Attachment 163374 Details for Bug 305162

[patch] cpufreq_fix_tickless_acounting.patch

cpufreq_fix_tickless_acounting.patch~ (text/plain), 2.56 KB, created by Thomas Renninger on 2007-09-11 22:04:30 UTC

(hide)

Description:

Filename:

MIME Type:

Creator: Thomas Renninger

Created: 2007-09-11 22:04:30 UTC

Size: 2.56 KB

patch

obsolete

>commit ea48761519bd40d7a881c587b5f3177664b2987e
>Author: Venki Pallipadi <venkatesh.pallipadi@intel.com>
>Date:   Wed Jun 20 14:26:24 2007 -0700
>
>    [CPUFREQ] ondemand: fix tickless accounting and software coordination bug
>    
>    With tickless kernel and software coordination os P-states, ondemand
>    can look at wrong idle statistics. This can happen when ondemand sampling
>    is happening on CPU 0 and due to software coordination sampling also looks at
>    utilization of CPU 1. If CPU 1 is in tickless state at that moment, its idle
>    statistics will not be uptodate and CPU 0 thinks CPU 1 is idle for less
>    amount of time than it actually is.
>    
>    This can be resolved by looking at all the busy times of CPUs, which is
>    accurate, even with tickless, and use that to determine idle time in a
>    round about way (total time - busy time).
>    
>    Thanks to Arjan for originally reporting the ondemand bug on
>    Lenovo T61.
>    
>    Signed-off-by: Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>
>    Signed-off-by: Dave Jones <davej@redhat.com>
>
>diff --git a/drivers/cpufreq/cpufreq_ondemand.c b/drivers/cpufreq/cpufreq_ondemand.c
>index dc6f357..e794527 100644
>--- a/drivers/cpufreq/cpufreq_ondemand.c
>+++ b/drivers/cpufreq/cpufreq_ondemand.c
>@@ -96,15 +96,25 @@ static struct dbs_tuners {
> 
> static inline cputime64_t get_cpu_idle_time(unsigned int cpu)
> {
>-	cputime64_t retval;
>+	cputime64_t idle_time;
>+	cputime64_t cur_jiffies;
>+	cputime64_t busy_time;
> 
>-	retval = cputime64_add(kstat_cpu(cpu).cpustat.idle,
>-			kstat_cpu(cpu).cpustat.iowait);
>+	cur_jiffies = jiffies64_to_cputime64(get_jiffies_64());
>+	busy_time = cputime64_add(kstat_cpu(cpu).cpustat.user,
>+			kstat_cpu(cpu).cpustat.system);
> 
>-	if (dbs_tuners_ins.ignore_nice)
>-		retval = cputime64_add(retval, kstat_cpu(cpu).cpustat.nice);
>+	busy_time = cputime64_add(busy_time, kstat_cpu(cpu).cpustat.irq);
>+	busy_time = cputime64_add(busy_time, kstat_cpu(cpu).cpustat.softirq);
>+	busy_time = cputime64_add(busy_time, kstat_cpu(cpu).cpustat.steal);
> 
>-	return retval;
>+	if (!dbs_tuners_ins.ignore_nice) {
>+		busy_time = cputime64_add(busy_time,
>+				kstat_cpu(cpu).cpustat.nice);
>+	}
>+
>+	idle_time = cputime64_sub(cur_jiffies, busy_time);
>+	return idle_time;
> }
> 
> /*
>@@ -339,7 +349,8 @@ static void dbs_check_cpu(struct cpu_dbs_info_s *this_dbs_info)
> 	cur_jiffies = jiffies64_to_cputime64(get_jiffies_64());
> 	total_ticks = (unsigned int) cputime64_sub(cur_jiffies,
> 			this_dbs_info->prev_cpu_wall);
>-	this_dbs_info->prev_cpu_wall = cur_jiffies;
>+	this_dbs_info->prev_cpu_wall = get_jiffies_64();
>+
> 	if (!total_ticks)
> 		return;
> 	/*

Actions: View | Diff

Attachments on bug 305162: 160171 | 161747 | 162896 | 163373 | 163374 | 163900 | 174436