On Tue, 1 Aug 2006, Daniel de Angelis Cordeiro wrote: > I have here a very specific workload that I want to test in Linux > without the "child-runs-first and in the parent's processor" behaviour. > I know that this behaviour is very good for general workloads because of > the cache affinity; but in my specific case I want that a short-lived > fork()ed process starts to run immediately after the fork(), in the > other CPU. Either modify cpuset of the newly created task (which is duplicated from parent in do_fork() -> copy_process() -> cpuset_fork()), or do it completely from userspace (hint: look at clone() syscall, CLONE_STOPPED flag (in 2.6 kernel) and sys_sched_setaffinity() syscall). -- JiKos. -- Kernelnewbies: Help each other learn about the Linux kernel. Archive: http://mail.nl.linux.org/kernelnewbies/ FAQ: http://kernelnewbies.org/faq/
