Additional Information for TDF Presentation
Many complex servers share code between multiple invocations of the server and between startups. This code is loaded once and not reloaded until it changes. Program identification allows unidentified code to the used in symbols. It can also be used to make CSA code eligible for traps. TDF never allows alteration of nucleus or LPA code. Except in the case of an SRB, TDF normally does not allow alteration of CSA code. However, program identification instructs TDF how to handle a request to alter a CSA program.
The pass through facility handles traps in shared code within the address space shared between tasks that are being debugged and tasks that are not being debugged. Though pass through overhead is minimal, in a dynamic interactive trace with numerous trace points, pass through could alter inter-task timing enough to affect the outcome. Even though this is unlikely, program identification allows the identification of code shared between tasks, particularly targeted code attached multiple times, allowing TDF to create copies of the shared code for the targeted tasks.
Watch a 16 minute video on using shared code and debugging PC routines