Groups > linux.kernel > #1256993 > unrolled thread

[PATCH 01/70] cpufreq: interactive: New 'interactive' governor

Started by	Bálint Czobor <czoborbalint@gmail.com>
First post	2015-10-27 18:40 +0100
Last post	2015-10-29 02:50 +0100
Articles	20 on this page of 72 — 3 participants

Back to article view | Back to linux.kernel

  [PATCH 01/70] cpufreq: interactive: New 'interactive' governor Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 57/70] cpufreq: interactive: fix compiling warnings Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 56/70] cpufreq: interactive: delete timers for GOV_START Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 41/70] cpufreq: interactive: fix race on governor start/stop Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 32/70] cpufreq: interactive: specify duration of CPU speed boost pulse Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 43/70] cpufreq: interactive: add io_is_busy interface Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 39/70] cpufreq: interactive: don't handle transition notification if not enabled Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 51/70] cpufreq: interactive: resched timer if max freq raised Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 09/70] cpufreq: interactive: Separate speed target revalidate time and initial set time Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 60/70] cpufreq: interactive: hold reference on global cpufreq kobject if needed Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 24/70] cpufreq: interactive: use deferrable timer by default Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 61/70] cpufreq: interactive: restructure CPUFREQ_GOV_LIMITS Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 27/70] cpufreq: interactive: change speed according to current speed and target load Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 07/70] cpufreq: interactive: adjust code and documentation to match Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 29/70] cpufreq: interactive: allow arbitrary speed / target load mappings Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 48/70] cpufreq: interactive: reduce chance of zero time delta on load eval Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 64/70] cpufreq: interactive: prevents the frequency to directly raise above the hispeed_freq from a lower frequency. Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 65/70] cpufreq: interactive: make common_tunables static Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 69/70] cpufreq: interactive: Don't set floor_validate_time during boost Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 37/70] cpufreq: interactive: default go_hispeed_load 99%, doc updates Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 58/70] cpufreq: interactive: fix NULL pointer dereference at sysfs ops Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 66/70] cpufreq: interactive: Fix compile errors in accordance with changes from 3.14 to 3.18 Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 17/70] cpufreq: interactive: take idle notifications only when active Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 35/70] cpufreq: interactive: fix racy timer stopping Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 67/70] cpufreq: interactive: Put global cpufreq kobject on failure Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 70/70] cpufreq: interactive: Round up timer_rate to match jiffy Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 20/70] cpufreq: interactive: remove input_boost handling Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 12/70] cpufreq: interactive: Add sysfs boost interface for hints from userspace Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 22/70] cpufreq: interactive: run at fraction of hispeed_freq when load is low Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 68/70] subsystem: CPU FREQUENCY DRIVERS- Set cpu_load calculation on current frequency Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 13/70] cpufreq: interactive: set floor for boosted speed Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 55/70] cpufreq: Interactive: Implement per policy instances of governor Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 52/70] cpufreq: interactive: fix show_target_loads and show_above_hispeed_delay Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 10/70] cpufreq: interactive: Boost frequency on touchscreen input Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:40 +0100
    [PATCH 62/70] cpufreq: interactive: turn boost_pulse off on boost off Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:50 +0100
    [PATCH 59/70] cpufreq: interactive: Use generic get_cpu_idle_time() from cpufreq.c Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:50 +0100
    [PATCH 54/70] cpufreq: interactive: Move definition of cpufreq_gov_interactive downwards Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:50 +0100
    [PATCH 53/70] cpufreq: interactive: Remove unnecessary cpu_online() check Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:50 +0100
    [PATCH 25/70] cpufreq: interactive: kick timer on idle exit past expiry Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:50 +0100
    [PATCH 47/70] cpufreq: interactive: handle errors from cpufreq_frequency_table_target Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:50 +0100
    [PATCH 63/70] cpufreq: interactive: remove compilation error from commit 49cc72365fb7ee87762a7ccc6a32ef68627216c5 Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:50 +0100
    [PATCH 44/70] cpufreq: interactive: fix crash on error paths in get_tokenized_data Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:50 +0100
    [PATCH 50/70] cpufreq: interactive: fix race on cpufreq TRANSITION notifier Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:50 +0100
    [PATCH 49/70] cpufreq: interactive: avoid underflow on active time calculation Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:50 +0100
    [PATCH 45/70] cpufreq: interactive: base above_hispeed_delay on target freq, not current Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:50 +0100
    [PATCH 46/70] cpufreq: interactive: fix uninitialized spinlock Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 18:50 +0100
    [PATCH 34/70] cpufreq: interactive: fix boosting logic Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 19:00 +0100
    [PATCH 38/70] cpufreq: interactive: init default values at compile time Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 19:00 +0100
    [PATCH 16/70] cpufreq: interactive: restart above_hispeed_delay at each hispeed load Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 19:00 +0100
    [PATCH 28/70] cpufreq: interactive: apply above_hispeed_delay to each step above hispeed Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 19:00 +0100
    [PATCH 23/70] cpufreq: interactive: pin timers to associated CPU Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 19:00 +0100
    [PATCH 18/70] cpufreq: interactive: keep freezer happy when not current governor Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 19:00 +0100
    [PATCH 21/70] cpufreq: interactive: always limit initial speed bump to hispeed Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 19:00 +0100
    [PATCH 36/70] cpufreq: interactive: fix race on timer restart on governor start Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 19:00 +0100
    [PATCH 19/70] cpufreq: interactive: handle speed up and down in the realtime task Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 19:00 +0100
    [PATCH 42/70] cpufreq: interactive: allow arbitrary speed / delay mappings Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 19:00 +0100
    [PATCH 26/70] cpufreq: interactive: trace actual speed in target speed decisions Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 19:00 +0100
    [PATCH 30/70] cpufreq: interactive: remove load since last speed change Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 19:00 +0100
    [PATCH 40/70] cpufreq: interactive: fix deadlock on spinlock in timer Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 19:00 +0100
    [PATCH 33/70] cpufreq: interactive: add timer slack to limit idle at speed > min Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 19:00 +0100
    [PATCH 31/70] cpufreq: interactive: adjust load for changes in speed Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 19:00 +0100
    [PATCH 04/70] cpufreq: interactive: set at least hispeed when above hispeed load Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 19:10 +0100
    [PATCH 02/70] cpufreq interactive governor: event tracing Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 19:10 +0100
    [PATCH 11/70] cpufreq: interactive: remove unused target_validate_time_in_idle Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 19:10 +0100
    [PATCH 05/70] cpufreq: interactive: don't drop speed if recently at higher load Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 19:10 +0100
    [PATCH 06/70] cpufreq: interactive: configurable delay before raising above hispeed Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 19:10 +0100
    [PATCH 08/70] cpufreq: interactive: base hispeed bump on target freq, not actual Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 19:10 +0100
    [PATCH 15/70] cpufreq-interactive: Compile fixup Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 19:10 +0100
    [PATCH 14/70] cpufreq: interactive: add boost pulse interface Bálint Czobor <czoborbalint@gmail.com> - 2015-10-27 19:10 +0100
    Re: [PATCH 01/70] cpufreq: interactive: New 'interactive' governor "Rafael J. Wysocki" <rjw@rjwysocki.net> - 2015-10-28 01:40 +0100
      Re: [PATCH 01/70] cpufreq: interactive: New 'interactive' governor Viresh Kumar <viresh.kumar@linaro.org> - 2015-10-28 04:10 +0100
        Re: [PATCH 01/70] cpufreq: interactive: New 'interactive' governor Viresh Kumar <viresh.kumar@linaro.org> - 2015-10-29 02:50 +0100

Page 1 of 4 [1] 2 3 4 Next page →

#1256993 — [PATCH 01/70] cpufreq: interactive: New 'interactive' governor

From	Bálint Czobor <czoborbalint@gmail.com>
Date	2015-10-27 18:40 +0100
Subject	[PATCH 01/70] cpufreq: interactive: New 'interactive' governor
Message-ID	<qogo1-84h-5@gated-at.bofh.it>

From: Mike Chan <mike@android.com>

This governor is designed for latency-sensitive workloads, such as
interactive user interfaces.  The interactive governor aims to be
significantly more responsive to ramp CPU quickly up when CPU-intensive
activity begins.

Existing governors sample CPU load at a particular rate, typically
every X ms.  This can lead to under-powering UI threads for the period of
time during which the user begins interacting with a previously-idle system
until the next sample period happens.

The 'interactive' governor uses a different approach. Instead of sampling
the CPU at a specified rate, the governor will check whether to scale the
CPU frequency up soon after coming out of idle.  When the CPU comes out of
idle, a timer is configured to fire within 1-2 ticks.  If the CPU is very
busy from exiting idle to when the timer fires then we assume the CPU is
underpowered and ramp to MAX speed.

If the CPU was not sufficiently busy to immediately ramp to MAX speed, then
the governor evaluates the CPU load since the last speed adjustment,
choosing the highest value between that longer-term load or the short-term
load since idle exit to determine the CPU speed to ramp to.

A realtime thread is used for scaling up, giving the remaining tasks the
CPU performance benefit, unlike existing governors which are more likely to
schedule rampup work to occur after your performance starved tasks have
completed.

The tuneables for this governor are:
/sys/devices/system/cpu/cpufreq/interactive/min_sample_time:
	The minimum amount of time to spend at the current frequency before
	ramping down. This is to ensure that the governor has seen enough
	historic CPU load data to determine the appropriate workload.
	Default is 80000 uS.
/sys/devices/system/cpu/cpufreq/interactive/go_maxspeed_load
	The CPU load at which to ramp to max speed.  Default is 85.

Change-Id: Ib2b362607c62f7c56d35f44a9ef3280f98c17585
Signed-off-by: Mike Chan <mike@android.com>
Signed-off-by: Todd Poynor <toddpoynor@google.com>
Bug: 3152864
Signed-off-by: Bálint Czobor <czoborbalint@gmail.com>
---
 Documentation/cpu-freq/governors.txt  |   37 ++
 drivers/cpufreq/Kconfig               |   27 ++
 drivers/cpufreq/Makefile              |    1 +
 drivers/cpufreq/cpufreq_interactive.c |  706 +++++++++++++++++++++++++++++++++
 include/linux/cpufreq.h               |    3 +
 5 files changed, 774 insertions(+)
 create mode 100644 drivers/cpufreq/cpufreq_interactive.c

diff --git a/Documentation/cpu-freq/governors.txt b/Documentation/cpu-freq/governors.txt
index c15aa75..b262c53 100644
--- a/Documentation/cpu-freq/governors.txt
+++ b/Documentation/cpu-freq/governors.txt
@@ -28,6 +28,7 @@ Contents:
 2.3  Userspace
 2.4  Ondemand
 2.5  Conservative
+2.6  Interactive
 
 3.   The Governor Interface in the CPUfreq Core
 
@@ -218,6 +219,42 @@ a decision on when to decrease the frequency while running in any
 speed. Load for frequency increase is still evaluated every
 sampling rate.
 
+2.6 Interactive
+---------------
+
+The CPUfreq governor "interactive" is designed for latency-sensitive,
+interactive workloads. This governor sets the CPU speed depending on
+usage, similar to "ondemand" and "conservative" governors.  However,
+the governor is more aggressive about scaling the CPU speed up in
+response to CPU-intensive activity.
+
+Sampling the CPU load every X ms can lead to under-powering the CPU
+for X ms, leading to dropped frames, stuttering UI, etc.  Instead of
+sampling the cpu at a specified rate, the interactive governor will
+check whether to scale the cpu frequency up soon after coming out of
+idle.  When the cpu comes out of idle, a timer is configured to fire
+within 1-2 ticks.  If the cpu is very busy between exiting idle and
+when the timer fires then we assume the cpu is underpowered and ramp
+to MAX speed.
+
+If the cpu was not sufficiently busy to immediately ramp to MAX speed,
+then governor evaluates the cpu load since the last speed adjustment,
+choosing the highest value between that longer-term load or the
+short-term load since idle exit to determine the cpu speed to ramp to.
+
+The tuneable values for this governor are:
+
+min_sample_time: The minimum amount of time to spend at the current
+frequency before ramping down. This is to ensure that the governor has
+seen enough historic cpu load data to determine the appropriate
+workload.  Default is 80000 uS.
+
+go_maxspeed_load: The CPU load at which to ramp to max speed.  Default
+is 85.
+
+timer_rate: Sample rate for reevaluating cpu load when the system is
+not idle.  Default is 30000 uS.
+
 3. The Governor Interface in the CPUfreq Core
 =============================================
 
diff --git a/drivers/cpufreq/Kconfig b/drivers/cpufreq/Kconfig
index 659879a..6e099e5 100644
--- a/drivers/cpufreq/Kconfig
+++ b/drivers/cpufreq/Kconfig
@@ -102,6 +102,16 @@ config CPU_FREQ_DEFAULT_GOV_CONSERVATIVE
 	  Be aware that not all cpufreq drivers support the conservative
 	  governor. If unsure have a look at the help section of the
 	  driver. Fallback governor will be the performance governor.
+
+config CPU_FREQ_DEFAULT_GOV_INTERACTIVE
+	bool "interactive"
+	select CPU_FREQ_GOV_INTERACTIVE
+	help
+	  Use the CPUFreq governor 'interactive' as default. This allows
+	  you to get a full dynamic cpu frequency capable system by simply
+	  loading your cpufreq low-level hardware driver, using the
+	  'interactive' governor for latency-sensitive workloads.
+
 endchoice
 
 config CPU_FREQ_GOV_PERFORMANCE
@@ -159,6 +169,23 @@ config CPU_FREQ_GOV_ONDEMAND
 
 	  If in doubt, say N.
 
+config CPU_FREQ_GOV_INTERACTIVE
+	tristate "'interactive' cpufreq policy governor"
+	help
+	  'interactive' - This driver adds a dynamic cpufreq policy governor
+	  designed for latency-sensitive workloads.
+
+	  This governor attempts to reduce the latency of clock
+	  increases so that the system is more responsive to
+	  interactive workloads.
+
+	  To compile this driver as a module, choose M here: the
+	  module will be called cpufreq_interactive.
+
+	  For details, take a look at linux/Documentation/cpu-freq.
+
+	  If in doubt, say N.
+
 config CPU_FREQ_GOV_CONSERVATIVE
 	tristate "'conservative' cpufreq governor"
 	depends on CPU_FREQ
diff --git a/drivers/cpufreq/Makefile b/drivers/cpufreq/Makefile
index 4134038..98667c8 100644
--- a/drivers/cpufreq/Makefile
+++ b/drivers/cpufreq/Makefile
@@ -11,6 +11,7 @@ obj-$(CONFIG_CPU_FREQ_GOV_POWERSAVE)	+= cpufreq_powersave.o
 obj-$(CONFIG_CPU_FREQ_GOV_USERSPACE)	+= cpufreq_userspace.o
 obj-$(CONFIG_CPU_FREQ_GOV_ONDEMAND)	+= cpufreq_ondemand.o
 obj-$(CONFIG_CPU_FREQ_GOV_CONSERVATIVE)	+= cpufreq_conservative.o
+obj-$(CONFIG_CPU_FREQ_GOV_INTERACTIVE)	+= cpufreq_interactive.o
 obj-$(CONFIG_CPU_FREQ_GOV_COMMON)		+= cpufreq_governor.o
 
 obj-$(CONFIG_CPUFREQ_DT)		+= cpufreq-dt.o
diff --git a/drivers/cpufreq/cpufreq_interactive.c b/drivers/cpufreq/cpufreq_interactive.c
new file mode 100644
index 0000000..8fc147c
--- /dev/null
+++ b/drivers/cpufreq/cpufreq_interactive.c
@@ -0,0 +1,706 @@
+/*
+ * drivers/cpufreq/cpufreq_interactive.c
+ *
+ * Copyright (C) 2010 Google, Inc.
+ *
+ * This software is licensed under the terms of the GNU General Public
+ * License version 2, as published by the Free Software Foundation, and
+ * may be copied, distributed, and modified under those terms.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * Author: Mike Chan (mike@android.com)
+ *
+ */
+
+#include <linux/cpu.h>
+#include <linux/cpumask.h>
+#include <linux/cpufreq.h>
+#include <linux/module.h>
+#include <linux/mutex.h>
+#include <linux/sched.h>
+#include <linux/sched/rt.h>
+#include <linux/tick.h>
+#include <linux/time.h>
+#include <linux/timer.h>
+#include <linux/workqueue.h>
+#include <linux/kthread.h>
+#include <linux/mutex.h>
+
+#include <asm/cputime.h>
+
+static atomic_t active_count = ATOMIC_INIT(0);
+
+struct cpufreq_interactive_cpuinfo {
+	struct timer_list cpu_timer;
+	int timer_idlecancel;
+	u64 time_in_idle;
+	u64 idle_exit_time;
+	u64 timer_run_time;
+	int idling;
+	u64 freq_change_time;
+	u64 freq_change_time_in_idle;
+	struct cpufreq_policy *policy;
+	struct cpufreq_frequency_table *freq_table;
+	unsigned int target_freq;
+	int governor_enabled;
+};
+
+static DEFINE_PER_CPU(struct cpufreq_interactive_cpuinfo, cpuinfo);
+
+/* Workqueues handle frequency scaling */
+static struct task_struct *up_task;
+static struct workqueue_struct *down_wq;
+static struct work_struct freq_scale_down_work;
+static cpumask_t up_cpumask;
+static spinlock_t up_cpumask_lock;
+static cpumask_t down_cpumask;
+static spinlock_t down_cpumask_lock;
+static struct mutex set_speed_lock;
+
+/* Hi speed to bump to from lo speed when load burst (default max) */
+static u64 hispeed_freq;
+
+/* Go to hi speed when CPU load at or above this value. */
+#define DEFAULT_GO_HISPEED_LOAD 95
+static unsigned long go_hispeed_load;
+
+/*
+ * The minimum amount of time to spend at a frequency before we can ramp down.
+ */
+#define DEFAULT_MIN_SAMPLE_TIME 20 * USEC_PER_MSEC
+static unsigned long min_sample_time;
+
+/*
+ * The sample rate of the timer used to increase frequency
+ */
+#define DEFAULT_TIMER_RATE 20 * USEC_PER_MSEC
+static unsigned long timer_rate;
+
+static int cpufreq_governor_interactive(struct cpufreq_policy *policy,
+		unsigned int event);
+
+#ifndef CONFIG_CPU_FREQ_DEFAULT_GOV_INTERACTIVE
+static
+#endif
+struct cpufreq_governor cpufreq_gov_interactive = {
+	.name = "interactive",
+	.governor = cpufreq_governor_interactive,
+	.max_transition_latency = 10000000,
+	.owner = THIS_MODULE,
+};
+
+static void cpufreq_interactive_timer(unsigned long data)
+{
+	unsigned int delta_idle;
+	unsigned int delta_time;
+	int cpu_load;
+	int load_since_change;
+	u64 time_in_idle;
+	u64 idle_exit_time;
+	struct cpufreq_interactive_cpuinfo *pcpu =
+		&per_cpu(cpuinfo, data);
+	u64 now_idle;
+	unsigned int new_freq;
+	unsigned int index;
+	unsigned long flags;
+
+	smp_rmb();
+
+	if (!pcpu->governor_enabled)
+		goto exit;
+
+	/*
+	 * Once pcpu->timer_run_time is updated to >= pcpu->idle_exit_time,
+	 * this lets idle exit know the current idle time sample has
+	 * been processed, and idle exit can generate a new sample and
+	 * re-arm the timer.  This prevents a concurrent idle
+	 * exit on that CPU from writing a new set of info at the same time
+	 * the timer function runs (the timer function can't use that info
+	 * until more time passes).
+	 */
+	time_in_idle = pcpu->time_in_idle;
+	idle_exit_time = pcpu->idle_exit_time;
+	now_idle = get_cpu_idle_time_us(data, &pcpu->timer_run_time);
+	smp_wmb();
+
+	/* If we raced with cancelling a timer, skip. */
+	if (!idle_exit_time)
+		goto exit;
+
+	delta_idle = (unsigned int)(now_idle - time_in_idle);
+	delta_time = (unsigned int)(pcpu->timer_run_time - idle_exit_time);
+
+	/*
+	 * If timer ran less than 1ms after short-term sample started, retry.
+	 */
+	if (delta_time < 1000)
+		goto rearm;
+
+	if (delta_idle > delta_time)
+		cpu_load = 0;
+	else
+		cpu_load = 100 * (delta_time - delta_idle) / delta_time;
+
+	delta_idle = (unsigned int)(now_idle - pcpu->freq_change_time_in_idle);
+	delta_time = (unsigned int)(pcpu->timer_run_time - pcpu->freq_change_time);
+
+	if ((delta_time == 0) || (delta_idle > delta_time))
+		load_since_change = 0;
+	else
+		load_since_change =
+			100 * (delta_time - delta_idle) / delta_time;
+
+	/*
+	 * Choose greater of short-term load (since last idle timer
+	 * started or timer function re-armed itself) or long-term load
+	 * (since last frequency change).
+	 */
+	if (load_since_change > cpu_load)
+		cpu_load = load_since_change;
+
+	if (cpu_load >= go_hispeed_load) {
+		if (pcpu->policy->cur == pcpu->policy->min)
+			new_freq = hispeed_freq;
+		else
+			new_freq = pcpu->policy->max * cpu_load / 100;
+	} else {
+		new_freq = pcpu->policy->cur * cpu_load / 100;
+	}
+
+	if (cpufreq_frequency_table_target(pcpu->policy, pcpu->freq_table,
+					   new_freq, CPUFREQ_RELATION_H,
+					   &index)) {
+		pr_warn_once("timer %d: cpufreq_frequency_table_target error\n",
+			     (int) data);
+		goto rearm;
+	}
+
+	new_freq = pcpu->freq_table[index].frequency;
+
+	if (pcpu->target_freq == new_freq)
+		goto rearm_if_notmax;
+
+	/*
+	 * Do not scale down unless we have been at this frequency for the
+	 * minimum sample time.
+	 */
+	if (new_freq < pcpu->target_freq) {
+		if (pcpu->timer_run_time - pcpu->freq_change_time
+		    < min_sample_time)
+			goto rearm;
+	}
+
+	if (new_freq < pcpu->target_freq) {
+		pcpu->target_freq = new_freq;
+		spin_lock_irqsave(&down_cpumask_lock, flags);
+		cpumask_set_cpu(data, &down_cpumask);
+		spin_unlock_irqrestore(&down_cpumask_lock, flags);
+		queue_work(down_wq, &freq_scale_down_work);
+	} else {
+		pcpu->target_freq = new_freq;
+		spin_lock_irqsave(&up_cpumask_lock, flags);
+		cpumask_set_cpu(data, &up_cpumask);
+		spin_unlock_irqrestore(&up_cpumask_lock, flags);
+		wake_up_process(up_task);
+	}
+
+rearm_if_notmax:
+	/*
+	 * Already set max speed and don't see a need to change that,
+	 * wait until next idle to re-evaluate, don't need timer.
+	 */
+	if (pcpu->target_freq == pcpu->policy->max)
+		goto exit;
+
+rearm:
+	if (!timer_pending(&pcpu->cpu_timer)) {
+		/*
+		 * If already at min: if that CPU is idle, don't set timer.
+		 * Else cancel the timer if that CPU goes idle.  We don't
+		 * need to re-evaluate speed until the next idle exit.
+		 */
+		if (pcpu->target_freq == pcpu->policy->min) {
+			smp_rmb();
+
+			if (pcpu->idling)
+				goto exit;
+
+			pcpu->timer_idlecancel = 1;
+		}
+
+		pcpu->time_in_idle = get_cpu_idle_time_us(
+			data, &pcpu->idle_exit_time);
+		mod_timer(&pcpu->cpu_timer,
+			  jiffies + usecs_to_jiffies(timer_rate));
+	}
+
+exit:
+	return;
+}
+
+static void cpufreq_interactive_idle_start(void)
+{
+	struct cpufreq_interactive_cpuinfo *pcpu =
+		&per_cpu(cpuinfo, smp_processor_id());
+	int pending;
+
+	if (!pcpu->governor_enabled)
+		return;
+
+	pcpu->idling = 1;
+	smp_wmb();
+	pending = timer_pending(&pcpu->cpu_timer);
+
+	if (pcpu->target_freq != pcpu->policy->min) {
+#ifdef CONFIG_SMP
+		/*
+		 * Entering idle while not at lowest speed.  On some
+		 * platforms this can hold the other CPU(s) at that speed
+		 * even though the CPU is idle. Set a timer to re-evaluate
+		 * speed so this idle CPU doesn't hold the other CPUs above
+		 * min indefinitely.  This should probably be a quirk of
+		 * the CPUFreq driver.
+		 */
+		if (!pending) {
+			pcpu->time_in_idle = get_cpu_idle_time_us(
+				smp_processor_id(), &pcpu->idle_exit_time);
+			pcpu->timer_idlecancel = 0;
+			mod_timer(&pcpu->cpu_timer,
+				  jiffies + usecs_to_jiffies(timer_rate));
+		}
+#endif
+	} else {
+		/*
+		 * If at min speed and entering idle after load has
+		 * already been evaluated, and a timer has been set just in
+		 * case the CPU suddenly goes busy, cancel that timer.  The
+		 * CPU didn't go busy; we'll recheck things upon idle exit.
+		 */
+		if (pending && pcpu->timer_idlecancel) {
+			del_timer(&pcpu->cpu_timer);
+			/*
+			 * Ensure last timer run time is after current idle
+			 * sample start time, so next idle exit will always
+			 * start a new idle sampling period.
+			 */
+			pcpu->idle_exit_time = 0;
+			pcpu->timer_idlecancel = 0;
+		}
+	}
+
+}
+
+static void cpufreq_interactive_idle_end(void)
+{
+	struct cpufreq_interactive_cpuinfo *pcpu =
+		&per_cpu(cpuinfo, smp_processor_id());
+
+	pcpu->idling = 0;
+	smp_wmb();
+
+	/*
+	 * Arm the timer for 1-2 ticks later if not already, and if the timer
+	 * function has already processed the previous load sampling
+	 * interval.  (If the timer is not pending but has not processed
+	 * the previous interval, it is probably racing with us on another
+	 * CPU.  Let it compute load based on the previous sample and then
+	 * re-arm the timer for another interval when it's done, rather
+	 * than updating the interval start time to be "now", which doesn't
+	 * give the timer function enough time to make a decision on this
+	 * run.)
+	 */
+	if (timer_pending(&pcpu->cpu_timer) == 0 &&
+	    pcpu->timer_run_time >= pcpu->idle_exit_time &&
+	    pcpu->governor_enabled) {
+		pcpu->time_in_idle =
+			get_cpu_idle_time_us(smp_processor_id(),
+					     &pcpu->idle_exit_time);
+		pcpu->timer_idlecancel = 0;
+		mod_timer(&pcpu->cpu_timer,
+			  jiffies + usecs_to_jiffies(timer_rate));
+	}
+
+}
+
+static int cpufreq_interactive_up_task(void *data)
+{
+	unsigned int cpu;
+	cpumask_t tmp_mask;
+	unsigned long flags;
+	struct cpufreq_interactive_cpuinfo *pcpu;
+
+	while (1) {
+		set_current_state(TASK_INTERRUPTIBLE);
+		spin_lock_irqsave(&up_cpumask_lock, flags);
+
+		if (cpumask_empty(&up_cpumask)) {
+			spin_unlock_irqrestore(&up_cpumask_lock, flags);
+			schedule();
+
+			if (kthread_should_stop())
+				break;
+
+			spin_lock_irqsave(&up_cpumask_lock, flags);
+		}
+
+		set_current_state(TASK_RUNNING);
+		tmp_mask = up_cpumask;
+		cpumask_clear(&up_cpumask);
+		spin_unlock_irqrestore(&up_cpumask_lock, flags);
+
+		for_each_cpu(cpu, &tmp_mask) {
+			unsigned int j;
+			unsigned int max_freq = 0;
+
+			pcpu = &per_cpu(cpuinfo, cpu);
+			smp_rmb();
+
+			if (!pcpu->governor_enabled)
+				continue;
+
+			mutex_lock(&set_speed_lock);
+
+			for_each_cpu(j, pcpu->policy->cpus) {
+				struct cpufreq_interactive_cpuinfo *pjcpu =
+					&per_cpu(cpuinfo, j);
+
+				if (pjcpu->target_freq > max_freq)
+					max_freq = pjcpu->target_freq;
+			}
+
+			if (max_freq != pcpu->policy->cur)
+				__cpufreq_driver_target(pcpu->policy,
+							max_freq,
+							CPUFREQ_RELATION_H);
+			mutex_unlock(&set_speed_lock);
+
+			pcpu->freq_change_time_in_idle =
+				get_cpu_idle_time_us(cpu,
+						     &pcpu->freq_change_time);
+		}
+	}
+
+	return 0;
+}
+
+static void cpufreq_interactive_freq_down(struct work_struct *work)
+{
+	unsigned int cpu;
+	cpumask_t tmp_mask;
+	unsigned long flags;
+	struct cpufreq_interactive_cpuinfo *pcpu;
+
+	spin_lock_irqsave(&down_cpumask_lock, flags);
+	tmp_mask = down_cpumask;
+	cpumask_clear(&down_cpumask);
+	spin_unlock_irqrestore(&down_cpumask_lock, flags);
+
+	for_each_cpu(cpu, &tmp_mask) {
+		unsigned int j;
+		unsigned int max_freq = 0;
+
+		pcpu = &per_cpu(cpuinfo, cpu);
+		smp_rmb();
+
+		if (!pcpu->governor_enabled)
+			continue;
+
+		mutex_lock(&set_speed_lock);
+
+		for_each_cpu(j, pcpu->policy->cpus) {
+			struct cpufreq_interactive_cpuinfo *pjcpu =
+				&per_cpu(cpuinfo, j);
+
+			if (pjcpu->target_freq > max_freq)
+				max_freq = pjcpu->target_freq;
+		}
+
+		if (max_freq != pcpu->policy->cur)
+			__cpufreq_driver_target(pcpu->policy, max_freq,
+						CPUFREQ_RELATION_H);
+
+		mutex_unlock(&set_speed_lock);
+		pcpu->freq_change_time_in_idle =
+			get_cpu_idle_time_us(cpu,
+					     &pcpu->freq_change_time);
+	}
+}
+
+static ssize_t show_hispeed_freq(struct kobject *kobj,
+				 struct attribute *attr, char *buf)
+{
+	return sprintf(buf, "%llu\n", hispeed_freq);
+}
+
+static ssize_t store_hispeed_freq(struct kobject *kobj,
+				  struct attribute *attr, const char *buf,
+				  size_t count)
+{
+	int ret;
+	u64 val;
+
+	ret = strict_strtoull(buf, 0, &val);
+	if (ret < 0)
+		return ret;
+	hispeed_freq = val;
+	return count;
+}
+
+static struct global_attr hispeed_freq_attr = __ATTR(hispeed_freq, 0644,
+		show_hispeed_freq, store_hispeed_freq);
+
+
+static ssize_t show_go_hispeed_load(struct kobject *kobj,
+				     struct attribute *attr, char *buf)
+{
+	return sprintf(buf, "%lu\n", go_hispeed_load);
+}
+
+static ssize_t store_go_hispeed_load(struct kobject *kobj,
+			struct attribute *attr, const char *buf, size_t count)
+{
+	int ret;
+	unsigned long val;
+
+	ret = strict_strtoul(buf, 0, &val);
+	if (ret < 0)
+		return ret;
+	go_hispeed_load = val;
+	return count;
+}
+
+static struct global_attr go_hispeed_load_attr = __ATTR(go_hispeed_load, 0644,
+		show_go_hispeed_load, store_go_hispeed_load);
+
+static ssize_t show_min_sample_time(struct kobject *kobj,
+				struct attribute *attr, char *buf)
+{
+	return sprintf(buf, "%lu\n", min_sample_time);
+}
+
+static ssize_t store_min_sample_time(struct kobject *kobj,
+			struct attribute *attr, const char *buf, size_t count)
+{
+	int ret;
+	unsigned long val;
+
+	ret = strict_strtoul(buf, 0, &val);
+	if (ret < 0)
+		return ret;
+	min_sample_time = val;
+	return count;
+}
+
+static struct global_attr min_sample_time_attr = __ATTR(min_sample_time, 0644,
+		show_min_sample_time, store_min_sample_time);
+
+static ssize_t show_timer_rate(struct kobject *kobj,
+			struct attribute *attr, char *buf)
+{
+	return sprintf(buf, "%lu\n", timer_rate);
+}
+
+static ssize_t store_timer_rate(struct kobject *kobj,
+			struct attribute *attr, const char *buf, size_t count)
+{
+	int ret;
+	unsigned long val;
+
+	ret = strict_strtoul(buf, 0, &val);
+	if (ret < 0)
+		return ret;
+	timer_rate = val;
+	return count;
+}
+
+static struct global_attr timer_rate_attr = __ATTR(timer_rate, 0644,
+		show_timer_rate, store_timer_rate);
+
+static struct attribute *interactive_attributes[] = {
+	&hispeed_freq_attr.attr,
+	&go_hispeed_load_attr.attr,
+	&min_sample_time_attr.attr,
+	&timer_rate_attr.attr,
+	NULL,
+};
+
+static struct attribute_group interactive_attr_group = {
+	.attrs = interactive_attributes,
+	.name = "interactive",
+};
+
+static int cpufreq_governor_interactive(struct cpufreq_policy *policy,
+		unsigned int event)
+{
+	int rc;
+	unsigned int j;
+	struct cpufreq_interactive_cpuinfo *pcpu;
+	struct cpufreq_frequency_table *freq_table;
+
+	switch (event) {
+	case CPUFREQ_GOV_START:
+		if (!cpu_online(policy->cpu))
+			return -EINVAL;
+
+		freq_table =
+			cpufreq_frequency_get_table(policy->cpu);
+
+		for_each_cpu(j, policy->cpus) {
+			pcpu = &per_cpu(cpuinfo, j);
+			pcpu->policy = policy;
+			pcpu->target_freq = policy->cur;
+			pcpu->freq_table = freq_table;
+			pcpu->freq_change_time_in_idle =
+				get_cpu_idle_time_us(j,
+					     &pcpu->freq_change_time);
+			pcpu->governor_enabled = 1;
+			smp_wmb();
+		}
+
+		if (!hispeed_freq)
+			hispeed_freq = policy->max;
+
+		/*
+		 * Do not register the idle hook and create sysfs
+		 * entries if we have already done so.
+		 */
+		if (atomic_inc_return(&active_count) > 1)
+			return 0;
+
+		rc = sysfs_create_group(cpufreq_global_kobject,
+				&interactive_attr_group);
+		if (rc)
+			return rc;
+
+		break;
+
+	case CPUFREQ_GOV_STOP:
+		for_each_cpu(j, policy->cpus) {
+			pcpu = &per_cpu(cpuinfo, j);
+			pcpu->governor_enabled = 0;
+			smp_wmb();
+			del_timer_sync(&pcpu->cpu_timer);
+
+			/*
+			 * Reset idle exit time since we may cancel the timer
+			 * before it can run after the last idle exit time,
+			 * to avoid tripping the check in idle exit for a timer
+			 * that is trying to run.
+			 */
+			pcpu->idle_exit_time = 0;
+		}
+
+		flush_work(&freq_scale_down_work);
+		if (atomic_dec_return(&active_count) > 0)
+			return 0;
+
+		sysfs_remove_group(cpufreq_global_kobject,
+				&interactive_attr_group);
+
+		break;
+
+	case CPUFREQ_GOV_LIMITS:
+		if (policy->max < policy->cur)
+			__cpufreq_driver_target(policy,
+					policy->max, CPUFREQ_RELATION_H);
+		else if (policy->min > policy->cur)
+			__cpufreq_driver_target(policy,
+					policy->min, CPUFREQ_RELATION_L);
+		break;
+	}
+	return 0;
+}
+
+static int cpufreq_interactive_idle_notifier(struct notifier_block *nb,
+					     unsigned long val,
+					     void *data)
+{
+	switch (val) {
+	case IDLE_START:
+		cpufreq_interactive_idle_start();
+		break;
+	case IDLE_END:
+		cpufreq_interactive_idle_end();
+		break;
+	}
+
+	return 0;
+}
+
+static struct notifier_block cpufreq_interactive_idle_nb = {
+	.notifier_call = cpufreq_interactive_idle_notifier,
+};
+
+static int __init cpufreq_interactive_init(void)
+{
+	unsigned int i;
+	struct cpufreq_interactive_cpuinfo *pcpu;
+	struct sched_param param = { .sched_priority = MAX_RT_PRIO-1 };
+
+	go_hispeed_load = DEFAULT_GO_HISPEED_LOAD;
+	min_sample_time = DEFAULT_MIN_SAMPLE_TIME;
+	timer_rate = DEFAULT_TIMER_RATE;
+
+	/* Initalize per-cpu timers */
+	for_each_possible_cpu(i) {
+		pcpu = &per_cpu(cpuinfo, i);
+		init_timer(&pcpu->cpu_timer);
+		pcpu->cpu_timer.function = cpufreq_interactive_timer;
+		pcpu->cpu_timer.data = i;
+	}
+
+	up_task = kthread_create(cpufreq_interactive_up_task, NULL,
+				 "kinteractiveup");
+	if (IS_ERR(up_task))
+		return PTR_ERR(up_task);
+
+	sched_setscheduler_nocheck(up_task, SCHED_FIFO, &param);
+	get_task_struct(up_task);
+
+	/* No rescuer thread, bind to CPU queuing the work for possibly
+	   warm cache (probably doesn't matter much). */
+	down_wq = alloc_workqueue("knteractive_down", 0, 1);
+
+	if (!down_wq)
+		goto err_freeuptask;
+
+	INIT_WORK(&freq_scale_down_work,
+		  cpufreq_interactive_freq_down);
+
+	spin_lock_init(&up_cpumask_lock);
+	spin_lock_init(&down_cpumask_lock);
+	mutex_init(&set_speed_lock);
+
+	idle_notifier_register(&cpufreq_interactive_idle_nb);
+
+	return cpufreq_register_governor(&cpufreq_gov_interactive);
+
+err_freeuptask:
+	put_task_struct(up_task);
+	return -ENOMEM;
+}
+
+#ifdef CONFIG_CPU_FREQ_DEFAULT_GOV_INTERACTIVE
+fs_initcall(cpufreq_interactive_init);
+#else
+module_init(cpufreq_interactive_init);
+#endif
+
+static void __exit cpufreq_interactive_exit(void)
+{
+	cpufreq_unregister_governor(&cpufreq_gov_interactive);
+	kthread_stop(up_task);
+	put_task_struct(up_task);
+	destroy_workqueue(down_wq);
+}
+
+module_exit(cpufreq_interactive_exit);
+
+MODULE_AUTHOR("Mike Chan <mike@android.com>");
+MODULE_DESCRIPTION("'cpufreq_interactive' - A cpufreq governor for "
+	"Latency sensitive workloads");
+MODULE_LICENSE("GPL");
diff --git a/include/linux/cpufreq.h b/include/linux/cpufreq.h
index dca22de..2ea8ab5 100644
--- a/include/linux/cpufreq.h
+++ b/include/linux/cpufreq.h
@@ -491,6 +491,9 @@ extern struct cpufreq_governor cpufreq_gov_ondemand;
 #elif defined(CONFIG_CPU_FREQ_DEFAULT_GOV_CONSERVATIVE)
 extern struct cpufreq_governor cpufreq_gov_conservative;
 #define CPUFREQ_DEFAULT_GOVERNOR	(&cpufreq_gov_conservative)
+#elif defined(CONFIG_CPU_FREQ_DEFAULT_GOV_INTERACTIVE)
+extern struct cpufreq_governor cpufreq_gov_interactive;
+#define CPUFREQ_DEFAULT_GOVERNOR	(&cpufreq_gov_interactive)
 #endif
 
 /*********************************************************************
-- 
1.7.9.5

--
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/

[toc] | [next] | [standalone]

#1256994 — [PATCH 57/70] cpufreq: interactive: fix compiling warnings

From	Bálint Czobor <czoborbalint@gmail.com>
Date	2015-10-27 18:40 +0100
Subject	[PATCH 57/70] cpufreq: interactive: fix compiling warnings
Message-ID	<qogo2-84h-15@gated-at.bofh.it>
In reply to	#1256993

From: Chih-Wei Huang <cwhuang@linux.org.tw>

The gcc warns like:

  cpufreq_interactive.c:745:6: warning: operation on 'ret' may be undefined [-Wsequence-point]

It was introduced by commit cf0fad49d17cb8273ce555dd5b7afab67d7923bf.

Since sprintf(...) just return 1 (one character) in this case, ret should not changed.
Just discarding the result of sprintf(...) leads to the result that
the committer of cf0fad49d17cb8273ce555dd5b7afab67d7923bf wants.

Change-Id: Ifed1cef6d6a31c3ed23dad03a567b3b9eddf3a57
Signed-off-by: Chih-Wei Huang <cwhuang@android-x86.org>
Signed-off-by: Bálint Czobor <czoborbalint@gmail.com>
---
 drivers/cpufreq/cpufreq_interactive.c |    4 ++--
 1 file changed, 2 insertions(+), 2 deletions(-)

diff --git a/drivers/cpufreq/cpufreq_interactive.c b/drivers/cpufreq/cpufreq_interactive.c
index d72e8c4..9f63bd1 100644
--- a/drivers/cpufreq/cpufreq_interactive.c
+++ b/drivers/cpufreq/cpufreq_interactive.c
@@ -752,7 +752,7 @@ static ssize_t show_target_loads(
 		ret += sprintf(buf + ret, "%u%s", tunables->target_loads[i],
 			       i & 0x1 ? ":" : " ");
 
-	ret += sprintf(buf + --ret, "\n");
+	sprintf(buf + ret - 1, "\n");
 	spin_unlock_irqrestore(&tunables->target_loads_lock, flags);
 	return ret;
 }
@@ -792,7 +792,7 @@ static ssize_t show_above_hispeed_delay(
 			       tunables->above_hispeed_delay[i],
 			       i & 0x1 ? ":" : " ");
 
-	ret += sprintf(buf + --ret, "\n");
+	sprintf(buf + ret - 1, "\n");
 	spin_unlock_irqrestore(&tunables->above_hispeed_delay_lock, flags);
 	return ret;
 }
-- 
1.7.9.5

--
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/

[toc] | [prev] | [next] | [standalone]

#1256997 — [PATCH 56/70] cpufreq: interactive: delete timers for GOV_START

From	Bálint Czobor <czoborbalint@gmail.com>
Date	2015-10-27 18:40 +0100
Subject	[PATCH 56/70] cpufreq: interactive: delete timers for GOV_START
Message-ID	<qogo2-84h-27@gated-at.bofh.it>
In reply to	#1256993

From: Shridhar Rasal <srasal@nvidia.com>

Make sure that timers cpu_timer and cpu_slack_timer
deactivated before addition of new.

Change-Id: If31c4049606871df6f00efdc24b1d713c86a6f69
Signed-off-by: Shridhar Rasal <srasal@nvidia.com>
Signed-off-by: Bharat Nihalani <bnihalani@nvidia.com>
Signed-off-by: Bálint Czobor <czoborbalint@gmail.com>
---
 drivers/cpufreq/cpufreq_interactive.c |    2 ++
 1 file changed, 2 insertions(+)

diff --git a/drivers/cpufreq/cpufreq_interactive.c b/drivers/cpufreq/cpufreq_interactive.c
index 66c096d..d72e8c4 100644
--- a/drivers/cpufreq/cpufreq_interactive.c
+++ b/drivers/cpufreq/cpufreq_interactive.c
@@ -1247,6 +1247,8 @@ static int cpufreq_governor_interactive(struct cpufreq_policy *policy,
 			pcpu->hispeed_validate_time =
 				pcpu->floor_validate_time;
 			down_write(&pcpu->enable_sem);
+			del_timer_sync(&pcpu->cpu_timer);
+			del_timer_sync(&pcpu->cpu_slack_timer);
 			cpufreq_interactive_timer_start(tunables, j);
 			pcpu->governor_enabled = 1;
 			up_write(&pcpu->enable_sem);
-- 
1.7.9.5

--
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/

[toc] | [prev] | [next] | [standalone]

#1256998 — [PATCH 41/70] cpufreq: interactive: fix race on governor start/stop

From	Bálint Czobor <czoborbalint@gmail.com>
Date	2015-10-27 18:40 +0100
Subject	[PATCH 41/70] cpufreq: interactive: fix race on governor start/stop
Message-ID	<qogo2-84h-29@gated-at.bofh.it>
In reply to	#1256993

From: Lianwei Wang <a22439@motorola.com>

There is race condition when both two cpu do CPUFREQ_GOV_STOP and one cpu
do CPUFREQ_GOV_START soon. The sysfs_remove_group is not done yet on one
cpu, but sysfs_create_group is called on another cpu, which cause governor
start failed and then kernel panic in timer callback because the policy and
cpu mask are all kfree in cpufreq driver.

Replace atomic with mutex to lock the whole START/STOP sequence.

