linux_dsm_epyc7002/arch/s390/kernel
Nick Piggin 64c7c8f885 [PATCH] sched: resched and cpu_idle rework
Make some changes to the NEED_RESCHED and POLLING_NRFLAG to reduce
confusion, and make their semantics rigid.  Improves efficiency of
resched_task and some cpu_idle routines.

* In resched_task:
- TIF_NEED_RESCHED is only cleared with the task's runqueue lock held,
  and as we hold it during resched_task, then there is no need for an
  atomic test and set there. The only other time this should be set is
  when the task's quantum expires, in the timer interrupt - this is
  protected against because the rq lock is irq-safe.

- If TIF_NEED_RESCHED is set, then we don't need to do anything. It
  won't get unset until the task get's schedule()d off.

- If we are running on the same CPU as the task we resched, then set
  TIF_NEED_RESCHED and no further action is required.

- If we are running on another CPU, and TIF_POLLING_NRFLAG is *not* set
  after TIF_NEED_RESCHED has been set, then we need to send an IPI.

Using these rules, we are able to remove the test and set operation in
resched_task, and make clear the previously vague semantics of
POLLING_NRFLAG.

* In idle routines:
- Enter cpu_idle with preempt disabled. When the need_resched() condition
  becomes true, explicitly call schedule(). This makes things a bit clearer
  (IMO), but haven't updated all architectures yet.

- Many do a test and clear of TIF_NEED_RESCHED for some reason. According
  to the resched_task rules, this isn't needed (and actually breaks the
  assumption that TIF_NEED_RESCHED is only cleared with the runqueue lock
  held). So remove that. Generally one less locked memory op when switching
  to the idle thread.

- Many idle routines clear TIF_POLLING_NRFLAG, and only set it in the inner
  most polling idle loops. The above resched_task semantics allow it to be
  set until before the last time need_resched() is checked before going into
  a halt requiring interrupt wakeup.

  Many idle routines simply never enter such a halt, and so POLLING_NRFLAG
  can be always left set, completely eliminating resched IPIs when rescheduling
  the idle task.

  POLLING_NRFLAG width can be increased, to reduce the chance of resched IPIs.

Signed-off-by: Nick Piggin <npiggin@suse.de>
Cc: Ingo Molnar <mingo@elte.hu>
Cc: Con Kolivas <kernel@kolivas.org>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>
2005-11-09 07:56:33 -08:00
..
asm-offsets.c Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
binfmt_elf32.c Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
bitmap.S Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
compat_exec_domain.c Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
compat_ioctl.c [PATCH] s390: 3270 fullscreen view 2005-10-30 17:37:16 -08:00
compat_linux.c [PATCH] x86-64: Fix 32bit sendfile 2005-09-12 10:49:57 -07:00
compat_linux.h [PATCH] compat: introduce compat_time_t 2005-06-23 09:45:32 -07:00
compat_ptrace.h Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
compat_signal.c [PATCH] s390 signal annotations 2005-09-29 08:46:27 -07:00
compat_wrapper.S [PATCH] s390: ioprio & inotify system calls. 2005-08-01 21:37:59 -07:00
cpcmd.c [PATCH] broken inline asm on s390 (misuse of labels) 2005-08-23 18:43:44 -07:00
crash.c [PATCH] kdump: Use real pt_regs from exception 2005-06-25 16:24:54 -07:00
debug.c [PATCH] s390: "extern inline" -> "static inline" 2005-11-09 07:55:52 -08:00
ebcdic.c Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
entry64.S [PATCH] s390: signal delivery 2005-11-07 07:53:32 -08:00
entry.S [PATCH] s390: signal delivery 2005-11-07 07:53:32 -08:00
head31.S [PATCH] s390: merge common parts of head.S and head64.S 2005-11-07 07:53:34 -08:00
head64.S [PATCH] s390: merge common parts of head.S and head64.S 2005-11-07 07:53:34 -08:00
head.S [PATCH] s390: merge common parts of head.S and head64.S 2005-11-07 07:53:34 -08:00
init_task.c Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
irq.c Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
machine_kexec.c [PATCH] s390: kexec fixes and improvements. 2005-08-01 21:37:59 -07:00
Makefile [PATCH] s390: merge common parts of head.S and head64.S 2005-11-07 07:53:34 -08:00
module.c Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
process.c [PATCH] sched: resched and cpu_idle rework 2005-11-09 07:56:33 -08:00
profile.c Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
ptrace.c [PATCH] s390: uml ptrace fixes 2005-06-04 17:13:00 -07:00
reipl64.S Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
reipl_diag.c [PATCH] s390: diag 0x308 reipl 2005-09-17 11:49:59 -07:00
reipl.S Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
relocate_kernel64.S [PATCH] s390: kexec fixes and improvements. 2005-08-01 21:37:59 -07:00
relocate_kernel.S [PATCH] s390: kexec fixes and improvements. 2005-08-01 21:37:59 -07:00
s390_ext.c [PATCH] s390: external call performance 2005-07-27 16:26:04 -07:00
s390_ksyms.c [PATCH] s390: default storage key 2005-05-01 08:58:58 -07:00
semaphore.c Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
setup.c [PATCH] s390: export ipl device parameters 2005-10-30 17:37:16 -08:00
signal.c [PATCH] s390 signal annotations 2005-09-29 08:46:27 -07:00
smp.c [PATCH] sched: disable preempt in idle tasks 2005-11-09 07:56:33 -08:00
sys_s390.c Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
syscalls.S [PATCH] s390: ioprio & inotify system calls. 2005-08-01 21:37:59 -07:00
time.c [PATCH] s390: stop_hz_timer vs. xtime updates 2005-11-07 07:53:32 -08:00
traps.c [PATCH] s390: remove pagex support 2005-11-07 07:53:33 -08:00
vmlinux.lds.S Linux-2.6.12-rc2 2005-04-16 15:20:36 -07:00
vtime.c [PATCH] remove timer debug field 2005-10-30 17:37:18 -08:00