cpufreq_conservative/ondemand: invert meaning of 'ignore nice'

The use of the 'ignore_nice' sysfs file is confusing to anyone using it. This removes the sysfs file 'ignore_nice' and in its place creates a 'ignore_nice_load' entry which defaults to '1'; meaning nice'd processes are not counted towards the 'business' caclulation.

WARNING: this obvious breaks any userland tools that expected 'ignore_nice' to exist, to draw attention to this fact it was concluded on the mailing list that the entry should be removed altogether so the userland app breaks and so the author can build simple to detect workaround. Having said that it seems currently very few tools even make use of this functionality; all I could find was a Gentoo Wiki entry.

diff -r -u -d linux-2.6.14-rc2.orig/drivers/cpufreq/cpufreq_conservative.c \
                linux-2.6.14-rc2/drivers/cpufreq/cpufreq_conservative.c
--- linux-2.6.14-rc2.orig/drivers/cpufreq/cpufreq_conservative.c	2005-10-03 \
                20:05:30.742334750 +0100

+++ linux-2.6.14-rc2/drivers/cpufreq/cpufreq_conservative.c 2005-10-06 \
21:10:47.785133750 +0100 @@ -93,7 +93,7 @@
{
return kstat_cpu(cpu).cpustat.idle +
kstat_cpu(cpu).cpustat.iowait +
- ( !dbs_tuners_ins.ignore_nice ?
+ ( dbs_tuners_ins.ignore_nice ?
kstat_cpu(cpu).cpustat.nice :
0);
}
@@ -127,7 +127,7 @@
show_one(sampling_down_factor, sampling_down_factor); show_one(up_threshold, up_threshold); show_one(down_threshold, down_threshold);
-show_one(ignore_nice, ignore_nice); +show_one(ignore_nice_load, ignore_nice);
show_one(freq_step, freq_step);

static ssize_t store_sampling_down_factor(struct cpufreq_policy *unused,
@@ -207,7 +207,7 @@
return count;
}

-static ssize_t store_ignore_nice(struct cpufreq_policy *policy, +static ssize_t store_ignore_nice_load(struct cpufreq_policy *policy,
const char *buf, size_t count)
{
unsigned int input;
@@ -272,7 +272,7 @@
define_one_rw(sampling_down_factor); define_one_rw(up_threshold); define_one_rw(down_threshold);
-define_one_rw(ignore_nice); +define_one_rw(ignore_nice_load);
define_one_rw(freq_step);

static struct attribute * dbs_attributes[] = {
@@ -282,7 +282,7 @@
&sampling_down_factor.attr,
&up_threshold.attr,
&down_threshold.attr,
- &ignore_nice.attr,
+ &ignore_nice_load.attr,
&freq_step.attr,
NULL
};
@@ -515,7 +515,7 @@
def_sampling_rate = (latency / 1000) *
DEF_SAMPLING_RATE_LATENCY_MULTIPLIER;
dbs_tuners_ins.sampling_rate = def_sampling_rate;
- dbs_tuners_ins.ignore_nice = 0;
+ dbs_tuners_ins.ignore_nice = 1;
dbs_tuners_ins.freq_step = 5;

dbs_timer_init();

diff -r -u -d linux-2.6.14-rc2.orig/drivers/cpufreq/cpufreq_ondemand.c \
                linux-2.6.14-rc2/drivers/cpufreq/cpufreq_ondemand.c
--- linux-2.6.14-rc2.orig/drivers/cpufreq/cpufreq_ondemand.c	2005-10-03 \
                20:05:30.742334750 +0100

+++ linux-2.6.14-rc2/drivers/cpufreq/cpufreq_ondemand.c 2005-10-06 21:10:23.979646000 \
+0100 @@ -86,7 +86,7 @@
{
return kstat_cpu(cpu).cpustat.idle +
kstat_cpu(cpu).cpustat.iowait +
- ( !dbs_tuners_ins.ignore_nice ?
+ ( dbs_tuners_ins.ignore_nice ?
kstat_cpu(cpu).cpustat.nice :
0);
}
@@ -119,7 +119,7 @@
show_one(sampling_rate, sampling_rate); show_one(sampling_down_factor, sampling_down_factor); show_one(up_threshold, up_threshold);
-show_one(ignore_nice, ignore_nice); +show_one(ignore_nice_load, ignore_nice);

static ssize_t store_sampling_down_factor(struct cpufreq_policy *unused,
const char *buf, size_t count)
@@ -179,7 +179,7 @@
return count;
}

-static ssize_t store_ignore_nice(struct cpufreq_policy *policy, +static ssize_t store_ignore_load_tasks(struct cpufreq_policy *policy,
const char *buf, size_t count)
{
unsigned int input;
@@ -220,7 +220,7 @@
define_one_rw(sampling_rate); define_one_rw(sampling_down_factor); define_one_rw(up_threshold);
-define_one_rw(ignore_nice); +define_one_rw(ignore_nice_load);

static struct attribute * dbs_attributes[] = {
&sampling_rate_max.attr,
@@ -228,7 +228,7 @@
&sampling_rate.attr,
&sampling_down_factor.attr,
&up_threshold.attr,
- &ignore_nice.attr,
+ &ignore_nice_load.attr,
NULL
};

@@ -424,7 +424,7 @@
def_sampling_rate = (latency / 1000) *
DEF_SAMPLING_RATE_LATENCY_MULTIPLIER;
dbs_tuners_ins.sampling_rate = def_sampling_rate;
- dbs_tuners_ins.ignore_nice = 0;
+ dbs_tuners_ins.ignore_nice = 1;

dbs_timer_init();
}

And just for the last time I'll argue that the default should be 0. I have yet to discuss this with any laptop user who thinks that 1 is the correct default for ondemand.

Regards,
Con

Con,

Con Kolivas <[email protected]> [20051111 02:48:57 +1100]:

....resubmitting with alternative defaults....

Cheers

Alex

 ____________________________________
/ "An ounce of prevention is worth a \
\ pound of purge."                   /
 ------------------------------------
        \   ^__^
         \  (oo)\_______
            (__)\       )\/\
                ||----w |
                ||     ||

Hi!

Me. I have graphics appp here (almara), that does user interaction in separate thread from real workers. Yet you want real workers to run...

And consider notebook on ac power, using ondemand for acoustic management.

64 bytes from 195.113.31.123: icmp_seq=28 ttl=51 time=448769.1 ms         

i think that 1 is the correct default for ondemand.
And i know that discussion is fruitless - everybody has its own preference, i prefer battery runtime before almost everything else :-)

Stefan Seyfried                  \ "I didn't want to write for pay. I
QA / R&D Team Mobile Devices      \ wanted to be paid for what I write."
SUSE LINUX Products GmbH, Nürnberg \                    -- Leonard Cohen

This patch is horribly word-wrapped. Please resend.

Dave