Change-Id: I3762b3d44315ae021b8275aca84f5ea9147cc540
Signed-off-by: Lianwei Wang <a22439@motorola.com>
Signed-off-by: Bálint Czobor <czoborbalint@gmail.com>
---
 drivers/cpufreq/cpufreq_interactive.c |   21 +++++++++++++++++----
 1 file changed, 17 insertions(+), 4 deletions(-)

diff --git a/drivers/cpufreq/cpufreq_interactive.c b/drivers/cpufreq/cpufreq_interactive.c
index e7f26aa..3447e58 100644
--- a/drivers/cpufreq/cpufreq_interactive.c
+++ b/drivers/cpufreq/cpufreq_interactive.c
@@ -35,7 +35,7 @@
 #define CREATE_TRACE_POINTS
 #include <trace/events/cpufreq_interactive.h>
 
-static atomic_t active_count = ATOMIC_INIT(0);
+static int active_count;
 
 struct cpufreq_interactive_cpuinfo {
 	struct timer_list cpu_timer;
@@ -61,6 +61,7 @@ static DEFINE_PER_CPU(struct cpufreq_interactive_cpuinfo, cpuinfo);
 static struct task_struct *speedchange_task;
 static cpumask_t speedchange_cpumask;
 static spinlock_t speedchange_cpumask_lock;
+static struct mutex gov_lock;
 
 /* Hi speed to bump to from lo speed when load burst (default max) */
 static unsigned int hispeed_freq;
@@ -914,6 +915,8 @@ static int cpufreq_governor_interactive(struct cpufreq_policy *policy,
 		if (!cpu_online(policy->cpu))
 			return -EINVAL;
 
+		mutex_lock(&gov_lock);
+
 		freq_table =
 			cpufreq_frequency_get_table(policy->cpu);
 		if (!hispeed_freq)
@@ -948,20 +951,26 @@ static int cpufreq_governor_interactive(struct cpufreq_policy *policy,
 		 * Do not register the idle hook and create sysfs
 		 * entries if we have already done so.
 		 */
-		if (atomic_inc_return(&active_count) > 1)
+		if (++active_count > 1) {
+			mutex_unlock(&gov_lock);
 			return 0;
+		}
 
 		rc = sysfs_create_group(cpufreq_global_kobject,
 				&interactive_attr_group);
-		if (rc)
+		if (rc) {
+			mutex_unlock(&gov_lock);
 			return rc;
+		}
 
 		idle_notifier_register(&cpufreq_interactive_idle_nb);
 		cpufreq_register_notifier(
 			&cpufreq_notifier_block, CPUFREQ_TRANSITION_NOTIFIER);
+		mutex_unlock(&gov_lock);
 		break;
 
 	case CPUFREQ_GOV_STOP:
+		mutex_lock(&gov_lock);
 		for_each_cpu(j, policy->cpus) {
 			pcpu = &per_cpu(cpuinfo, j);
 			down_write(&pcpu->enable_sem);
@@ -971,14 +980,17 @@ static int cpufreq_governor_interactive(struct cpufreq_policy *policy,
 			up_write(&pcpu->enable_sem);
 		}
 
-		if (atomic_dec_return(&active_count) > 0)
+		if (--active_count > 0) {
+			mutex_unlock(&gov_lock);
 			return 0;
+		}
 
 		cpufreq_unregister_notifier(
 			&cpufreq_notifier_block, CPUFREQ_TRANSITION_NOTIFIER);
 		idle_notifier_unregister(&cpufreq_interactive_idle_nb);
 		sysfs_remove_group(cpufreq_global_kobject,
 				&interactive_attr_group);
+		mutex_unlock(&gov_lock);
 
 		break;
 
@@ -1018,6 +1030,7 @@ static int __init cpufreq_interactive_init(void)
 
 	spin_lock_init(&target_loads_lock);
 	spin_lock_init(&speedchange_cpumask_lock);
+	mutex_init(&gov_lock);
 	speedchange_task =
 		kthread_create(cpufreq_interactive_speedchange_task, NULL,
 			       "cfinteractive");
-- 
1.7.9.5

--
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/

[toc] | [prev] | [next] | [standalone]

#1256999 — [PATCH 32/70] cpufreq: interactive: specify duration of CPU speed boost pulse

From	Bálint Czobor <czoborbalint@gmail.com>
Date	2015-10-27 18:40 +0100
Subject	[PATCH 32/70] cpufreq: interactive: specify duration of CPU speed boost pulse
Message-ID	<qogo2-84h-35@gated-at.bofh.it>
In reply to	#1256993

From: Todd Poynor <toddpoynor@google.com>

Sysfs attribute boostpulse_duration specifies the duration of boosting CPU
speed in response to bootpulse events.  Duration is specified in usecs,
default 80ms.

Change-Id: Ifd41625574891a44f1787a4e85d1e7b4f2afb52b
Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Bálint Czobor <czoborbalint@gmail.com>
---
 drivers/cpufreq/cpufreq_interactive.c |   52 ++++++++++++++++++++++++++++-----
 1 file changed, 45 insertions(+), 7 deletions(-)

diff --git a/drivers/cpufreq/cpufreq_interactive.c b/drivers/cpufreq/cpufreq_interactive.c
index d0d51ee..8b44a82 100644
--- a/drivers/cpufreq/cpufreq_interactive.c
+++ b/drivers/cpufreq/cpufreq_interactive.c
@@ -95,11 +95,12 @@ static unsigned long timer_rate;
 #define DEFAULT_ABOVE_HISPEED_DELAY DEFAULT_TIMER_RATE
 static unsigned long above_hispeed_delay_val;
 
-/*
- * Non-zero means longer-term speed boost active.
- */
-
+/* Non-zero means indefinite speed boost active */
 static int boost_val;
+/* Duration of a boot pulse in usecs */
+static int boostpulse_duration_val = DEFAULT_MIN_SAMPLE_TIME;
+/* End time of boost pulse in ktime converted to usecs */
+static u64 boostpulse_endtime;
 
 static bool governidle;
 module_param(governidle, bool, S_IWUSR | S_IRUGO);
@@ -268,6 +269,7 @@ static void cpufreq_interactive_timer(unsigned long data)
 	unsigned int loadadjfreq;
 	unsigned int index;
 	unsigned long flags;
+	bool boosted;
 
 	smp_rmb();
 
@@ -286,8 +288,9 @@ static void cpufreq_interactive_timer(unsigned long data)
 	do_div(cputime_speedadj, delta_time);
 	loadadjfreq = (unsigned int)cputime_speedadj * 100;
 	cpu_load = loadadjfreq / pcpu->target_freq;
+	boosted = boost_val || now < boostpulse_endtime;
 
-	if ((cpu_load >= go_hispeed_load || boost_val) &&
+	if ((cpu_load >= go_hispeed_load || boosted) &&
 	    pcpu->target_freq < hispeed_freq)
 		new_freq = hispeed_freq;
 	else
@@ -327,8 +330,18 @@ static void cpufreq_interactive_timer(unsigned long data)
 		}
 	}
 
-	pcpu->floor_freq = new_freq;
-	pcpu->floor_validate_time = now;
+	/*
+	 * Update the timestamp for checking whether speed has been held at
+	 * or above the selected frequency for a minimum of min_sample_time,
+	 * if not boosted to hispeed_freq.  If boosted to hispeed_freq then we
+	 * allow the speed to drop as soon as the boostpulse duration expires
+	 * (or the indefinite boost is turned off).
+	 */
+
+	if (!boosted || new_freq > hispeed_freq) {
+		pcpu->floor_freq = new_freq;
+		pcpu->floor_validate_time = now;
+	}
 
 	if (pcpu->target_freq == new_freq) {
 		trace_cpufreq_interactive_already(
@@ -774,6 +787,7 @@ static ssize_t store_boostpulse(struct kobject *kobj, struct attribute *attr,
 	if (ret < 0)
 		return ret;
 
+	boostpulse_endtime = ktime_to_us(ktime_get()) + boostpulse_duration_val;
 	trace_cpufreq_interactive_boost("pulse");
 	cpufreq_interactive_boost();
 	return count;
@@ -782,6 +796,29 @@ static ssize_t store_boostpulse(struct kobject *kobj, struct attribute *attr,
 static struct global_attr boostpulse =
 	__ATTR(boostpulse, 0200, NULL, store_boostpulse);
 
+static ssize_t show_boostpulse_duration(
+	struct kobject *kobj, struct attribute *attr, char *buf)
+{
+	return sprintf(buf, "%d\n", boostpulse_duration_val);
+}
+
+static ssize_t store_boostpulse_duration(
+	struct kobject *kobj, struct attribute *attr, const char *buf,
+	size_t count)
+{
+	int ret;
+	unsigned long val;
+
+	ret = kstrtoul(buf, 0, &val);
+	if (ret < 0)
+		return ret;
+
+	boostpulse_duration_val = val;
+	return count;
+}
+
+define_one_global_rw(boostpulse_duration);
+
 static struct attribute *interactive_attributes[] = {
 	&target_loads_attr.attr,
 	&hispeed_freq_attr.attr,
@@ -791,6 +828,7 @@ static struct attribute *interactive_attributes[] = {
 	&timer_rate_attr.attr,
 	&boost.attr,
 	&boostpulse.attr,
+	&boostpulse_duration.attr,
 	NULL,
 };
 
-- 
1.7.9.5

--
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/

[toc] | [prev] | [next] | [standalone]

#1257000 — [PATCH 43/70] cpufreq: interactive: add io_is_busy interface

From	Bálint Czobor <czoborbalint@gmail.com>
Date	2015-10-27 18:40 +0100
Subject	[PATCH 43/70] cpufreq: interactive: add io_is_busy interface
Message-ID	<qogo2-84h-37@gated-at.bofh.it>
In reply to	#1256993

From: Lianwei Wang <lian-wei.wang@motorola.com>

Previously the idle time returned from get_cpu_idle_time_us included the
iowait time. So the iowait time was always calculated as idle time.

But now the idle time returned from get_cpu_idle_time_us does not include
the iowait time anymore because of below commit which cause the iowait time
always calculated as busy time:
    6beea0c nohz: Fix update_ts_time_stat idle accounting

Add the io_is_busy interface, as does the ondemand governor, and let the user
configure the iowait time as busy or idle through the io_is_busy sysfs
interface.

By default, io_is_busy is disabled.

[toddpoynor@google.com: minor updates]
Change-Id: If7d70ff864c43bc9c8d7fd7cfc66f930d339f9b4
Signed-off-by: Lianwei Wang <lian-wei.wang@motorola.com>
Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Bálint Czobor <czoborbalint@gmail.com>
---
 drivers/cpufreq/cpufreq_interactive.c |   66 ++++++++++++++++++++++++++++++++-
 1 file changed, 64 insertions(+), 2 deletions(-)

diff --git a/drivers/cpufreq/cpufreq_interactive.c b/drivers/cpufreq/cpufreq_interactive.c
index 620b46c..7c734fa 100644
--- a/drivers/cpufreq/cpufreq_interactive.c
+++ b/drivers/cpufreq/cpufreq_interactive.c
@@ -30,6 +30,7 @@
 #include <linux/workqueue.h>
 #include <linux/kthread.h>
 #include <linux/slab.h>
+#include <linux/kernel_stat.h>
 #include <asm/cputime.h>
 
 #define CREATE_TRACE_POINTS
@@ -114,6 +115,8 @@ static u64 boostpulse_endtime;
 #define DEFAULT_TIMER_SLACK (4 * DEFAULT_TIMER_RATE)
 static int timer_slack_val = DEFAULT_TIMER_SLACK;
 
+static bool io_is_busy;
+
 static int cpufreq_governor_interactive(struct cpufreq_policy *policy,
 		unsigned int event);
 
@@ -127,6 +130,42 @@ struct cpufreq_governor cpufreq_gov_interactive = {
 	.owner = THIS_MODULE,
 };
 
+static inline cputime64_t get_cpu_idle_time_jiffy(unsigned int cpu,
+						  cputime64_t *wall)
+{
+	u64 idle_time;
+	u64 cur_wall_time;
+	u64 busy_time;
+
+	cur_wall_time = jiffies64_to_cputime64(get_jiffies_64());
+
+	busy_time  = kcpustat_cpu(cpu).cpustat[CPUTIME_USER];
+	busy_time += kcpustat_cpu(cpu).cpustat[CPUTIME_SYSTEM];
+	busy_time += kcpustat_cpu(cpu).cpustat[CPUTIME_IRQ];
+	busy_time += kcpustat_cpu(cpu).cpustat[CPUTIME_SOFTIRQ];
+	busy_time += kcpustat_cpu(cpu).cpustat[CPUTIME_STEAL];
+	busy_time += kcpustat_cpu(cpu).cpustat[CPUTIME_NICE];
+
+	idle_time = cur_wall_time - busy_time;
+	if (wall)
+		*wall = jiffies_to_usecs(cur_wall_time);
+
+	return jiffies_to_usecs(idle_time);
+}
+
+static inline cputime64_t get_cpu_idle_time(unsigned int cpu,
+					    cputime64_t *wall)
+{
+	u64 idle_time = get_cpu_idle_time_us(cpu, wall);
+
+	if (idle_time == -1ULL)
+		idle_time = get_cpu_idle_time_jiffy(cpu, wall);
+	else if (!io_is_busy)
+		idle_time += get_cpu_iowait_time_us(cpu, wall);
+
+	return idle_time;
+}
+
 static void cpufreq_interactive_timer_resched(
 	struct cpufreq_interactive_cpuinfo *pcpu)
 {
@@ -141,7 +180,7 @@ static void cpufreq_interactive_timer_resched(
 
 	spin_lock_irqsave(&pcpu->load_lock, flags);
 	pcpu->time_in_idle =
-		get_cpu_idle_time_us(smp_processor_id(),
+		get_cpu_idle_time(smp_processor_id(),
 				     &pcpu->time_in_idle_timestamp);
 	pcpu->cputime_speedadj = 0;
 	pcpu->cputime_speedadj_timestamp = pcpu->time_in_idle_timestamp;
@@ -278,7 +317,7 @@ static u64 update_load(int cpu)
 	unsigned int delta_time;
 	u64 active_time;
 
-	now_idle = get_cpu_idle_time_us(cpu, &now);
+	now_idle = get_cpu_idle_time(cpu, &now);
 	delta_idle = (unsigned int)(now_idle - pcpu->time_in_idle);
 	delta_time = (unsigned int)(now - pcpu->time_in_idle_timestamp);
 	active_time = delta_time - delta_idle;
@@ -924,6 +963,28 @@ static ssize_t store_boostpulse_duration(
 
 define_one_global_rw(boostpulse_duration);
 
+static ssize_t show_io_is_busy(struct kobject *kobj,
+			struct attribute *attr, char *buf)
+{
+	return sprintf(buf, "%u\n", io_is_busy);
+}
+
+static ssize_t store_io_is_busy(struct kobject *kobj,
+			struct attribute *attr, const char *buf, size_t count)
+{
+	int ret;
+	unsigned long val;
+
+	ret = kstrtoul(buf, 0, &val);
+	if (ret < 0)
+		return ret;
+	io_is_busy = val;
+	return count;
+}
+
+static struct global_attr io_is_busy_attr = __ATTR(io_is_busy, 0644,
+		show_io_is_busy, store_io_is_busy);
+
 static struct attribute *interactive_attributes[] = {
 	&target_loads_attr.attr,
 	&above_hispeed_delay_attr.attr,
@@ -935,6 +996,7 @@ static struct attribute *interactive_attributes[] = {
 	&boost.attr,
 	&boostpulse.attr,
 	&boostpulse_duration.attr,
+	&io_is_busy_attr.attr,
 	NULL,
 };
 
-- 
1.7.9.5

--
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/

[toc] | [prev] | [next] | [standalone]

#1257001 — [PATCH 39/70] cpufreq: interactive: don't handle transition notification if not enabled

From	Bálint Czobor <czoborbalint@gmail.com>
Date	2015-10-27 18:40 +0100
Subject	[PATCH 39/70] cpufreq: interactive: don't handle transition notification if not enabled
Message-ID	<qogo2-84h-43@gated-at.bofh.it>
In reply to	#1256993

From: Todd Poynor <toddpoynor@google.com>

If multiple governors are in use then avoid processing frequency transition
notifications for CPUs on which the interactive governor is not enabled.

