Minutes Verilog-AMS committee meeting - 3rd April 2008

Present:
- Geoffrey Coram, Analog Devices
- Dave Miller, Freescale
- Patrick O'Halloran, Tiburon DA
- Marek Mierzwinski, Tiburon DA
- Marq Kole, NXP

Agenda:
- continue review of Draft 3

Note:
- an issue was raised whether the timing of the call should be shifted to a
slightly later time to move the people on the West Coast from 6:30 AM to a
more attainable 7:30 AM. The main impact will be for Australia whose
starting time moves well into the night. The reason for the change would be
to have more attendance on the conference call.

Review:
- p.89 - as noted on the reflector an underscore is missing in the analog
initial block syntax rule.

- p.91 - the defparam that was struck out was not allowed according to the
syntax description above it. Question was whether the example should be
updated to have the myscope.pl parameter inherit its value from a
module-level parameter:

module example;
  parameter integer p1 = 1;
  real moduleVar;

  analog begin
    begin: myscope
      parameter real p2 = p1;
      real localVar = 1.5 * p2;
    end
    moduleVar = myscope.localVar;
  end

endmodule

module top;
  example #(.p1(4)) inst1(); // allowed
  example #(.myscope.p2(4)) inst2(); // error
endmodule

- p.113 - The addition of the analog filter function restriction for the
dir and enable arguments of the cross event was questioned. What is exactly
meant with this restriction - ae you not allowed to put in here a variable
whose value depends on an analog filter function, or are you not allowed to
use an analog filter function directly. Why is this restriction here? It
was conceded that an analog filter function for the dir or enable arguments
would probably have a bad impact on efficiency of the event operators but
in itself is not sufficient reason to make this restriction. Moreover, the
same restriction is not present on the enable argument in the above and
timer event functions - that seems inconsistent.

As per request on the reflector the analog_expression and
constant_expression for the second and later arguments should be
analog_expression_or_null and constant_expression_or_null:

analog_event_functions ::=
cross ( analog_expression [ , analog_expression_or_null
[ , constant_expression_or_null [ , constant_expression_or_null [ ,
analog_expression ] ] ] ] )
| above ( analog_expression [ , constant_expression_or_null
[ , constant_expression_or_null [ , analog_expression ] ] ] )
| timer ( analog_expression [ , analog_expression_or_null
[ , constant_expression_or_null [ , analog_expression ] ] ] )

- p.145 - In the heading the word OOMR is used for the first time in this
document - it should be spelled out here followed by the abbreviation in
brackets. In the first bullet the '&' should be replaced by the word 'and'.
In the second bullet the word 'hierarchical' should be changed to
'hierarchically'. The bullets are inconsistently terminated - some have a
period, some don't. This needs to be fixed by providing a period at the end
of each  bullet item.

- p.151 - Regarding the editor's note: bit is here not used as a keyword
but in the technical sense of the word. In the rightmost column in the
bottom row, there is a lot of superfluous courier text - this should not be
courier but the regular font. In the second column of the bottom the
definition 'reg r[0:66];' is in a larger font size than the rest.

- p.152 - The name 'AnalogVar' in this example should be replaced by 'avar'
for consistency.

- p.190 - A space is missing after the inserted 'ddiscrete'.

- p.212 - On the first line of the first paragraph the text 'and $strobe'
should be replaced by '$strobe, and $debug'. On the third line the text '
and $fstrobe' should be replaced by '$fstrobe, and $fdebug'.

- p.217 - This is a completely new section on assertions - there was
insufficient time and not enough people present to go over this. This is
postponed to the next conference call. It is suggested that a discussion on
this is started on the reflector to find out how well this is accepted for
inclusion in this release.

- p.249 - In response to the editor's note the last sentence of section
10.4 should be: 'The ‘undef compiler directive shall have no effect on
these predefined macros and can issue a warning.'.

- p.250 - It was felt that there should also be a "VAMS-2.2" version
specifier as there are a number of keyword differences between Verilog-AMS
2.2 and Verilog-AMS 2.3 that may need to be taken into account.

- p.345 - an underscore is missing in the analog initial block syntax rule
(first rule of A.6.2, third line).

- p.347 - the analog_expression and constant_expression for the second and
later arguments of the analgo event functiosn listed in the
analog_event_functions rule should be analog_expression_or_null and
constant_expression_or_null:

analog_event_functions ::=
cross ( analog_expression [ , analog_expression_or_null
[ , constant_expression_or_null [ , constant_expression_or_null [ ,
analog_expression ] ] ] ] )
| above ( analog_expression [ , constant_expression_or_null
[ , constant_expression_or_null [ , analog_expression ] ] ] )
| timer ( analog_expression [ , analog_expression_or_null
[ , constant_expression_or_null [ , analog_expression ] ] ] )

- p.367 - Need to look up the correct references in this page as per the
editor's note and pass them to Stu.

- p.386 - Annex G should be reviewed and updated to reflect the open items
for LRM 2.3 instead of LRM 2.1.

- p.392 - In the heading of Table H.3 the text should read 'Changes from
v2.1 to v2.2' (the final '.2' is missing in the draft).

- p.394 - The table needs to be extended with items introduced between
Draft 2 and Draft 3.

This concludes the review of Draft 3.


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