Our system consists of about 100 software modules. Each module is a process with up to about 15 threads each.
We run them under Linux with the RT patch, and schedule them with SCHED_FIFO.
When a certain thread gets interrupted I would like to be able to tell why.
1. Modifying the source of the modules is not an option.
2. Hacking the kernel is not an option.
So what i need is a user space program which can monitor a bunch of threads. I have already written a program which runs as a daemon monitoring the proc files for each process, so it would be nice if I could somehow add this monitoring capability to this program.
Anybody here have any information regarding this or know where I should ask?