Change-Id: Ibd75255b921d887501a64774a8c4f62302f2d4e4
Reported-by: Francisco Franco <francisco.franco@cloudcar.com>
Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Bálint Czobor <czoborbalint@gmail.com>
---
 drivers/cpufreq/cpufreq_interactive.c |    9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)

diff --git a/drivers/cpufreq/cpufreq_interactive.c b/drivers/cpufreq/cpufreq_interactive.c
index c8358a3..51c34bf 100644
--- a/drivers/cpufreq/cpufreq_interactive.c
+++ b/drivers/cpufreq/cpufreq_interactive.c
@@ -552,6 +552,12 @@ static int cpufreq_interactive_notifier(
 
 	if (val == CPUFREQ_POSTCHANGE) {
 		pcpu = &per_cpu(cpuinfo, freq->cpu);
+		if (!down_read_trylock(&pcpu->enable_sem))
+			return 0;
+		if (!pcpu->governor_enabled) {
+			up_read(&pcpu->enable_sem);
+			return 0;
+		}
 
 		for_each_cpu(cpu, pcpu->policy->cpus) {
 			struct cpufreq_interactive_cpuinfo *pjcpu =
@@ -560,8 +566,9 @@ static int cpufreq_interactive_notifier(
 			update_load(cpu);
 			spin_unlock(&pjcpu->load_lock);
 		}
-	}
 
+		up_read(&pcpu->enable_sem);
+	}
 	return 0;
 }
 
-- 
1.7.9.5

--
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/

[toc] | [prev] | [next] | [standalone]

#1257002 — [PATCH 51/70] cpufreq: interactive: resched timer if max freq raised

From	Bálint Czobor <czoborbalint@gmail.com>
Date	2015-10-27 18:40 +0100
Subject	[PATCH 51/70] cpufreq: interactive: resched timer if max freq raised
Message-ID	<qogo1-84h-7@gated-at.bofh.it>
In reply to	#1256993

From: Lianwei Wang <a22439@motorola.com>

When the policy max freq is raised, and before the timer is
rescheduled in idle callback, the cpu freq may stuck at a
lower freq.

The target_freq shall be updated too, else on a high load
situation, the new_freq is always equal to target_freq and
which will cause freq stuck at a lower freq too.

Reschedule the timer on gov limits callback.

Change-Id: I6c187001ab43e859731429b64f75a74eebc37a24
Signed-off-by: Lianwei Wang <a22439@motorola.com>
Signed-off-by: Bálint Czobor <czoborbalint@gmail.com>
---
 drivers/cpufreq/cpufreq_interactive.c |   64 +++++++++++++++++++++++++++------
 1 file changed, 54 insertions(+), 10 deletions(-)

diff --git a/drivers/cpufreq/cpufreq_interactive.c b/drivers/cpufreq/cpufreq_interactive.c
index 691923d..90958fd 100644
--- a/drivers/cpufreq/cpufreq_interactive.c
+++ b/drivers/cpufreq/cpufreq_interactive.c
@@ -189,6 +189,32 @@ static void cpufreq_interactive_timer_resched(
 	spin_unlock_irqrestore(&pcpu->load_lock, flags);
 }
 
+/* The caller shall take enable_sem write semaphore to avoid any timer race.
+ * The cpu_timer and cpu_slack_timer must be deactivated when calling this
+ * function.
+ */
+static void cpufreq_interactive_timer_start(int cpu)
+{
+	struct cpufreq_interactive_cpuinfo *pcpu = &per_cpu(cpuinfo, cpu);
+	unsigned long expires = jiffies + usecs_to_jiffies(timer_rate);
+	unsigned long flags;
+
+	pcpu->cpu_timer.expires = expires;
+	add_timer_on(&pcpu->cpu_timer, cpu);
+	if (timer_slack_val >= 0 && pcpu->target_freq > pcpu->policy->min) {
+		expires += usecs_to_jiffies(timer_slack_val);
+		pcpu->cpu_slack_timer.expires = expires;
+		add_timer_on(&pcpu->cpu_slack_timer, cpu);
+	}
+
+	spin_lock_irqsave(&pcpu->load_lock, flags);
+	pcpu->time_in_idle =
+		get_cpu_idle_time(cpu, &pcpu->time_in_idle_timestamp);
+	pcpu->cputime_speedadj = 0;
+	pcpu->cputime_speedadj_timestamp = pcpu->time_in_idle_timestamp;
+	spin_unlock_irqrestore(&pcpu->load_lock, flags);
+}
+
 static unsigned int freq_to_above_hispeed_delay(unsigned int freq)
 {
 	int i;
@@ -1058,8 +1084,6 @@ static int cpufreq_governor_interactive(struct cpufreq_policy *policy,
 			hispeed_freq = policy->max;
 
 		for_each_cpu(j, policy->cpus) {
-			unsigned long expires;
-
 			pcpu = &per_cpu(cpuinfo, j);
 			pcpu->policy = policy;
 			pcpu->target_freq = policy->cur;
@@ -1070,14 +1094,7 @@ static int cpufreq_governor_interactive(struct cpufreq_policy *policy,
 			pcpu->hispeed_validate_time =
 				pcpu->floor_validate_time;
 			down_write(&pcpu->enable_sem);
-			expires = jiffies + usecs_to_jiffies(timer_rate);
-			pcpu->cpu_timer.expires = expires;
-			add_timer_on(&pcpu->cpu_timer, j);
-			if (timer_slack_val >= 0) {
-				expires += usecs_to_jiffies(timer_slack_val);
-				pcpu->cpu_slack_timer.expires = expires;
-				add_timer_on(&pcpu->cpu_slack_timer, j);
-			}
+			cpufreq_interactive_timer_start(j);
 			pcpu->governor_enabled = 1;
 			up_write(&pcpu->enable_sem);
 		}
@@ -1136,6 +1153,33 @@ static int cpufreq_governor_interactive(struct cpufreq_policy *policy,
 		else if (policy->min > policy->cur)
 			__cpufreq_driver_target(policy,
 					policy->min, CPUFREQ_RELATION_L);
+		for_each_cpu(j, policy->cpus) {
+			pcpu = &per_cpu(cpuinfo, j);
+
+			/* hold write semaphore to avoid race */
+			down_write(&pcpu->enable_sem);
+			if (pcpu->governor_enabled == 0) {
+				up_write(&pcpu->enable_sem);
+				continue;
+			}
+
+			/* update target_freq firstly */
+			if (policy->max < pcpu->target_freq)
+				pcpu->target_freq = policy->max;
+			else if (policy->min > pcpu->target_freq)
+				pcpu->target_freq = policy->min;
+
+			/* Reschedule timer.
+			 * Delete the timers, else the timer callback may
+			 * return without re-arm the timer when failed
+			 * acquire the semaphore. This race may cause timer
+			 * stopped unexpectedly.
+			 */
+			del_timer_sync(&pcpu->cpu_timer);
+			del_timer_sync(&pcpu->cpu_slack_timer);
+			cpufreq_interactive_timer_start(j);
+			up_write(&pcpu->enable_sem);
+		}
 		break;
 	}
 	return 0;
-- 
1.7.9.5

--
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/

[toc] | [prev] | [next] | [standalone]

#1257003 — [PATCH 09/70] cpufreq: interactive: Separate speed target revalidate time and initial set time

From	Bálint Czobor <czoborbalint@gmail.com>
Date	2015-10-27 18:40 +0100
Subject	[PATCH 09/70] cpufreq: interactive: Separate speed target revalidate time and initial set time
Message-ID	<qogo1-84h-9@gated-at.bofh.it>
In reply to	#1256993

From: Todd Poynor <toddpoynor@google.com>

Allow speed drop after min_sample_time elapses from last time
the current speed was last re-validated as appropriate for
current load / input boost.

Allow speed bump after min_sample_time (or above_hispeed_delay)
elapses from the time the current speed was originally set.

Change-Id: Ic25687a7a53d25e6544c30c47d7ab6f27a47bee8
Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Bálint Czobor <czoborbalint@gmail.com>
---
 drivers/cpufreq/cpufreq_interactive.c |   16 ++++++++++++----
 1 file changed, 12 insertions(+), 4 deletions(-)

diff --git a/drivers/cpufreq/cpufreq_interactive.c b/drivers/cpufreq/cpufreq_interactive.c
index b78174f..083f790 100644
--- a/drivers/cpufreq/cpufreq_interactive.c
+++ b/drivers/cpufreq/cpufreq_interactive.c
@@ -46,6 +46,8 @@ struct cpufreq_interactive_cpuinfo {
 	int idling;
 	u64 target_set_time;
 	u64 target_set_time_in_idle;
+	u64 target_validate_time;
+	u64 target_validate_time_in_idle;
 	struct cpufreq_policy *policy;
 	struct cpufreq_frequency_table *freq_table;
 	unsigned int target_freq;
@@ -208,10 +210,10 @@ static void cpufreq_interactive_timer(unsigned long data)
 
 	/*
 	 * Do not scale down unless we have been at this frequency for the
-	 * minimum sample time.
+	 * minimum sample time since last validated.
 	 */
 	if (new_freq < pcpu->target_freq) {
-		if (pcpu->timer_run_time - pcpu->target_set_time
+		if (pcpu->timer_run_time - pcpu->target_validate_time
 		    < min_sample_time) {
 			trace_cpufreq_interactive_notyet(data, cpu_load,
 					 pcpu->target_freq, new_freq);
@@ -219,8 +221,8 @@ static void cpufreq_interactive_timer(unsigned long data)
 		}
 	}
 
-	pcpu->target_set_time_in_idle = now_idle;
-	pcpu->target_set_time = pcpu->timer_run_time;
+	pcpu->target_validate_time_in_idle = now_idle;
+	pcpu->target_validate_time = pcpu->timer_run_time;
 
 	if (pcpu->target_freq == new_freq) {
 		trace_cpufreq_interactive_already(data, cpu_load,
@@ -230,6 +232,8 @@ static void cpufreq_interactive_timer(unsigned long data)
 
 	trace_cpufreq_interactive_target(data, cpu_load, pcpu->target_freq,
 					 new_freq);
+	pcpu->target_set_time_in_idle = now_idle;
+	pcpu->target_set_time = pcpu->timer_run_time;
 
 	if (new_freq < pcpu->target_freq) {
 		pcpu->target_freq = new_freq;
@@ -614,6 +618,10 @@ static int cpufreq_governor_interactive(struct cpufreq_policy *policy,
 			pcpu->target_set_time_in_idle =
 				get_cpu_idle_time_us(j,
 					     &pcpu->target_set_time);
+			pcpu->target_validate_time =
+				pcpu->target_set_time;
+			pcpu->target_validate_time_in_idle =
+				pcpu->target_set_time_in_idle;
 			pcpu->governor_enabled = 1;
 			smp_wmb();
 		}
-- 
1.7.9.5

--
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/

[toc] | [prev] | [next] | [standalone]

#1257004 — [PATCH 60/70] cpufreq: interactive: hold reference on global cpufreq kobject if needed

From	Bálint Czobor <czoborbalint@gmail.com>
Date	2015-10-27 18:40 +0100
Subject	[PATCH 60/70] cpufreq: interactive: hold reference on global cpufreq kobject if needed
Message-ID	<qogo2-84h-45@gated-at.bofh.it>
In reply to	#1256993

From: Greg Hackmann <ghackmann@google.com>

2361be23666232dbb4851a527f466c4cbf5340fc changed cpufreq to add the
global cpufreq kobject to sysfs on demand.

To ensure this happens, cpufreq_interactive must hold a reference on
this object on devices where it intends to use it (i.e., devices where
have_governor_per_policy() returns false).  Otherwise a parentless
kobject will be passed to sysfs_create_group() which will subsequently
BUG().

Change-Id: I7dd03956e1d3c6c3c0cc17c799882c235804ae09
Signed-off-by: Greg Hackmann <ghackmann@google.com>
Signed-off-by: Bálint Czobor <czoborbalint@gmail.com>
---
 drivers/cpufreq/cpufreq_interactive.c |    8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)

diff --git a/drivers/cpufreq/cpufreq_interactive.c b/drivers/cpufreq/cpufreq_interactive.c
index ff77b30..786fd01 100644
--- a/drivers/cpufreq/cpufreq_interactive.c
+++ b/drivers/cpufreq/cpufreq_interactive.c
@@ -1154,8 +1154,10 @@ static int cpufreq_governor_interactive(struct cpufreq_policy *policy,
 		spin_lock_init(&tunables->above_hispeed_delay_lock);
 
 		policy->governor_data = tunables;
-		if (!have_governor_per_policy())
+		if (!have_governor_per_policy()) {
 			common_tunables = tunables;
+			WARN_ON(cpufreq_get_global_kobject());
+		}
 
 		rc = sysfs_create_group(get_governor_parent_kobj(policy),
 				get_sysfs_attr());
@@ -1185,6 +1187,10 @@ static int cpufreq_governor_interactive(struct cpufreq_policy *policy,
 
 			sysfs_remove_group(get_governor_parent_kobj(policy),
 					get_sysfs_attr());
+
+			if (!have_governor_per_policy())
+				cpufreq_put_global_kobject();
+
 			kfree(tunables);
 			common_tunables = NULL;
 		}
-- 
1.7.9.5

--
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/

[toc] | [prev] | [next] | [standalone]

#1257005 — [PATCH 24/70] cpufreq: interactive: use deferrable timer by default

From	Bálint Czobor <czoborbalint@gmail.com>
Date	2015-10-27 18:40 +0100
Subject	[PATCH 24/70] cpufreq: interactive: use deferrable timer by default
Message-ID	<qogo2-84h-47@gated-at.bofh.it>
In reply to	#1256993

From: Lianwei Wang <a22439@motorola.com>

Avoid wakeups only to handle the governor timer when the system is otherwise
idle.

For platforms where the power cost of remaining in idle at higher CPU
speed may outweigh the cost of a governor wakeup from idle to lower the speed,
set parameter cpufreq_interactive.governidle=1.

