All, From a review of LRM 2.2 and the sections updated and reviewed so far for LRM 2.3, I find that most of the text needed for the support of multiple analog blocks needs to be added to section 6 (Analog Behavior). In particular, I suggest to start from the latest version of merged_beh.pdf from the public documents section of the Verilog-AMS committee. As a suggestion for the sections and kinds of changes, here is an overview. I've also tried to note whether most issues are known and a solution has been proposed. 6.1 Analog Procedural Block --> remove text stating the limitation to a single analog block [no issues] 6.1.1 Concurrent Analog Blocks [NEW] --> describe multiple (concurrent) analog blocks [no issues] ... 6.2.1 Sequential Blocks --> add the concatenate sequential block syntax to Syntax 6-2 [use concatenate syntax as agreed in the committee telecon of April 12, 2007] --> allow branch declaration at the top of an analog sequential block [branch declarations are module level declarations in LRM 2.2, but need to be allowed in analog sequential blocks as well, both named and unnamed. However, it needs to be restricted to the top-level analog sequential blocks. Need to assess the implications for the Verilog-AMS 2.3 syntax.] 6.2.3 Concatenated Analog Blocks [NEW] --> describe the concatenation of named and unnamed analog blocks [use concatenate syntax as agreed in the committee telecon of April 12, 2007] ... 6.3.1.3 Value Retention --> add a description for value retention in case of multiple concurrent analog blocks [refer to section 6.1.1 to describe value retention for unnamed branches in case of multiple concurrent analog blocks. Similarly explain use of analog block level named branches in cases where the unnamed module-level branches cannot be used.] 6.3.1.4 Switch Branches --> add a limitation to allow switch branches only in a single analog block [disallow switch branches distributed over multiple concurrent analog blocks, as agreed in the committee telecon of April 12, 2007] 6.4.1 Concurrent Assignment --> describe the limitations to variable assignments in case of multiple concurrent analog blocks [open issue: should the concurrent analog blocks be evaluated atomically as suggested as an optional requirement in the Multiple Analog Blocks V4 document section 3.1?] [open issue: add a requirement for the run-time detection of race conditions in case of assignment to the same variable in multiple concurrent analog blocks?] 6.4.1.1 Concurrent Variable Access [NEW] --> related to the concurrent variable assignment there is also the issue of variable access in case of multiple concurrent analog blocks [open issue: no decision yet on how to handle reliable access to a variable that has been assigned a value in another concurrent analog block] With the multiple analog blocks defined mainly in section 6, there are few isolated places in the rest of the LRM that need to be updated. 7.1 Modules --> add text describing the support of multiple analog blocks. [no issues] Comments are welcome. Cheers, Marq Marq Kole Domain Leader Robust Design Research NXP Semiconductors -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.Received on Wed, 18 Apr 2007 22:13:04 +0200
This archive was generated by hypermail 2.1.8 : Wed Apr 18 2007 - 13:14:40 PDT