On Mon, Apr 20, 2020 at 02:05:57PM -0700, Matthew Wilcox wrote: > On Mon, Apr 20, 2020 at 11:57:29PM +0300, Alexey Dobriyan wrote: > > I don't anyone have been crazy enough to spawn 2^32 threads. > > It'd require absurd amounts of physical memory, and bump into futex pid > > limit anyway. > > > > Meanwhile save few bits on REX prefixes and some stack space for upcoming > > print_integer() stuff. > > > > And remove "extern" from prototypes while I'm at it. > > It seems like there's a few more places to fix in this regard? > > kernel/sched/fair.c:static u64 __sched_period(unsigned long nr_running) > kernel/sched/sched.h: unsigned long dl_nr_running; > kernel/sched/core.c:unsigned long nr_iowait_cpu(int cpu) > kernel/sched/core.c:unsigned long nr_iowait(void) > kernel/sched/loadavg.c: long nr_active, delta = 0; > kernel/sched/sched.h: unsigned long rt_nr_migratory; > kernel/sched/sched.h: unsigned long rt_nr_total; > kernel/sched/sched.h: unsigned long rt_nr_boosted; > kernel/sched/sched.h: unsigned long dl_nr_running; > kernel/sched/sched.h: unsigned long dl_nr_migratory; > kernel/sched/sched.h: unsigned long nr_uninterruptible; Sure. I changed nr_running() and nr_iowait() because they're in format strings in /proc as %lu.
