Observation by a debug host with memory model and timing offset calculation between instruction and data traces of software execution carried on in a debug target having a main memory and a cache arrangement
Abstract:
A method, performed in a debug host, for observing software execution on a computer having one or more processor cores, a cache attached to the one or more processor cores via respective execution pipelines forming a cache arrangement, and a memory, comprises obtaining an instruction trace of the cache arrangement and a data trace for data being loaded from the memory into the cache. The instruction trace is synchronized with the data trace to generate a synchronized data trace and/or a synchronized instruction trace. A state of a memory model, representing a memory readable by the one or more processor cores via a respective instruction is updated using the synchronized data trace and the synchronized instruction trace.
Information query
Patent Agency Ranking
0/0