regression-wally needs to search for non-mismatch errors during lockstep

[rosethompson]

Lockstep does not detect all types of failures. The current WALLY-COV-fsh.S functional coverage test causes a trap without a trap handler. ImperasDV and Wally both inter an infinite loop of traps to address 0. We restored a check to the testbench but regression-wally need to identify this as failure rather than success. The current grep string is "Mismatch: 0". A multiline grep string is likely needed.