Mechanism for lossless tracing in an architecture having a delay slot

A method for tracing an instrumented program using a thread, including transferring control of the instrumented program to a trap handler to obtain an original instruction associated with a probe, loading the original instruction into a scratch space, setting a program counter to point to the scratch space, setting a next program counter to point to a next instruction, and executing the original instruction in the scratch space using the thread, wherein executing the original instruction results in placing the instrumented program in a state equivalent to natively executing the original instruction.