Commit 1e185b97 authored by Chen, Kenneth W's avatar Chen, Kenneth W Committed by Linus Torvalds
Browse files

[PATCH] ia64: cpu_idle performance bug fix

Our performance validation on 2.6.15-rc1 caught a disastrous performance
regression on ia64 with netperf (-98%) and volanomark (-58%) compares to
previous kernel version 2.6.14-git7.  See the following chart (result
group 1 & 2).

We have root caused it to commit 64c7c8f8

This changeset broke the ia64 task resched notification.  In
sched.c:resched_task(), a reschedule IPI is conditioned upon
TIF_POLLING_NRFLAG.  However, the above changeset unconditionally set
the polling thread flag for idle tasks regardless whether pal_halt_light
is in use or not.  As a result, resched IPI is not sent from
resched_task().  And since the default behavior on ia64 is to use
pal_halt_light, we end up delaying the rescheduling task until next
timer tick, and thus cause the performance regression.

This fixes the performance bug.  I'm glad our performance suite is
turning up bad performance bug like this in time.
Signed-off-by: default avatarKen Chen <>
Signed-off-by: default avatarLinus Torvalds <>
parent 47227d50
...@@ -202,12 +202,9 @@ default_idle (void) ...@@ -202,12 +202,9 @@ default_idle (void)
{ {
local_irq_enable(); local_irq_enable();
while (!need_resched()) { while (!need_resched()) {
if (can_do_pal_halt) { if (can_do_pal_halt)
if (!need_resched())
safe_halt(); safe_halt();
local_irq_enable(); else
} else
cpu_relax(); cpu_relax();
} }
} }
...@@ -272,10 +269,14 @@ cpu_idle (void) ...@@ -272,10 +269,14 @@ cpu_idle (void)
{ {
void (*mark_idle)(int) = ia64_mark_idle; void (*mark_idle)(int) = ia64_mark_idle;
int cpu = smp_processor_id(); int cpu = smp_processor_id();
/* endless idle loop with no priority at all */ /* endless idle loop with no priority at all */
while (1) { while (1) {
if (can_do_pal_halt)
if (!need_resched()) { if (!need_resched()) {
void (*idle)(void); void (*idle)(void);
Supports Markdown
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment