Self-Modifying Code Detector for the IBM Series/1 Processor
Original Publication Date: 1985-Aug-01
Included in the Prior Art Database: 2005-Feb-19
This article describes a technique which allows a programmer to detect self-modifying code in existing software without spending substantial time to track down the error. The code detector disclosed herein is utilized with an instruction buffer [Ibuffer (not shown)] which is divided into two separate buffers: a two-word buffer on the processor card and a one-word buffer on the storage card. When a program starts to execute logic on the processor card, it will attempt to keep its two-word buffer full by prefetching instructions from the instruction stream. The storage card will fill its buffer, whenever it is empty, on the next instruction fetch by simultaneously fetching two instructions, placing one instruction in the buffer and sending the other to the processor.