sukadev@xxxxxxxxxx writes: > Yes. I can break that up into two patches, but I missed and Badari > pointed the other caller to copy_process() > > struct task_struct * __cpuinit fork_idle(int cpu) > { > struct task_struct *task; > struct pt_regs regs; > > task = copy_process(CLONE_VM, 0, idle_regs(®s), 0, NULL, NULL, 0); > if (!IS_ERR(task)) > init_idle(task, cpu); > > return task; > } > > Now this is passing a null struct pid which would not be good > if I remove the if (likely(p->pid)) check in copy_process(). > > Does this copy_process() mean there can be multiple tasks with > pid_t == 0 (one per cpu on an SMP system) ? Yes. The idle processes. > Can we simply attach all those tasks to init_struct_pid by passing > in &init_struct_pid to the above copy_process() ? Yes. Eric _______________________________________________ Containers mailing list Containers@xxxxxxxxxxxxxx https://lists.osdl.org/mailman/listinfo/containers