Answers Database

XABEL 6: Feedback signals interpreted differently in M1 (EDIF) than XACT(PLD)


Record #3101

Product Family:  Software

Product Line:  Data I/O

Problem Title:
XABEL 6: Feedback signals interpreted differently in M1 (EDIF) than
XACT(PLD)



Problem Description:
Keywords: XABEL, feedback

Urgency: standard

General Description:

In XABEL, a signal is used to drive an output pin and also used
as feedback in other equations. The signal feedback is
interpreted as a pin feedback (signal.pin) using the M1 EDIF
flow, but it was interpreted as an internal feedback
(signal.fb) using the XACT6 or XABEL-CPLD flow. This problem
occurs when the signal name has no dot-extension when used as
feedback in other equations. The output equation for the signal
also has an output-enable condition (signal.oe) defined.


Solution 1:

Specify signal.fb (for internal feedback) or signal.pin (for
pin feedback) wherever the signal name appears in other
equations.

When an output signal is also used as feedback in other
equations, the ABEL language sometimes requires the use of dot
extensions (signal.fb or signal.pin) to avoid ambiguity. If the
signal has no output-enable equation (signal.oe=...), ABEL
assumes internal feedback (.fb) and the implementation will
match for both XACT6 (XABEL-CPLD) and M1 (EDIF flow). If the
signal does have a .oe equation, ABEL assumes pin feedback
instead. The Plusasm netlist writer (pla2eqn) in XACT6 (or
XABEL-CPLD) incorrectly assumes that the unqualified signal
name always implies internal feedback. The M1 EDIF flow
(abl2edif) correctly interprets the pin feedback structure
created by ABEL. To avoid ambiguity, always use .fb or .pin
extensions on all feedback instances whenever the signal drives
a 3-state output pin.



End of Record #3101