The vulnerability lies in the pypdf.filters.RunLengthDecode.decode function, which is responsible for handling Run-Length Decode streams in PDF files. The provided commit patch clearly shows the addition of a size check within this function. Specifically, the code now tracks the total_length of the decompressed data and compares it against a new limit, RUN_LENGTH_MAX_OUTPUT_LENGTH. If the decompressed size exceeds this limit, a LimitReachedError is raised. This directly addresses the reported vulnerability of potential memory exhaustion from a manipulated stream. Therefore, the RunLengthDecode.decode function is the exact location of the vulnerability, and it would be the function appearing in a runtime profile when the vulnerability is triggered.