On Thu, 29 Jun 2006, Christoph Lameter wrote:
> > Did you consider my earlier suggestion about these counters? That, over the
> > short-term, they tend to count in only one direction? So we can do
> Uhh... We are overcompensating right? Pretty funky idea that is new to me
> and that would require some thought.
>
> This would basically increase the stepping by 50% if we are only going in
> one direction.
A patch that does this:
ZVC: overcompensate while incrementing ZVC counters
Overcompensate by a balance factor when incrementing or decrementing
ZVC counters anticipating continual increase in the same direction.
Note that I have not been able to see any effect off this approach on
an 8p system where I tested this.
I probably will have a chance to test it on larger systems (160p) tomorrow.
Signed-off-by: Christoph Lameter <[email protected]>
Index: linux-2.6.17-mm4/mm/vmstat.c
===================================================================
--- linux-2.6.17-mm4.orig/mm/vmstat.c 2006-06-29 13:35:16.959161608 -0700
+++ linux-2.6.17-mm4/mm/vmstat.c 2006-06-29 13:54:46.361438715 -0700
@@ -167,6 +167,9 @@ EXPORT_SYMBOL(mod_zone_page_state);
* in between and therefore the atomicity vs. interrupt cannot be exploited
* in a useful way here.
*/
+
+#define balance_factor(x) (x->stat_threshold / 2)
+
static void __inc_zone_state(struct zone *zone, enum zone_stat_item item)
{
struct per_cpu_pageset *pcp = zone_pcp(zone, smp_processor_id());
@@ -175,8 +178,8 @@ static void __inc_zone_state(struct zone
(*p)++;
if (unlikely(*p > pcp->stat_threshold)) {
- zone_page_state_add(*p, zone, item);
- *p = 0;
+ zone_page_state_add(*p + balance_factor(pcp), zone, item);
+ *p = -balance_factor(pcp);
}
}
@@ -195,8 +198,8 @@ void __dec_zone_page_state(struct page *
(*p)--;
if (unlikely(*p < -pcp->stat_threshold)) {
- zone_page_state_add(*p, zone, item);
- *p = 0;
+ zone_page_state_add(*p - balance_factor(pcp), zone, item);
+ *p = balance_factor(pcp);
}
}
EXPORT_SYMBOL(__dec_zone_page_state);
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to [email protected]
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
[Index of Archives]
[Kernel Newbies]
[Netfilter]
[Bugtraq]
[Photo]
[Stuff]
[Gimp]
[Yosemite News]
[MIPS Linux]
[ARM Linux]
[Linux Security]
[Linux RAID]
[Video 4 Linux]
[Linux for the blind]
[Linux Resources]