Change-Id: Id6c43eb35caecf9b0574fcdd5b769711bc7e6de6
Signed-off-by: LianWei WANG <a22439@motorola.com>
Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Bálint Czobor <czoborbalint@gmail.com>
---
 drivers/cpufreq/cpufreq_interactive.c |   22 +++++++++++++++-------
 1 file changed, 15 insertions(+), 7 deletions(-)

diff --git a/drivers/cpufreq/cpufreq_interactive.c b/drivers/cpufreq/cpufreq_interactive.c
index e53eae2..71887ca 100644
--- a/drivers/cpufreq/cpufreq_interactive.c
+++ b/drivers/cpufreq/cpufreq_interactive.c
@@ -20,6 +20,7 @@
 #include <linux/cpumask.h>
 #include <linux/cpufreq.h>
 #include <linux/module.h>
+#include <linux/moduleparam.h>
 #include <linux/mutex.h>
 #include <linux/sched.h>
 #include <linux/sched/rt.h>
@@ -92,6 +93,11 @@ static unsigned long above_hispeed_delay_val;
 
 static int boost_val;
 
+static bool governidle;
+module_param(governidle, bool, S_IWUSR | S_IRUGO);
+MODULE_PARM_DESC(governidle,
+	"Set to 1 to wake up CPUs from idle to reduce speed (default 0)");
+
 static int cpufreq_governor_interactive(struct cpufreq_policy *policy,
 		unsigned int event);
 
@@ -240,10 +246,11 @@ rearm_if_notmax:
 rearm:
 	if (!timer_pending(&pcpu->cpu_timer)) {
 		/*
-		 * If already at min, cancel the timer if that CPU goes idle.
-		 * We don't need to re-evaluate speed until the next idle exit.
+		 * If governing speed in idle and already at min, cancel the
+		 * timer if that CPU goes idle.  We don't need to re-evaluate
+		 * speed until the next idle exit.
 		 */
-		if (pcpu->target_freq == pcpu->policy->min)
+		if (governidle && pcpu->target_freq == pcpu->policy->min)
 			pcpu->timer_idlecancel = 1;
 
 		pcpu->time_in_idle = get_cpu_idle_time_us(
@@ -268,7 +275,6 @@ static void cpufreq_interactive_idle_start(void)
 	pending = timer_pending(&pcpu->cpu_timer);
 
 	if (pcpu->target_freq != pcpu->policy->min) {
-#ifdef CONFIG_SMP
 		/*
 		 * Entering idle while not at lowest speed.  On some
 		 * platforms this can hold the other CPU(s) at that speed
@@ -285,8 +291,7 @@ static void cpufreq_interactive_idle_start(void)
 				&pcpu->cpu_timer,
 				jiffies + usecs_to_jiffies(timer_rate));
 		}
-#endif
-	} else {
+	} else if (governidle) {
 		/*
 		 * If at min speed and entering idle after load has
 		 * already been evaluated, and a timer has been set just in
@@ -708,7 +713,10 @@ static int __init cpufreq_interactive_init(void)
 	/* Initalize per-cpu timers */
 	for_each_possible_cpu(i) {
 		pcpu = &per_cpu(cpuinfo, i);
-		init_timer(&pcpu->cpu_timer);
+		if (governidle)
+			init_timer(&pcpu->cpu_timer);
+		else
+			init_timer_deferrable(&pcpu->cpu_timer);
 		pcpu->cpu_timer.function = cpufreq_interactive_timer;
 		pcpu->cpu_timer.data = i;
 	}
-- 
1.7.9.5

--
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/

[toc] | [prev] | [next] | [standalone]

#1257006 — [PATCH 61/70] cpufreq: interactive: restructure CPUFREQ_GOV_LIMITS

From	Bálint Czobor <czoborbalint@gmail.com>
Date	2015-10-27 18:40 +0100
Subject	[PATCH 61/70] cpufreq: interactive: restructure CPUFREQ_GOV_LIMITS
Message-ID	<qogo3-84h-49@gated-at.bofh.it>
In reply to	#1256993

From: Badhri Jagan Sridharan <Badhri@google.com>

The cpufreq_interactive_timer gets cancelled and rescheduled
whenever the cpufreq_policy is changed. When the cpufreq policy is
changed at a rate faster than the sampling_rate of the interactive
governor, then the governor misses to change the target frequency
for long duration. The patch removes the need of cancelling the
timers when policy->min is changed.

Signed-off-by: Badhri Jagan Sridharan <Badhri@google.com>
Change-Id: Ibd98d151e1c73b8bd969484583ff98ee9f1135ef
Signed-off-by: Bálint Czobor <czoborbalint@gmail.com>
---
 drivers/cpufreq/cpufreq_interactive.c |   48 ++++++++++++++++++++++++---------
 1 file changed, 35 insertions(+), 13 deletions(-)

diff --git a/drivers/cpufreq/cpufreq_interactive.c b/drivers/cpufreq/cpufreq_interactive.c
index 786fd01..0e3e45e 100644
--- a/drivers/cpufreq/cpufreq_interactive.c
+++ b/drivers/cpufreq/cpufreq_interactive.c
@@ -44,8 +44,10 @@ struct cpufreq_interactive_cpuinfo {
 	u64 cputime_speedadj_timestamp;
 	struct cpufreq_policy *policy;
 	struct cpufreq_frequency_table *freq_table;
+	spinlock_t target_freq_lock; /*protects target freq */
 	unsigned int target_freq;
 	unsigned int floor_freq;
+	unsigned int max_freq;
 	u64 floor_validate_time;
 	u64 hispeed_validate_time;
 	struct rw_semaphore enable_sem;
@@ -358,6 +360,7 @@ static void cpufreq_interactive_timer(unsigned long data)
 	if (WARN_ON_ONCE(!delta_time))
 		goto rearm;
 
+	spin_lock_irqsave(&pcpu->target_freq_lock, flags);
 	do_div(cputime_speedadj, delta_time);
 	loadadjfreq = (unsigned int)cputime_speedadj * 100;
 	cpu_load = loadadjfreq / pcpu->target_freq;
@@ -383,6 +386,7 @@ static void cpufreq_interactive_timer(unsigned long data)
 		trace_cpufreq_interactive_notyet(
 			data, cpu_load, pcpu->target_freq,
 			pcpu->policy->cur, new_freq);
+		spin_unlock_irqrestore(&pcpu->target_freq_lock, flags);
 		goto rearm;
 	}
 
@@ -390,8 +394,10 @@ static void cpufreq_interactive_timer(unsigned long data)
 
 	if (cpufreq_frequency_table_target(pcpu->policy, pcpu->freq_table,
 					   new_freq, CPUFREQ_RELATION_L,
-					   &index))
+					   &index)) {
+		spin_unlock_irqrestore(&pcpu->target_freq_lock, flags);
 		goto rearm;
+	}
 
 	new_freq = pcpu->freq_table[index].frequency;
 
@@ -405,6 +411,7 @@ static void cpufreq_interactive_timer(unsigned long data)
 			trace_cpufreq_interactive_notyet(
 				data, cpu_load, pcpu->target_freq,
 				pcpu->policy->cur, new_freq);
+			spin_unlock_irqrestore(&pcpu->target_freq_lock, flags);
 			goto rearm;
 		}
 	}
@@ -426,6 +433,7 @@ static void cpufreq_interactive_timer(unsigned long data)
 		trace_cpufreq_interactive_already(
 			data, cpu_load, pcpu->target_freq,
 			pcpu->policy->cur, new_freq);
+		spin_unlock_irqrestore(&pcpu->target_freq_lock, flags);
 		goto rearm_if_notmax;
 	}
 
@@ -433,6 +441,7 @@ static void cpufreq_interactive_timer(unsigned long data)
 					 pcpu->policy->cur, new_freq);
 
 	pcpu->target_freq = new_freq;
+	spin_unlock_irqrestore(&pcpu->target_freq_lock, flags);
 	spin_lock_irqsave(&speedchange_cpumask_lock, flags);
 	cpumask_set_cpu(data, &speedchange_cpumask);
 	spin_unlock_irqrestore(&speedchange_cpumask_lock, flags);
@@ -576,16 +585,17 @@ static void cpufreq_interactive_boost(void)
 {
 	int i;
 	int anyboost = 0;
-	unsigned long flags;
+	unsigned long flags[2];
 	struct cpufreq_interactive_cpuinfo *pcpu;
 	struct cpufreq_interactive_tunables *tunables;
 
-	spin_lock_irqsave(&speedchange_cpumask_lock, flags);
+	spin_lock_irqsave(&speedchange_cpumask_lock, flags[0]);
 
 	for_each_online_cpu(i) {
 		pcpu = &per_cpu(cpuinfo, i);
 		tunables = pcpu->policy->governor_data;
 
+		spin_lock_irqsave(&pcpu->target_freq_lock, flags[1]);
 		if (pcpu->target_freq < tunables->hispeed_freq) {
 			pcpu->target_freq = tunables->hispeed_freq;
 			cpumask_set_cpu(i, &speedchange_cpumask);
@@ -601,9 +611,10 @@ static void cpufreq_interactive_boost(void)
 
 		pcpu->floor_freq = tunables->hispeed_freq;
 		pcpu->floor_validate_time = ktime_to_us(ktime_get());
+		spin_unlock_irqrestore(&pcpu->target_freq_lock, flags[1]);
 	}
 
-	spin_unlock_irqrestore(&speedchange_cpumask_lock, flags);
+	spin_unlock_irqrestore(&speedchange_cpumask_lock, flags[0]);
 
 	if (anyboost)
 		wake_up_process(speedchange_task);
@@ -1114,6 +1125,7 @@ static int cpufreq_governor_interactive(struct cpufreq_policy *policy,
 	struct cpufreq_interactive_cpuinfo *pcpu;
 	struct cpufreq_frequency_table *freq_table;
 	struct cpufreq_interactive_tunables *tunables;
+	unsigned long flags;
 
 	if (have_governor_per_policy())
 		tunables = policy->governor_data;
@@ -1215,6 +1227,7 @@ static int cpufreq_governor_interactive(struct cpufreq_policy *policy,
 				ktime_to_us(ktime_get());
 			pcpu->hispeed_validate_time =
 				pcpu->floor_validate_time;
+			pcpu->max_freq = policy->max;
 			down_write(&pcpu->enable_sem);
 			del_timer_sync(&pcpu->cpu_timer);
 			del_timer_sync(&pcpu->cpu_slack_timer);
@@ -1250,29 +1263,37 @@ static int cpufreq_governor_interactive(struct cpufreq_policy *policy,
 		for_each_cpu(j, policy->cpus) {
 			pcpu = &per_cpu(cpuinfo, j);
 
-			/* hold write semaphore to avoid race */
-			down_write(&pcpu->enable_sem);
+			down_read(&pcpu->enable_sem);
 			if (pcpu->governor_enabled == 0) {
-				up_write(&pcpu->enable_sem);
+				up_read(&pcpu->enable_sem);
 				continue;
 			}
 
-			/* update target_freq firstly */
+			spin_lock_irqsave(&pcpu->target_freq_lock, flags);
 			if (policy->max < pcpu->target_freq)
 				pcpu->target_freq = policy->max;
 			else if (policy->min > pcpu->target_freq)
 				pcpu->target_freq = policy->min;
 
-			/* Reschedule timer.
+			spin_unlock_irqrestore(&pcpu->target_freq_lock, flags);
+			up_read(&pcpu->enable_sem);
+
+			/* Reschedule timer only if policy->max is raised.
 			 * Delete the timers, else the timer callback may
 			 * return without re-arm the timer when failed
 			 * acquire the semaphore. This race may cause timer
 			 * stopped unexpectedly.
 			 */
-			del_timer_sync(&pcpu->cpu_timer);
-			del_timer_sync(&pcpu->cpu_slack_timer);
-			cpufreq_interactive_timer_start(tunables, j);
-			up_write(&pcpu->enable_sem);
+
+			if (policy->max > pcpu->max_freq) {
+				down_write(&pcpu->enable_sem);
+				del_timer_sync(&pcpu->cpu_timer);
+				del_timer_sync(&pcpu->cpu_slack_timer);
+				cpufreq_interactive_timer_start(tunables, j);
+				up_write(&pcpu->enable_sem);
+			}
+
+			pcpu->max_freq = policy->max;
 		}
 		break;
 	}
@@ -1308,6 +1329,7 @@ static int __init cpufreq_interactive_init(void)
 		init_timer(&pcpu->cpu_slack_timer);
 		pcpu->cpu_slack_timer.function = cpufreq_interactive_nop_timer;
 		spin_lock_init(&pcpu->load_lock);
+		spin_lock_init(&pcpu->target_freq_lock);
 		init_rwsem(&pcpu->enable_sem);
 	}
 
-- 
1.7.9.5

--
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/

[toc] | [prev] | [next] | [standalone]

#1257007 — [PATCH 27/70] cpufreq: interactive: change speed according to current speed and target load

From	Bálint Czobor <czoborbalint@gmail.com>
Date	2015-10-27 18:40 +0100
Subject	[PATCH 27/70] cpufreq: interactive: change speed according to current speed and target load
Message-ID	<qogo3-84h-51@gated-at.bofh.it>
In reply to	#1256993

From: Todd Poynor <toddpoynor@google.com>

Add a target_load attribute that specifies how aggressively the governor is
to adjust speed to meet the observed load.  New target speed is calculated
as the current actual speed (may be higher than target speed on SMP) times
the CPU load (as a fraction) divided by target load (fraction).

cpufreq_frequency_table_target() call use CPUFREQ_RELATION_L to set
the next higher speed rather than next lower speed.

Change-Id: If432451da82f5fed12e15c9421d7d27792376150
Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Bálint Czobor <czoborbalint@gmail.com>
---
 drivers/cpufreq/cpufreq_interactive.c |   35 ++++++++++++++++++++++++++++++---
 1 file changed, 32 insertions(+), 3 deletions(-)

diff --git a/drivers/cpufreq/cpufreq_interactive.c b/drivers/cpufreq/cpufreq_interactive.c
index 2b4aad9..c4293c5 100644
--- a/drivers/cpufreq/cpufreq_interactive.c
+++ b/drivers/cpufreq/cpufreq_interactive.c
@@ -68,6 +68,10 @@ static unsigned int hispeed_freq;
 #define DEFAULT_GO_HISPEED_LOAD 85
 static unsigned long go_hispeed_load;
 
+/* Target load.  Lower values result in higher CPU speeds. */
+#define DEFAULT_TARGET_LOAD 90
+static unsigned long target_load = DEFAULT_TARGET_LOAD;
+
 /*
  * The minimum amount of time to spend at a frequency before we can ramp down.
  */
@@ -177,7 +181,7 @@ static void cpufreq_interactive_timer(unsigned long data)
 		    hispeed_freq < pcpu->policy->max) {
 			new_freq = hispeed_freq;
 		} else {
-			new_freq = pcpu->policy->max * cpu_load / 100;
+			new_freq = pcpu->policy->cur * cpu_load / target_load;
 
 			if (new_freq < hispeed_freq)
 				new_freq = hispeed_freq;
@@ -193,14 +197,14 @@ static void cpufreq_interactive_timer(unsigned long data)
 			}
 		}
 	} else {
-		new_freq = hispeed_freq * cpu_load / 100;
+		new_freq = pcpu->policy->cur * cpu_load / target_load;
 	}
 
 	if (new_freq <= hispeed_freq)
 		pcpu->hispeed_validate_time = now;
 
 	if (cpufreq_frequency_table_target(pcpu->policy, pcpu->freq_table,
-					   new_freq, CPUFREQ_RELATION_H,
+					   new_freq, CPUFREQ_RELATION_L,
 					   &index)) {
 		pr_warn_once("timer %d: cpufreq_frequency_table_target error\n",
 			     (int) data);
@@ -420,6 +424,30 @@ static void cpufreq_interactive_boost(void)
 		wake_up_process(speedchange_task);
 }
 
+static ssize_t show_target_load(
+	struct kobject *kobj, struct attribute *attr, char *buf)
+{
+	return sprintf(buf, "%lu\n", target_load);
+}
+
+static ssize_t store_target_load(
+	struct kobject *kobj, struct attribute *attr, const char *buf,
+	size_t count)
+{
+	int ret;
+	unsigned long val;
+
+	ret = strict_strtoul(buf, 0, &val);
+	if (ret < 0)
+		return ret;
+	target_load = val;
+	return count;
+}
+
+static struct global_attr target_load_attr =
+	__ATTR(target_load, S_IRUGO | S_IWUSR,
+		show_target_load, store_target_load);
+
 static ssize_t show_hispeed_freq(struct kobject *kobj,
 				 struct attribute *attr, char *buf)
 {
@@ -581,6 +609,7 @@ static struct global_attr boostpulse =
 	__ATTR(boostpulse, 0200, NULL, store_boostpulse);
 
 static struct attribute *interactive_attributes[] = {
+	&target_load_attr.attr,
 	&hispeed_freq_attr.attr,
 	&go_hispeed_load_attr.attr,
 	&above_hispeed_delay.attr,
-- 
1.7.9.5

--
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/

[toc] | [prev] | [next] | [standalone]

#1257009 — [PATCH 07/70] cpufreq: interactive: adjust code and documentation to match

From	Bálint Czobor <czoborbalint@gmail.com>
Date	2015-10-27 18:40 +0100
Subject	[PATCH 07/70] cpufreq: interactive: adjust code and documentation to match
Message-ID	<qogo3-84h-59@gated-at.bofh.it>
In reply to	#1256993

From: Todd Poynor <toddpoynor@google.com>

Change-Id: If59c668d514a29febe5c35404fd9d01df8548eb1
Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Bálint Czobor <czoborbalint@gmail.com>
---
 Documentation/cpu-freq/governors.txt  |   16 +++++++++++++---
 drivers/cpufreq/cpufreq_interactive.c |    6 +++---
 2 files changed, 16 insertions(+), 6 deletions(-)

diff --git a/Documentation/cpu-freq/governors.txt b/Documentation/cpu-freq/governors.txt
index b262c53..40894da 100644
--- a/Documentation/cpu-freq/governors.txt
+++ b/Documentation/cpu-freq/governors.txt
@@ -249,11 +249,21 @@ frequency before ramping down. This is to ensure that the governor has
 seen enough historic cpu load data to determine the appropriate
 workload.  Default is 80000 uS.
 
