I have bumped the number of threads in the t_ptrace_wait* LWP tests from 3 to 100. Each thread attempts to generate 1 or 2 SIGTRAP. 100 threads struggling to emit 100 sigtraps generates complexity that affects timing of execution.
Please report back whether the execution on low-end setups or software emulation runs without timeouts. If there are timeouts please reduce TRACE_THREADS_NUM from 100 to a smaller value and please report back what is the reliable number. If there are any failures in this setup, please report to me. The purpose of keeping the TRACE_THREADS_NUM number as high as possible is to torture the kernel and more frequently trigger any kernel anomalies.
