   #define vpiArrayMember            607
> 
> 
> The following constants originally had values with a single 
> bit set (1,2,4).
> The revised header file (from #158) uses simple binary count 
> values (1,2,3).
> Erratum #267 adds two more constants, following the old 
> convention (sort
> of).  Are these values OK?  Was the original one-hot pattern 
> needed?  Since
> the original values were changed by 158 anyway, can I just 
> re-sequence all
> the values after adding #267 (1,2,3,4,5)?
> 
> WAS:
>   #define vpiQualifier 621 
>   #define vpiUniqueQualifier 1 
>   #define vpiPriorityQualifier 2 
>   #define vpiTaggedQualifier 4 
> 
> AFTER #158 and #267
>   #define vpiQualifier 621 
>   #define vpiNoQualifier 0 
>   #define vpiUniqueQualifier 1 
>   #define vpiPriorityQualifier 2 
>   #define vpiTaggedQualifier 3 
>   #define vpiRandQualifier 8
> 
> Please respond ASAP.  All editing closes this afternoon, Pacific time.
> 
> Stu
> ~~~~~~~~~~~~~~~~~~~~~~~~~
> Stuart Sutherland
> stuart@sutherland-hdl.com
> +1-503-692-0898
>  
> 
Received on Fri Jan 14 15:49:16 2005
This archive was generated by hypermail 2.1.8 : Fri Jan 14 2005 - 15:49:20 PST