Verilog-AMS committee meeting minutes - 11 Dec 2008

Minutes of the meeting: 11 Dec 2008, 6.30am US Pacific time

Attendees:
Graham Helwig, ASTC
Dave Miller, Freescale
Martin O'Leary, Cadence
Patrick O'Halloran, Tiburon
Boris Troyanovsky, Tiburon
Geoffrey Coram, Analog Devices
Dave Cronauer, Synopsys
Sri Chandra, Freescale


Discussion on Mantis items:
===========================

(Please feel free to add/correct if i have not represented it correctly)

Mantis item: 2497
-----------------
* Issue #1:
  - Typo. Accepted. Will be fixed along with typos pointed by Shalom on 
the notes.

* Issue #2:
   - The production of file_path is a 1364 issue; may be addressed in P1800
   - The string_parameter_declaration and 
local_string_parameter_declaration can be removed from the grammar as 
its already covered
   - net/port identifier is a typo, will be fixed as part of issue #1

* Issue #3:
   - specparam does not require ;
   - Probably need to maintain consistency with localparam (this also 
does not require ; in module item decl)
   - It was agreed that the port_declaration syntax in LRM2.3 is correct

* Issue #4:
   - Agreed that the text_macro_identifier is not mentioned anywhere in 
the document. This should be added to Section 10.5

* Issue #5:
   - This issue is a straight production from 1364. This has been 
resolved in P1800 and will be fixed as part of SV integration

* Issue #6:
   - will be fixed in next version

* Issue #7/#8/#11:
   - This issue was discussed and it was agreed that the BNF is not 
intended for this purpose. The identifiers were separated out for 
clarity resulting in ambiguities in certain places. It was agreed that 
the current BNF (having separate identifiers) is more clearer than just 
using "identifier"
   - For issue #8, analog_variable_lvalue needs to be added to Syntax 
5-14 and Syntax 7-3

* Issue #9/#10:
   - Agreed

* Issue #12:
   - (From Shalom's note): "\n" has not appeared in red in 1364 or 1800, 
as it does not represent the characters "\" and "\n", but rather a 
carriage-return. That is, it is not typed exactly as it appears. Hence 
not in red.

* Issue #13:
   - It was agreed to have a separate production for analysis function 
and the test case in the example to be modified. Probably should come 
out of the constant_primary.
   - Instead of documenting illegal analysis() usage in each of the 
different declarations (parameter declaration etc), might be easier to 
add a separate section specifying where analysis() function is allowed.
   - According to P1800 rules of constant_primary and primary; analysis 
should be analog_primary, however it can't be in primary which is purely 
digital.


Mantis Item: 2498
-----------------
* Missing keywords:
   - Some of the keywords got deleted when editing LRM2.3, the extra 
column spilled over after "while" keyword.


Mantis Item: 2535
-----------------
* Multiline strings:
   - This particular syntax production was copied directly from 1364. 
However, P1800 in the latest draft has resolved this issue which can be 
used in the Verilog-AMS production also.


Mantis Item: 2536
-----------------
   - The examples need to be fixed so that they are legal. 1364, AMS and 
1800 requires a digit to appear explicitly before the decimal point.
   - The description in the LRM should say "without leading 
'unsigned_number'" instead of "without trailing". Needs to be fixed.
   - Geoffrey coram has pointed out additional mistake as part of the 
note, the footnote mark is not referenced in the BNF when it was copied 
from A.8.7. This needs to be fixed.


Mantis Item: 2537
-----------------
* Issue #1/#2/#3/#4/#5/#6
   - agreed to be fixed as part of next revision

* Issue #7:
   - It was agreed to have a special production for paramset RHS 
expression rather than making hierarchical parameters part of the 
constant_expression production; probably 
hierarchical_local_parameter_identifier which will reference only 
localparams (as instance paramters cannot be used in RHS expressions for 
these paramsets)
   - Last 3 lines should be wrapped in a module

* Issue #8:
   - the BNF for real_declaration should not have a ";"
   - the port direction and parameter declaration do not have trailing 
";" since these declarations can be standalone or within a list. Hence 
these statements need a separate ";" when they are used in statement form.
   - Agreed in the BNF referred in the mantis item the parameter 
declaration needs a trailing ";"
   - In the BNF mentioned in the mantis item, the port direction 
requires a trailing ";" for the analog_function_item_declaration syntax. 
Fix in 4.7.1 and A.2.6

* Issue #9:
   - Agreed.


Mantis Item: 2538
-----------------
* Strings:
   - change "parameter_type" from keyword to syntax-item in sections 
A.2.1.1 and 3.4
   - remove "string_parameter_declaration" syntax item from section 6.4 
and A.1.9
   - remove "local_string_parameter_declaration" syntax item from 
section 6.4 and A.1.9
   - add ";" symbol after "parameter_declaration syntax item in section 
A.2.8, A.1.9, 6.4, 5.3 and 4.7.1
   - add ";" symbol after "local_parameter_declaration syntax item in 
section A.1.9 and 6.4
   - In A.1.4, Syntax 2-7 and Syntax 6-1, the semicolon after 
parameter_declaration in non_port_module_item should be red.


Mantis Item: 2539
-----------------
* Switch Branch Syntax:
   - accept geoffrey coram suggestion documented as part of Mantis note 
reproduced verbatim.
   - Change from:
     The if-else statement is used to determine whether a statement is 
executed or not. The syntax of a analog conditional statement is shown 
in Syntax 5-8. The conditional expression shall be a constant 
expression. (See the discussion in 4.5.15 regarding restrictions on the 
usage of analog operators.)
   To:
     The if-else statement is used to determine whether a statement is 
executed or not. The syntax of a analog conditional statement is shown 
in Syntax 5-8. If any of the conditionally-executed statements 
(analog_statement_or_null) contains an analog operator, the conditional 
expression (analog_expression) shall be a constant expression. (See the 
discussion in 4.5.15 regarding restrictions on the usage of analog 
operators.)
   - Also, in 5.6.5, after "As a result contribution statements are 
allowed within conditional statements but are not allowed within event 
control statements." add the following note "Note that the contribution 
statements shall not use analog operators when the condition can change 
during the course of a simulation."
   - The indentation in the example (relay) of 5.6.5 is bad; it looks 
like the if statement is "inside" the cross; it would be better to move 
the ";" to the next line (see note in Mantis item for fixed example)

Next Meeting: Scheduled for 18th Dec 2008, 6.30am US Pacific.

Regards,
Sri
-- 
Srikanth Chandrasekaran
Design Technology (Tools Development)
Freescale Semiconductor Inc.
T:+91-120-439 5000 p:x3824 f: x5199

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