From: Stuart Sutherland (stuart@sutherland-hdl.com)
Date: Tue Aug 14 2001 - 05:43:47 PDT
Chris,
The IEEE spec for tf_error() and tf_message() state that these routines
only cause the simulator to exit when then they are called from a checktf
routine. When called from a calltf, misctf or sizetf routine, they print a
message but do not cause the simulator to exit.
I do wish that how the routines cause simulators to exit from a checktf
routine was better defined in the LRM, but that is left up to the
simulator. That vagueness is deliberate, because the standard came out
after products had implemented different ways to exit. On the PLI task
force, we chose to allow all exit behaviors to be legal, rather than
breaking existing products by over-constraining the standard. From outward
appearances, it appears to me that some simulators schedule the exit at the
end of one of their passes of compilation/elaboration, and other simulators
exit immediately without even completing the checktf routine that called
tf_error() or tf_message(). I prefer the scheduled approach.
Stu
At 07:41 AM 8/9/2001, Chris Spear wrote:
>Precedence: bulk
>
>Can I get a little more clarification on how tf_error and tf_message work
>outside of the checktf? I suggest the following addition, based on what
>VCS does:
>
>25.13 tf_error
>
>If tf_error is called from the calltf or misctf application associated with
>the user-defined system task or function, the following rules shall apply:
>
>- The calltf or misctf application will be allowed to complete.
>
>- No further events in the model will be processed.
>
>- All the misctf applications will be called with reason_finish to allow
>the user-defined system tasks and functions to close.
>
>
>
>25.31 tf_message
>
>If tf_message is called from the calltf or misctf application associated
>with the user-defined system task or function, and the level is ERR_ERROR,
>ERR_SYSTEM, ERR_INTERNAL, the following rules shall apply:
>
>- The calltf or misctf application will be allowed to complete.
>
>- No further events in the model will be processed.
>
>- All the misctf applications will be called with reason_finish to allow
>the user-defined system tasks and functions to close.
>
>
>Okay, I am a dinosaur for still using these tasks!
>
>
>/*****************************************************************
>Chris Spear Verification Specialist .. __o
>Synopsys, Inc. Email Chris@Spear.net _`\<,_
>154 Crane Meadow Road Phone 508-486-5214 .. (*)/ (*)
>Marlboro, MA 01752 Fax 508-486-5223 My cogs go to 11
>http://chris.spear.net Car 508-254-7223
>*****************************************************************/
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
Stuart Sutherland Sutherland HDL Inc.
stuart@sutherland-hdl.com 22805 SW 92nd Place
phone: 503-692-0898 Tualatin, OR 97062
www.sutherland-hdl.com
~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
This archive was generated by hypermail 2.1.4
: Mon Jul 08 2002 - 12:54:44 PDT
and
sponsored by Boyd Technology, Inc.