You can adjust the memory overcommit policy system-wide via sysctl calls, but is it possible to adjust this policy on a per-process basis from within the process itself?
I want to make one particular real-time process never use overcommit, though the rest of the system can overcommit.