This release does not support automatic collection of performance data under dbx on descendant processes. If you want to collect data on descendant processes automatically, use the collect command.
The following sections describe how to collect data under dbx on a descendant processes, and what happens to data collection for the parent process, for the cases of calls to fork and calls to exec.
Data collection continues for the parent process after it calls fork. To collect data on the descendant process, attach dbx to the descendant process and enable data collection. See Collecting Performance Data on a Running Process for more information.
If the parent process calls exec and the call fails, data collection continues as if nothing happened. If the parent process successfully calls exec, the experiment on the parent is terminated abnormally, regardless of which tool you use for data collection. This experiment can still be read by the Performance Analyzer. However, if you want to ensure normal termination of the parent experiment when you are using dbx, set a breakpoint before the call to exec and close the experiment when the program stops. You can open a new experiment on the descendant process when dbx stops the process at the beginning of its execution.
Collecting Performance Data |