lkml.org 
[lkml]   [2002]   [Jan]   [28]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
Patch in this message
/
Date
From
Subject[patch] [sched] misc cleanups, 2.5.3-pre5

the attached patch cleans up a few more things:

- initialize the scheduler a bit earlier in start_secondary().

- mark more code with unlikely() conditions.

- make the load-balancer CONFIG_SMP. We used to run it on UP as well, but
people noticed :-)

with the unlikely() improvements schedule() is almost fall-through for the
common context-switch case, only the prev->state switch creates a branch.

with all patches applied to 2.5.3-pre5, lat_ctx context switch times
improve from 1.25 usecs to 1.12 usecs.

Ingo

diff -rNu linux/arch/i386/kernel/smpboot.c linux/arch/i386/kernel/smpboot.c
--- linux/arch/i386/kernel/smpboot.c Mon Jan 28 15:21:39 2002
+++ linux/arch/i386/kernel/smpboot.c Mon Jan 28 15:24:44 2002
@@ -456,13 +456,14 @@
*/
int __init start_secondary(void *unused)
{
+ __cli();
+ init_idle();
/*
* Dont put anything before smp_callin(), SMP
* booting is too fragile that we want to limit the
* things done here to the most necessary things.
*/
cpu_init();
- init_idle();
smp_callin();
while (!atomic_read(&smp_commenced))
rep_nop();
diff -rNu linux/include/asm-i386/mmu_context.h linux/include/asm-i386/mmu_context.h
--- linux/include/asm-i386/mmu_context.h Mon Jan 28 15:21:41 2002
+++ linux/include/asm-i386/mmu_context.h Mon Jan 28 15:24:44 2002
@@ -49,13 +51,13 @@

static inline void switch_mm(struct mm_struct *prev, struct mm_struct *next, struct task_struct *tsk, unsigned cpu)
{
- if (prev != next) {
+ if (likely(prev != next)) {
/* stop flush ipis for the previous mm */
clear_bit(cpu, &prev->cpu_vm_mask);
/*
* Re-load LDT if necessary
*/
- if (prev->context.segments != next->context.segments)
+ if (unlikely(prev->context.segments != next->context.segments))
load_LDT(next);
#ifdef CONFIG_SMP
cpu_tlbstate[cpu].state = TLBSTATE_OK;
diff -rNu linux/kernel/sched.c linux/kernel/sched.c
--- linux/kernel/sched.c Mon Jan 28 15:23:50 2002
+++ linux/kernel/sched.c Mon Jan 28 15:24:44 2002
@@ -315,14 +315,14 @@

prepare_to_switch();

- if (!mm) {
+ if (unlikely(!mm)) {
next->active_mm = oldmm;
atomic_inc(&oldmm->mm_count);
enter_lazy_tlb(oldmm, next, smp_processor_id());
} else
switch_mm(oldmm, mm, next, smp_processor_id());

- if (!prev->mm) {
+ if (unlikely(!prev->mm)) {
prev->active_mm = NULL;
mmdrop(oldmm);
}
@@ -360,6 +360,7 @@
return sum;
}

+#if CONFIG_SMP
/*
* Lock the busiest runqueue as well, this_rq is locked already.
* Recalculate nr_running if we have to drop the runqueue lock.
@@ -539,17 +555,21 @@
spin_unlock(&this_rq()->lock);
}

+#endif
+
/*
* This function gets called by the timer code, with HZ frequency.
* We call it with interrupts disabled.
*/
void scheduler_tick(task_t *p)
{
- unsigned long now = jiffies;
runqueue_t *rq = this_rq();
+#if CONFIG_SMP
+ unsigned long now = jiffies;

if (p == rq->idle)
return idle_tick();
+#endif
/* Task might have expired already, but not scheduled off yet */
if (p->array != rq->active) {
p->need_resched = 1;
@@ -595,8 +615,10 @@
enqueue_task(p, rq->active);
}
out:
+#if CONFIG_SMP
if (!(now % BUSY_REBALANCE_TICK))
load_balance(rq, 0);
+#endif
spin_unlock(&rq->lock);
}

@@ -863,9 +892,8 @@
goto out_unlock;
}
array = p->array;
- if (array) {
+ if (array)
dequeue_task(p, array);
- }
p->__nice = nice;
p->prio = NICE_TO_PRIO(nice);
if (array) {
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/

\
 
 \ /
  Last update: 2005-03-22 13:15    [W:0.023 / U:0.380 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site