-go_maxspeed_load: The CPU load at which to ramp to max speed.  Default
-is 85.
+hispeed_freq: An intermediate "hi speed" at which to initially ramp
+when CPU load hits the value specified in go_hispeed_load.  If load
+stays high for the amount of time specified in above_hispeed_delay,
+then speed may be bumped higher.  Default is maximum speed.
+
+go_hispeed_load: The CPU load at which to ramp to the intermediate "hi
+speed".  Default is 85%.
+
+above_hispeed_delay: Once speed is set to hispeed_freq, wait for this
+long before bumping speed higher in response to continued high load.
+Default is 20000 uS.
 
 timer_rate: Sample rate for reevaluating cpu load when the system is
-not idle.  Default is 30000 uS.
+not idle.  Default is 20000 uS.
+
 
 3. The Governor Interface in the CPUfreq Core
 =============================================
diff --git a/drivers/cpufreq/cpufreq_interactive.c b/drivers/cpufreq/cpufreq_interactive.c
index 154f7bd..bf8d772 100644
--- a/drivers/cpufreq/cpufreq_interactive.c
+++ b/drivers/cpufreq/cpufreq_interactive.c
@@ -68,19 +68,19 @@ static struct mutex set_speed_lock;
 static u64 hispeed_freq;
 
 /* Go to hi speed when CPU load at or above this value. */
-#define DEFAULT_GO_HISPEED_LOAD 95
+#define DEFAULT_GO_HISPEED_LOAD 85
 static unsigned long go_hispeed_load;
 
 /*
  * The minimum amount of time to spend at a frequency before we can ramp down.
  */
-#define DEFAULT_MIN_SAMPLE_TIME 20 * USEC_PER_MSEC
+#define DEFAULT_MIN_SAMPLE_TIME (80 * USEC_PER_MSEC)
 static unsigned long min_sample_time;
 
 /*
  * The sample rate of the timer used to increase frequency
  */
-#define DEFAULT_TIMER_RATE 20 * USEC_PER_MSEC
+#define DEFAULT_TIMER_RATE (20 * USEC_PER_MSEC)
 static unsigned long timer_rate;
 
 /*
-- 
1.7.9.5

--
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/

[toc] | [prev] | [next] | [standalone]

#1257010 — [PATCH 29/70] cpufreq: interactive: allow arbitrary speed / target load mappings

From	Bálint Czobor <czoborbalint@gmail.com>
Date	2015-10-27 18:40 +0100
Subject	[PATCH 29/70] cpufreq: interactive: allow arbitrary speed / target load mappings
Message-ID	<qogo3-84h-65@gated-at.bofh.it>
In reply to	#1256993

From: Todd Poynor <toddpoynor@google.com>

Accept a string of target loads and speeds at which to apply the
target loads, per the documentation update in this patch.  For example,
"85 1000000:90 1700000:99" targets CPU load 85% below speed 1GHz,  90%
at or above 1GHz, until 1.7GHz and above, at which load 99% is targeted.

Attempt to avoid oscillations by evaluating the current speed
weighted by current load against each new choice of speed, choosing a
higher speed if the current load requires a higher speed.

Change-Id: Ie3300206047c84eca5a26b0b63ea512e5207550e
Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Bálint Czobor <czoborbalint@gmail.com>
---
 Documentation/cpu-freq/governors.txt  |   17 +++
 drivers/cpufreq/cpufreq_interactive.c |  186 ++++++++++++++++++++++++++++++---
 2 files changed, 189 insertions(+), 14 deletions(-)

diff --git a/Documentation/cpu-freq/governors.txt b/Documentation/cpu-freq/governors.txt
index 509b179..b15f6d2 100644
--- a/Documentation/cpu-freq/governors.txt
+++ b/Documentation/cpu-freq/governors.txt
@@ -244,6 +244,23 @@ short-term load since idle exit to determine the cpu speed to ramp to.
 
 The tuneable values for this governor are:
 
+target_loads: CPU load values used to adjust speed to influence the
+current CPU load toward that value.  In general, the lower the target
+load, the more often the governor will raise CPU speeds to bring load
+below the target.  The format is a single target load, optionally
+followed by pairs of CPU speeds and CPU loads to target at or above
+those speeds.  Colons can be used between the speeds and associated
+target loads for readability.  For example:
+
+   85 1000000:90 1700000:99
+
+targets CPU load 85% below speed 1GHz, 90% at or above 1GHz, until
+1.7GHz and above, at which load 99% is targeted.  If speeds are
+specified these must appear in ascending order.  Higher target load
+values are typically specified for higher speeds, that is, target load
+values also usually appear in an ascending order. The default is
+target load 90% for all speeds.
+
 min_sample_time: The minimum amount of time to spend at the current
 frequency before ramping down. This is to ensure that the governor has
 seen enough historic cpu load data to determine the appropriate
diff --git a/drivers/cpufreq/cpufreq_interactive.c b/drivers/cpufreq/cpufreq_interactive.c
index 17c42cc..6ea77a0 100644
--- a/drivers/cpufreq/cpufreq_interactive.c
+++ b/drivers/cpufreq/cpufreq_interactive.c
@@ -70,7 +70,10 @@ static unsigned long go_hispeed_load;
 
 /* Target load.  Lower values result in higher CPU speeds. */
 #define DEFAULT_TARGET_LOAD 90
-static unsigned long target_load = DEFAULT_TARGET_LOAD;
+static unsigned int default_target_loads[] = {DEFAULT_TARGET_LOAD};
+static spinlock_t target_loads_lock;
+static unsigned int *target_loads = default_target_loads;
+static int ntarget_loads = ARRAY_SIZE(default_target_loads);
 
 /*
  * The minimum amount of time to spend at a frequency before we can ramp down.
@@ -125,6 +128,110 @@ static void cpufreq_interactive_timer_resched(
 				     &pcpu->time_in_idle_timestamp);
 }
 
+static unsigned int freq_to_targetload(unsigned int freq)
+{
+	int i;
+	unsigned int ret;
+
+	spin_lock(&target_loads_lock);
+
+	for (i = 0; i < ntarget_loads - 1 && freq >= target_loads[i+1]; i += 2)
+		;
+
+	ret = target_loads[i];
+	spin_unlock(&target_loads_lock);
+	return ret;
+}
+
+/*
+ * If increasing frequencies never map to a lower target load then
+ * choose_freq() will find the minimum frequency that does not exceed its
+ * target load given the current load.
+ */
+
+static unsigned int choose_freq(
+	struct cpufreq_interactive_cpuinfo *pcpu, unsigned int curload)
+{
+	unsigned int freq = pcpu->policy->cur;
+	unsigned int loadadjfreq = freq * curload;
+	unsigned int prevfreq, freqmin, freqmax;
+	unsigned int tl;
+	int index;
+
+	freqmin = 0;
+	freqmax = UINT_MAX;
+
+	do {
+		prevfreq = freq;
+		tl = freq_to_targetload(freq);
+
+		/*
+		 * Find the lowest frequency where the computed load is less
+		 * than or equal to the target load.
+		 */
+
+		cpufreq_frequency_table_target(
+			pcpu->policy, pcpu->freq_table, loadadjfreq / tl,
+			CPUFREQ_RELATION_L, &index);
+		freq = pcpu->freq_table[index].frequency;
+
+		if (freq > prevfreq) {
+			/* The previous frequency is too low. */
+			freqmin = prevfreq;
+
+			if (freq >= freqmax) {
+				/*
+				 * Find the highest frequency that is less
+				 * than freqmax.
+				 */
+				cpufreq_frequency_table_target(
+					pcpu->policy, pcpu->freq_table,
+					freqmax - 1, CPUFREQ_RELATION_H,
+					&index);
+				freq = pcpu->freq_table[index].frequency;
+
+				if (freq == freqmin) {
+					/*
+					 * The first frequency below freqmax
+					 * has already been found to be too
+					 * low.  freqmax is the lowest speed
+					 * we found that is fast enough.
+					 */
+					freq = freqmax;
+					break;
+				}
+			}
+		} else if (freq < prevfreq) {
+			/* The previous frequency is high enough. */
+			freqmax = prevfreq;
+
+			if (freq <= freqmin) {
+				/*
+				 * Find the lowest frequency that is higher
+				 * than freqmin.
+				 */
+				cpufreq_frequency_table_target(
+					pcpu->policy, pcpu->freq_table,
+					freqmin + 1, CPUFREQ_RELATION_L,
+					&index);
+				freq = pcpu->freq_table[index].frequency;
+
+				/*
+				 * If freqmax is the first frequency above
+				 * freqmin then we have already found that
+				 * this speed is fast enough.
+				 */
+				if (freq == freqmax)
+					break;
+			}
+		}
+
+		/* If same frequency chosen as previous then done. */
+	} while (freq != prevfreq);
+
+	return freq;
+}
+
 static void cpufreq_interactive_timer(unsigned long data)
 {
 	u64 now;
@@ -180,7 +287,7 @@ static void cpufreq_interactive_timer(unsigned long data)
 	    pcpu->target_freq < hispeed_freq)
 		new_freq = hispeed_freq;
 	else
-		new_freq = pcpu->policy->cur * cpu_load / target_load;
+		new_freq = choose_freq(pcpu, cpu_load);
 
 	if (pcpu->target_freq >= hispeed_freq &&
 	    new_freq > pcpu->target_freq &&
@@ -414,29 +521,79 @@ static void cpufreq_interactive_boost(void)
 		wake_up_process(speedchange_task);
 }
 
-static ssize_t show_target_load(
+static ssize_t show_target_loads(
 	struct kobject *kobj, struct attribute *attr, char *buf)
 {
-	return sprintf(buf, "%lu\n", target_load);
+	int i;
+	ssize_t ret = 0;
+
+	spin_lock(&target_loads_lock);
+
+	for (i = 0; i < ntarget_loads; i++)
+		ret += sprintf(buf + ret, "%u%s", target_loads[i],
+			       i & 0x1 ? ":" : " ");
+
+	ret += sprintf(buf + ret, "\n");
+	spin_unlock(&target_loads_lock);
+	return ret;
 }
 
-static ssize_t store_target_load(
+static ssize_t store_target_loads(
 	struct kobject *kobj, struct attribute *attr, const char *buf,
 	size_t count)
 {
 	int ret;
-	unsigned long val;
+	const char *cp;
+	unsigned int *new_target_loads = NULL;
+	int ntokens = 1;
+	int i;
 
-	ret = strict_strtoul(buf, 0, &val);
-	if (ret < 0)
-		return ret;
-	target_load = val;
+	cp = buf;
+	while ((cp = strpbrk(cp + 1, " :")))
+		ntokens++;
+
+	if (!(ntokens & 0x1))
+		goto err_inval;
+
+	new_target_loads = kmalloc(ntokens * sizeof(unsigned int), GFP_KERNEL);
+	if (!new_target_loads) {
+		ret = -ENOMEM;
+		goto err;
+	}
+
+	cp = buf;
+	i = 0;
+	while (i < ntokens) {
+		if (sscanf(cp, "%u", &new_target_loads[i++]) != 1)
+			goto err_inval;
+
+		cp = strpbrk(cp, " :");
+		if (!cp)
+			break;
+		cp++;
+	}
+
+	if (i != ntokens)
+		goto err_inval;
+
+	spin_lock(&target_loads_lock);
+	if (target_loads != default_target_loads)
+		kfree(target_loads);
+	target_loads = new_target_loads;
+	ntarget_loads = ntokens;
+	spin_unlock(&target_loads_lock);
 	return count;
+
+err_inval:
+	ret = -EINVAL;
+err:
+	kfree(new_target_loads);
+	return ret;
 }
 
-static struct global_attr target_load_attr =
-	__ATTR(target_load, S_IRUGO | S_IWUSR,
-		show_target_load, store_target_load);
+static struct global_attr target_loads_attr =
+	__ATTR(target_loads, S_IRUGO | S_IWUSR,
+		show_target_loads, store_target_loads);
 
 static ssize_t show_hispeed_freq(struct kobject *kobj,
 				 struct attribute *attr, char *buf)
@@ -599,7 +756,7 @@ static struct global_attr boostpulse =
 	__ATTR(boostpulse, 0200, NULL, store_boostpulse);
 
 static struct attribute *interactive_attributes[] = {
-	&target_load_attr.attr,
+	&target_loads_attr.attr,
 	&hispeed_freq_attr.attr,
 	&go_hispeed_load_attr.attr,
 	&above_hispeed_delay.attr,
@@ -739,6 +896,7 @@ static int __init cpufreq_interactive_init(void)
 		pcpu->cpu_timer.data = i;
 	}
 
+	spin_lock_init(&target_loads_lock);
 	spin_lock_init(&speedchange_cpumask_lock);
 	speedchange_task =
 		kthread_create(cpufreq_interactive_speedchange_task, NULL,
-- 
1.7.9.5

--
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/

[toc] | [prev] | [next] | [standalone]

#1257011 — [PATCH 48/70] cpufreq: interactive: reduce chance of zero time delta on load eval

From	Bálint Czobor <czoborbalint@gmail.com>
Date	2015-10-27 18:40 +0100
Subject	[PATCH 48/70] cpufreq: interactive: reduce chance of zero time delta on load eval
Message-ID	<qogo3-84h-69@gated-at.bofh.it>
In reply to	#1256993

From: Todd Poynor <toddpoynor@google.com>

Reschedule load sampling timer after timestamp of sample start taken,
hold spinlock across entire sequence to avoid preemption.  Avoid the
WARN for zero time delta in the load sampling timer function.

Change-Id: Idc10a756f09141decb6df92669521a1ebf0dbc10
Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Bálint Czobor <czoborbalint@gmail.com>
---
 drivers/cpufreq/cpufreq_interactive.c |   16 +++++++++-------
 1 file changed, 9 insertions(+), 7 deletions(-)

diff --git a/drivers/cpufreq/cpufreq_interactive.c b/drivers/cpufreq/cpufreq_interactive.c
index 1b5d930..b4d12b2 100644
--- a/drivers/cpufreq/cpufreq_interactive.c
+++ b/drivers/cpufreq/cpufreq_interactive.c
@@ -169,21 +169,23 @@ static inline cputime64_t get_cpu_idle_time(unsigned int cpu,
 static void cpufreq_interactive_timer_resched(
 	struct cpufreq_interactive_cpuinfo *pcpu)
 {
-	unsigned long expires = jiffies + usecs_to_jiffies(timer_rate);
+	unsigned long expires;
 	unsigned long flags;
 
-	mod_timer_pinned(&pcpu->cpu_timer, expires);
-	if (timer_slack_val >= 0 && pcpu->target_freq > pcpu->policy->min) {
-		expires += usecs_to_jiffies(timer_slack_val);
-		mod_timer_pinned(&pcpu->cpu_slack_timer, expires);
-	}
-
 	spin_lock_irqsave(&pcpu->load_lock, flags);
 	pcpu->time_in_idle =
 		get_cpu_idle_time(smp_processor_id(),
 				     &pcpu->time_in_idle_timestamp);
 	pcpu->cputime_speedadj = 0;
 	pcpu->cputime_speedadj_timestamp = pcpu->time_in_idle_timestamp;
+	expires = jiffies + usecs_to_jiffies(timer_rate);
+	mod_timer_pinned(&pcpu->cpu_timer, expires);
+
+	if (timer_slack_val >= 0 && pcpu->target_freq > pcpu->policy->min) {
+		expires += usecs_to_jiffies(timer_slack_val);
+		mod_timer_pinned(&pcpu->cpu_slack_timer, expires);
+	}
+
 	spin_unlock_irqrestore(&pcpu->load_lock, flags);
 }
 
-- 
1.7.9.5

--
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/

[toc] | [prev] | [next] | [standalone]

#1257012 — [PATCH 64/70] cpufreq: interactive: prevents the frequency to directly raise above the hispeed_freq from a lower frequency.

From	Bálint Czobor <czoborbalint@gmail.com>
Date	2015-10-27 18:40 +0100
Subject	[PATCH 64/70] cpufreq: interactive: prevents the frequency to directly raise above the hispeed_freq from a lower frequency.
Message-ID	<qogo3-84h-71@gated-at.bofh.it>
In reply to	#1256993

From: Ruchi Kandoi <kandoiruchi@google.com>

When the load was below go_hispeed_load, there is a possibility that
choose_freq() would return a frequency which would be higher than the
hispeed_freq. According to the policy we should first jump to the
hispeed_freq, stay there for above_hispeed_delay and then be allowed to
raise higher than that.

Added a check to prevent the frequency to be directly raised to
something higher than the hispeed_freq.

Change-Id: Icda5d848dd9beadcc18835082ddf269732c75bd0
Signed-off-by: Ruchi Kandoi <kandoiruchi@google.com>
Signed-off-by: Bálint Czobor <czoborbalint@gmail.com>
---
 drivers/cpufreq/cpufreq_interactive.c |    3 +++
 1 file changed, 3 insertions(+)

diff --git a/drivers/cpufreq/cpufreq_interactive.c b/drivers/cpufreq/cpufreq_interactive.c
index 6c4b6a8..f0e6fe2 100644
--- a/drivers/cpufreq/cpufreq_interactive.c
+++ b/drivers/cpufreq/cpufreq_interactive.c
@@ -377,6 +377,9 @@ static void cpufreq_interactive_timer(unsigned long data)
 		}
 	} else {
 		new_freq = choose_freq(pcpu, loadadjfreq);
+		if (new_freq > tunables->hispeed_freq &&
+				pcpu->target_freq < tunables->hispeed_freq)
+			new_freq = tunables->hispeed_freq;
 	}
 
 	if (pcpu->target_freq >= tunables->hispeed_freq &&
-- 
1.7.9.5

--
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/

[toc] | [prev] | [next] | [standalone]

#1257013 — [PATCH 65/70] cpufreq: interactive: make common_tunables static

From	Bálint Czobor <czoborbalint@gmail.com>
Date	2015-10-27 18:40 +0100
Subject	[PATCH 65/70] cpufreq: interactive: make common_tunables static
Message-ID	<qogo3-84h-73@gated-at.bofh.it>
In reply to	#1256993

From: Todd Poynor <toddpoynor@google.com>

From: Cylen Yao <cylen.yao@mediatek.com>

common_tunables should be static.

Change-Id: I502ee3062bece5082fea7861eff2f6237e25cede
Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Bálint Czobor <czoborbalint@gmail.com>
---
 drivers/cpufreq/cpufreq_interactive.c |    2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/drivers/cpufreq/cpufreq_interactive.c b/drivers/cpufreq/cpufreq_interactive.c
index f0e6fe2..a0a7c3a 100644
--- a/drivers/cpufreq/cpufreq_interactive.c
+++ b/drivers/cpufreq/cpufreq_interactive.c
@@ -115,7 +115,7 @@ struct cpufreq_interactive_tunables {
 };
 
 /* For cases where we have single governor instance for system */
-struct cpufreq_interactive_tunables *common_tunables;
+static struct cpufreq_interactive_tunables *common_tunables;
 
 static struct attribute_group *get_sysfs_attr(void);
 
-- 
1.7.9.5

--
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/

[toc] | [prev] | [next] | [standalone]

#1257014 — [PATCH 69/70] cpufreq: interactive: Don't set floor_validate_time during boost

From	Bálint Czobor <czoborbalint@gmail.com>
Date	2015-10-27 18:40 +0100
Subject	[PATCH 69/70] cpufreq: interactive: Don't set floor_validate_time during boost
Message-ID	<qogo3-84h-75@gated-at.bofh.it>
In reply to	#1256993

From: Junjie Wu <junjiew@codeaurora.org>

Frequency selection algorithm guarantees its chosen frequency
is not lower than hispeed_freq as long as boost is enabled.

Setting floor_freq and floor_validate_time during boost could block
CPU frequency from going below hispeed_freq even after
boostpulse_duration expires, if min_sample_time is higher than
boostpulse_duration. This conflicts with the intention of commit
de091367ead15b6e95dd1d0743a18f0da5a07ee5
(cpufreq: interactive: specify duration of CPU speed boost pulse)
to allow CPU to ramp down immediately after boost expires. It also
makes boost behavior inconsistent since it depends on min_sample_time.

Avoid setting floor_freq and floor_validate_time when boost starts.

Change-Id: I12852998af46cfbfaf8661eb5e8d5301b6f631e7
Signed-off-by: Junjie Wu <junjiew@codeaurora.org>
Signed-off-by: Bálint Czobor <czoborbalint@gmail.com>
---
 drivers/cpufreq/cpufreq_interactive.c |    8 --------
 1 file changed, 8 deletions(-)

diff --git a/drivers/cpufreq/cpufreq_interactive.c b/drivers/cpufreq/cpufreq_interactive.c
index 064ab56..95525b3 100644
--- a/drivers/cpufreq/cpufreq_interactive.c
+++ b/drivers/cpufreq/cpufreq_interactive.c
@@ -606,14 +606,6 @@ static void cpufreq_interactive_boost(void)
 				ktime_to_us(ktime_get());
 			anyboost = 1;
 		}
-
-		/*
-		 * Set floor freq and (re)start timer for when last
-		 * validated.
-		 */
-
-		pcpu->floor_freq = tunables->hispeed_freq;
-		pcpu->floor_validate_time = ktime_to_us(ktime_get());
 		spin_unlock_irqrestore(&pcpu->target_freq_lock, flags[1]);
 	}
 
-- 
1.7.9.5

--
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/

[toc] | [prev] | [next] | [standalone]

#1257015 — [PATCH 37/70] cpufreq: interactive: default go_hispeed_load 99%, doc updates

From	Bálint Czobor <czoborbalint@gmail.com>
Date	2015-10-27 18:40 +0100
Subject	[PATCH 37/70] cpufreq: interactive: default go_hispeed_load 99%, doc updates
Message-ID	<qogo3-84h-77@gated-at.bofh.it>
In reply to	#1256993

From: Todd Poynor <toddpoynor@google.com>

Update default go_hispeed_load from 85% to 99%.  Recent changes to the
governor now use a default target_load of 90%.  go_hispeed_load should
not be lower than the target load for hispeed_freq, which could lead
to oscillating speed decisions.  Other recent changes reduce the need
to dampen speed jumps on load spikes, while input event boosts from
userspace are the preferred method for anticipating load spikes with
UI impacts.

General update to the documentation to reflect recent changes.

Change-Id: I1b92f3091f42c04b10503cd1169a943b5dfd6faf
Signed-off-by: Todd Poynor <toddpoynor@google.com>
Signed-off-by: Bálint Czobor <czoborbalint@gmail.com>
---
 Documentation/cpu-freq/governors.txt  |   64 ++++++++++++++++-----------------
 drivers/cpufreq/cpufreq_interactive.c |    2 +-
 2 files changed, 33 insertions(+), 33 deletions(-)

diff --git a/Documentation/cpu-freq/governors.txt b/Documentation/cpu-freq/governors.txt
index b15f6d2..4d8ae34 100644
--- a/Documentation/cpu-freq/governors.txt
+++ b/Documentation/cpu-freq/governors.txt
@@ -224,23 +224,8 @@ sampling rate.
 
 The CPUfreq governor "interactive" is designed for latency-sensitive,
 interactive workloads. This governor sets the CPU speed depending on
-usage, similar to "ondemand" and "conservative" governors.  However,
-the governor is more aggressive about scaling the CPU speed up in
-response to CPU-intensive activity.
-
-Sampling the CPU load every X ms can lead to under-powering the CPU
-for X ms, leading to dropped frames, stuttering UI, etc.  Instead of
-sampling the cpu at a specified rate, the interactive governor will
-check whether to scale the cpu frequency up soon after coming out of
-idle.  When the cpu comes out of idle, a timer is configured to fire
-within 1-2 ticks.  If the cpu is very busy between exiting idle and
-when the timer fires then we assume the cpu is underpowered and ramp
-to MAX speed.
-
-If the cpu was not sufficiently busy to immediately ramp to MAX speed,
-then governor evaluates the cpu load since the last speed adjustment,
-choosing the highest value between that longer-term load or the
-short-term load since idle exit to determine the cpu speed to ramp to.
+usage, similar to "ondemand" and "conservative" governors, but with a
+different set of configurable behaviors.
 
 The tuneable values for this governor are:
 
@@ -262,36 +247,51 @@ values also usually appear in an ascending order. The default is
 target load 90% for all speeds.
 
 min_sample_time: The minimum amount of time to spend at the current
-frequency before ramping down. This is to ensure that the governor has
-seen enough historic cpu load data to determine the appropriate
-workload.  Default is 80000 uS.
+frequency before ramping down. Default is 80000 uS.
 
 hispeed_freq: An intermediate "hi speed" at which to initially ramp
 when CPU load hits the value specified in go_hispeed_load.  If load
 stays high for the amount of time specified in above_hispeed_delay,
-then speed may be bumped higher.  Default is maximum speed.
+then speed may be bumped higher.  Default is the maximum speed
+allowed by the policy at governor initialization time.
 
-go_hispeed_load: The CPU load at which to ramp to the intermediate "hi
-speed".  Default is 85%.
+go_hispeed_load: The CPU load at which to ramp to hispeed_freq.
+Default is 99%.
 
-above_hispeed_delay: Once speed is set to hispeed_freq, wait for this
-long before bumping speed higher in response to continued high load.
+above_hispeed_delay: When speed is at or above hispeed_freq, wait for
+this long before raising speed in response to continued high load.
 Default is 20000 uS.
 
-timer_rate: Sample rate for reevaluating cpu load when the system is
-not idle.  Default is 20000 uS.
-
-input_boost: If non-zero, boost speed of all CPUs to hispeed_freq on
-touchscreen activity.  Default is 0.
+timer_rate: Sample rate for reevaluating CPU load when the CPU is not
+idle.  A deferrable timer is used, such that the CPU will not be woken
+from idle to service this timer until something else needs to run.
+(The maximum time to allow deferring this timer when not running at
+minimum speed is configurable via timer_slack.)  Default is 20000 uS.
+
+timer_slack: Maximum additional time to defer handling the governor
+sampling timer beyond timer_rate when running at speeds above the
+minimum.  For platforms that consume additional power at idle when
+CPUs are running at speeds greater than minimum, this places an upper
+bound on how long the timer will be deferred prior to re-evaluating
+load and dropping speed.  For example, if timer_rate is 20000uS and
+timer_slack is 10000uS then timers will be deferred for up to 30msec
+when not at lowest speed.  A value of -1 means defer timers
+indefinitely at all speeds.  Default is 80000 uS.
 
 boost: If non-zero, immediately boost speed of all CPUs to at least
 hispeed_freq until zero is written to this attribute.  If zero, allow
 CPU speeds to drop below hispeed_freq according to load as usual.
+Default is zero.
 
-boostpulse: Immediately boost speed of all CPUs to hispeed_freq for
-min_sample_time, after which speeds are allowed to drop below
+boostpulse: On each write, immediately boost speed of all CPUs to
+hispeed_freq for at least the period of time specified by
+boostpulse_duration, after which speeds are allowed to drop below
 hispeed_freq according to load as usual.
 
+boostpulse_duration: Length of time to hold CPU speed at hispeed_freq
+on a write to boostpulse, before allowing speed to drop according to
+load as usual.  Default is 80000 uS.
+
 
 3. The Governor Interface in the CPUfreq Core
 =============================================
diff --git a/drivers/cpufreq/cpufreq_interactive.c b/drivers/cpufreq/cpufreq_interactive.c
index d22d162..fdc2a47 100644
--- a/drivers/cpufreq/cpufreq_interactive.c
+++ b/drivers/cpufreq/cpufreq_interactive.c
@@ -66,7 +66,7 @@ static spinlock_t speedchange_cpumask_lock;
 static unsigned int hispeed_freq;
 
 /* Go to hi speed when CPU load at or above this value. */
-#define DEFAULT_GO_HISPEED_LOAD 85
+#define DEFAULT_GO_HISPEED_LOAD 99
 static unsigned long go_hispeed_load;
 
 /* Target load.  Lower values result in higher CPU speeds. */
-- 
1.7.9.5

--
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/

[toc] | [prev] | [next] | [standalone]

Page 1 of 4 [1] 2 3 4 Next page →

csiph-web