[ibis-macro] DLL debugging and validation

  • From: Essaid BENSOUDANE <essaid.bensoudane@xxxxxx>
  • To: ibis-macro@xxxxxxxxxxxxx
  • Date: Tue, 31 Jul 2007 16:16:34 -0400

Subject: Debugging and validation

A final product for SERDES modeling using the BIRD should allow the
model developer to debug it's own source code under any platform. It's
not true that the only way to debug the Dll is to have access to source
code that load it. The C compiler allows individual file to be compiled
on various mode (see gcc switches). Therefore you could choose to
generate files symbol tables or not. 

Adding any license fee for an environment to allow debugging is not
acceptable at all. By default debugging capabilities should be included
in the tools. Tools provider should find a solution to be able to hock
gdb to their tools as a client. I already used similar stuff in the past
to integrate processor C-ISS models (as slave) in some EDA tools and I
was able to debug my code without need to have access to EDA tools
source code. 
 
Having access to both test environment and model for the current
environment is understandable, but it should not be the rule.


Functional validation is another issue, where the model developer should
make sure his model much the desired functionalities. In some cases the
model should be validated to reflect an already existing SERDES. It's
the responsibility of the model developer to make sure the DLL he
delivers is validated and conform to ATM specification. 

Best regards,

       
-- 
Essaid Bensoudane           | System-on-Chip Platform Automation
essaid.bensoudane@xxxxxx    | STMicroelectronics, AST
Phone: +1-613-768-9006      | 16 Fitzgerald Road, Suite 300
Fax:   +1-613-768-9025      | Nepean, Ontario K2H 8R6 Canada

---------------------------------------------------------------------
IBIS Macro website  :  http://www.eda.org/pub/ibis/macromodel_wip/
IBIS Macro reflector:  http://www.freelists.org/list/ibis-macro
To unsubscribe send an email:
  To: ibis-macro-request@xxxxxxxxxxxxx
  Subject: unsubscribe

Other related posts: