On Wed, Apr 24, 2013 at 5:12 PM, Viresh Kumar viresh.kumar@linaro.org wrote:

Workqueues can be performance or power oriented. For performance we may want to keep them running on a single cpu, so that it remains cache hot. For power we can give scheduler the liberty to choose target cpu for running work handler.

Consider the following rewording:

Workqueues can be performance or power-oriented. Currently, most workqueues are bound to the CPU they were created on. This gives good performance (due to cache effects) at the cost of potentially waking up otherwise idle cores just to process some work. To save power, we can allow the work to be rescheduled on a core that is already awake.

Later one (Power oriented WQ) can be achieved if the workqueue is allocated with WQ_UNBOUND flag. Enabling CONFIG_WQ_POWER_EFFICIENT will set 'wq_power_efficient' to 'true'. Setting 'power_efficient' boot param will override value of 'wq_power_efficient' variable. When 'wq_power_efficient' is set to 'true', we will convert WQ_POWER_EFFICIENT flag to WQ_UNBOUND on wq allocation. And so scheduler will have the liberty to choose where to run this work.

Consider the following rewording which is easier to understand IMO:

Workqueues created with the WQ_UNBOUND flag will allow some power savings. However, we don't change the default behaviour of the system. To enable power-saving behaviour, a new config option CONFIG_WQ_POWER_EFFICIENT needs to be turned on. This option can also be overridden by the workqueue.power_efficient boot parameter.

Signed-off-by: Viresh Kumar viresh.kumar@linaro.org

Documentation/kernel-parameters.txt | 17 +++++++++++++++++ include/linux/workqueue.h | 3 +++ kernel/power/Kconfig | 19 +++++++++++++++++++ kernel/workqueue.c | 11 +++++++++++ 4 files changed, 50 insertions(+)

diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index c4fa000..a9040fa 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -3312,6 +3312,23 @@ bytes respectively. Such letter suffixes can also be entirely omitted. that this also can be controlled per-workqueue for workqueues visible under /sys/bus/workqueue/.

•   workqueue.power_efficient
  

•                   Workqueues can be performance or power oriented. For
  

•                   performance we may want to keep them running on a single
  

•                   cpu, so that it remains cache hot. For power we can give
  

•                   scheduler the liberty to choose target cpu for running
  

•                   work handler.
  

•                   Later one (Power oriented WQ) can be achieved if the
  

•                   workqueue is allocated with WQ_UNBOUND flag. Enabling
  

•                   CONFIG_WQ_POWER_EFFICIENT will set 'wq_power_efficient'
  

•                   to 'true'. Setting 'power_efficient' boot param will
  

•                   override value of 'wq_power_efficient' variable. When
  

•                   'wq_power_efficient' is set to 'true', we will convert
  

•                   WQ_POWER_EFFICIENT flag to WQ_UNBOUND on wq allocation.
  

•                   And so scheduler will have the liberty to choose where
  

•                   to run this work.
  

Rewrite as above.

    x2apic_phys     [X86-64,APIC] Use x2apic physical mode instead of
                    default x2apic cluster mode on platforms
                    supporting x2apic.

diff --git a/include/linux/workqueue.h b/include/linux/workqueue.h index 623488f..83fa570 100644 --- a/include/linux/workqueue.h +++ b/include/linux/workqueue.h @@ -302,6 +302,9 @@ enum { WQ_HIGHPRI = 1 << 4, /* high priority */ WQ_CPU_INTENSIVE = 1 << 5, /* cpu instensive workqueue */ WQ_SYSFS = 1 << 6, /* visible in sysfs, see wq_sysfs_register() */

•   WQ_POWER_EFFICIENT      = 1 << 7, /* WQ_UNBOUND, for power
  

•                                      * saving, if wq_power_efficient is
  

•                                      * enabled. Unused otherwise. */
  
    __WQ_DRAINING           = 1 << 16, /* internal: workqueue is draining */
    __WQ_ORDERED            = 1 << 17, /* internal: workqueue is ordered */
  

diff --git a/kernel/power/Kconfig b/kernel/power/Kconfig index 5dfdc9e..e1e9c8b 100644 --- a/kernel/power/Kconfig +++ b/kernel/power/Kconfig @@ -263,6 +263,25 @@ config PM_GENERIC_DOMAINS bool depends on PM

+config WQ_POWER_EFFICIENT

•   bool "Workqueue allocated as UNBOUND (by default) for power efficiency"
  

•   depends on PM
  

default n

•   help
  

•     Workqueues can be performance or power oriented. For performance we
  

•     may want to keep them running on a single cpu, so that it remains
  

•     cache hot. For power we can give scheduler the liberty to choose
  

•     target cpu for running work handler.
  

•     Later one (Power oriented WQ) can be achieved if the workqueue is
  

•     allocated with WQ_UNBOUND flag. Enabling CONFIG_WQ_POWER_EFFICIENT
  

•     will set 'wq_power_efficient' to 'true'. Setting 'power_efficient'
  

•     boot param will override value of 'wq_power_efficient' variable. When
  

•     'wq_power_efficient' is set to 'true', we will convert
  

•     WQ_POWER_EFFICIENT flag to WQ_UNBOUND on wq allocation. And so
  

•     scheduler will have the liberty to choose where to run this work.
  

•     If in doubt, say N.
  

Rewrite as above

config PM_GENERIC_DOMAINS_SLEEP def_bool y depends on PM_SLEEP && PM_GENERIC_DOMAINS diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 4aa9f5b..a327027 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -272,6 +272,14 @@ static cpumask_var_t *wq_numa_possible_cpumask; static bool wq_disable_numa; module_param_named(disable_numa, wq_disable_numa, bool, 0444);

+#ifdef CONFIG_WQ_POWER_EFFICIENT +static bool wq_power_efficient = true; +#else +static bool wq_power_efficient; +#endif

+module_param_named(power_efficient, wq_power_efficient, bool, 0444);

static bool wq_numa_enabled; /* unbound NUMA affinity enabled */

/* buf for wq_update_unbound_numa_attrs(), protected by CPU hotplug exclusion */ @@ -4085,6 +4093,9 @@ struct workqueue_struct *__alloc_workqueue_key(const char *fmt, struct workqueue_struct *wq; struct pool_workqueue *pwq;

•   if ((flags & WQ_POWER_EFFICIENT) && wq_power_efficient)
  

•           flags |= WQ_UNBOUND;
  

•   /* allocate wq and format name */
    if (flags & WQ_UNBOUND)
            tbl_size = wq_numa_tbl_len * sizeof(wq->numa_pwq_tbl[0]);
  

-- 1.7.12.rc2.18.g61b472e

---

linaro-kernel mailing list linaro-kernel@lists.linaro.org http://lists.linaro.org/mailman/listinfo/linaro-kernel

On 25 April 2013 09:00, Amit Kucheria amit.kucheria@linaro.org wrote:

Yes. That was my intention - preventing a prompt on existing defconfigs and there by maintaining current behavior.

Hmm... Following is the version after fixing all problems you reported. @Tejun: I have attached it too as gmail's copy-paste may break it. Please consider applying this series if it looks fine to you.

---------------x----------------x---------------------

From: Viresh Kumar viresh.kumar@linaro.org Date: Mon, 8 Apr 2013 16:45:40 +0530 Subject: [PATCH V5 1/5] workqueues: Introduce new flag WQ_POWER_EFFICIENT for power oriented workqueues

Workqueues can be performance or power-oriented. Currently, most workqueues are bound to the CPU they were created on. This gives good performance (due to cache effects) at the cost of potentially waking up otherwise idle cores just to process some work. To save power, we can allow the work to be rescheduled on a core that is already awake.

Workqueues created with the WQ_UNBOUND flag will allow some power savings. However, we don't change the default behaviour of the system. To enable power-saving behaviour, a new config option CONFIG_WQ_POWER_EFFICIENT needs to be turned on. This option can also be overridden by the workqueue.power_efficient boot parameter.

Signed-off-by: Viresh Kumar viresh.kumar@linaro.org --- Documentation/kernel-parameters.txt | 17 +++++++++++++++++ include/linux/workqueue.h | 3 +++ kernel/power/Kconfig | 19 +++++++++++++++++++ kernel/workqueue.c | 11 +++++++++++ 4 files changed, 50 insertions(+)

diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index c4fa000..22edc83 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -3312,6 +3312,23 @@ bytes respectively. Such letter suffixes can also be entirely omitted. that this also can be controlled per-workqueue for workqueues visible under /sys/bus/workqueue/.

+ workqueue.power_efficient + Workqueues can be performance or power-oriented. + Currently, most workqueues are bound to the CPU they + were created on. This gives good performance (due to + cache effects) at the cost of potentially waking up + otherwise idle cores just to process some work. To save + power, we can allow the work to be rescheduled on a core + that is already awake. + + Workqueues created with the WQ_UNBOUND flag will allow + some power savings. However, we don't change the + default behaviour of the system. To enable power-saving + behaviour, a new config option CONFIG_WQ_POWER_EFFICIENT + needs to be turned on. This option can also be + overridden by the workqueue.power_efficient boot + parameter. + x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of default x2apic cluster mode on platforms supporting x2apic. diff --git a/include/linux/workqueue.h b/include/linux/workqueue.h index 623488f..83fa570 100644 --- a/include/linux/workqueue.h +++ b/include/linux/workqueue.h @@ -302,6 +302,9 @@ enum { WQ_HIGHPRI = 1 << 4, /* high priority */ WQ_CPU_INTENSIVE = 1 << 5, /* cpu instensive workqueue */ WQ_SYSFS = 1 << 6, /* visible in sysfs, see wq_sysfs_register() */ + WQ_POWER_EFFICIENT = 1 << 7, /* WQ_UNBOUND, for power + * saving, if wq_power_efficient is + * enabled. Unused otherwise. */

__WQ_DRAINING = 1 << 16, /* internal: workqueue is draining */ __WQ_ORDERED = 1 << 17, /* internal: workqueue is ordered */ diff --git a/kernel/power/Kconfig b/kernel/power/Kconfig index 5dfdc9e..018f039 100644 --- a/kernel/power/Kconfig +++ b/kernel/power/Kconfig @@ -263,6 +263,25 @@ config PM_GENERIC_DOMAINS bool depends on PM

+config WQ_POWER_EFFICIENT + bool "Workqueue allocated as UNBOUND (by default) for power efficiency" + depends on PM + default n + help + Workqueues can be performance or power-oriented. Currently, most + workqueues are bound to the CPU they were created on. This gives good + performance (due to cache effects) at the cost of potentially waking + up otherwise idle cores just to process some work. To save power, we + can allow the work to be rescheduled on a core that is already awake. + + Workqueues created with the WQ_UNBOUND flag will allow some power + savings. However, we don't change the default behaviour of the + system. To enable power-saving behaviour, a new config option + CONFIG_WQ_POWER_EFFICIENT needs to be turned on. This option can also + be overridden by the workqueue.power_efficient boot parameter. + + If in doubt, say N. + config PM_GENERIC_DOMAINS_SLEEP def_bool y depends on PM_SLEEP && PM_GENERIC_DOMAINS diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 4aa9f5b..a327027 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -272,6 +272,14 @@ static cpumask_var_t *wq_numa_possible_cpumask; static bool wq_disable_numa; module_param_named(disable_numa, wq_disable_numa, bool, 0444);

+#ifdef CONFIG_WQ_POWER_EFFICIENT +static bool wq_power_efficient = true; +#else +static bool wq_power_efficient; +#endif + +module_param_named(power_efficient, wq_power_efficient, bool, 0444); + static bool wq_numa_enabled; /* unbound NUMA affinity enabled */

/* buf for wq_update_unbound_numa_attrs(), protected by CPU hotplug exclusion */ @@ -4085,6 +4093,9 @@ struct workqueue_struct *__alloc_workqueue_key(const char *fmt, struct workqueue_struct *wq; struct pool_workqueue *pwq;

+ if ((flags & WQ_POWER_EFFICIENT) && wq_power_efficient) + flags |= WQ_UNBOUND; + /* allocate wq and format name */ if (flags & WQ_UNBOUND) tbl_size = wq_numa_tbl_len * sizeof(wq->numa_pwq_tbl[0]);

From cee22a15052faa817e3ec8985a28154d3fabc7aa Mon Sep 17 00:00:00 2001

From: Viresh Kumar viresh.kumar@linaro.org Date: Mon, 8 Apr 2013 16:45:40 +0530 Subject: workqueues: Introduce new flag WQ_POWER_EFFICIENT for power oriented workqueues

Workqueues can be performance or power-oriented. Currently, most workqueues are bound to the CPU they were created on. This gives good performance (due to cache effects) at the cost of potentially waking up otherwise idle cores (Idle from scheduler's perspective. Which may or may not be physically idle) just to process some work. To save power, we can allow the work to be rescheduled on a core that is already awake.

Workqueues created with the WQ_UNBOUND flag will allow some power savings. However, we don't change the default behaviour of the system. To enable power-saving behaviour, a new config option CONFIG_WQ_POWER_EFFICIENT needs to be turned on. This option can also be overridden by the workqueue.power_efficient boot parameter.

tj: Updated config description and comments. Renamed CONFIG_WQ_POWER_EFFICIENT to CONFIG_WQ_POWER_EFFICIENT_DEFAULT.

Signed-off-by: Viresh Kumar viresh.kumar@linaro.org Reviewed-by: Amit Kucheria amit.kucheria@linaro.org Signed-off-by: Tejun Heo tj@kernel.org --- Documentation/kernel-parameters.txt | 15 +++++++++++++++ include/linux/workqueue.h | 27 +++++++++++++++++++++++++++ kernel/power/Kconfig | 20 ++++++++++++++++++++ kernel/workqueue.c | 13 +++++++++++++ 4 files changed, 75 insertions(+)

diff --git a/Documentation/kernel-parameters.txt b/Documentation/kernel-parameters.txt index c3bfacb..37dfd72 100644 --- a/Documentation/kernel-parameters.txt +++ b/Documentation/kernel-parameters.txt @@ -3320,6 +3320,21 @@ bytes respectively. Such letter suffixes can also be entirely omitted. that this also can be controlled per-workqueue for workqueues visible under /sys/bus/workqueue/.

+ workqueue.power_efficient + Per-cpu workqueues are generally preferred because + they show better performance thanks to cache + locality; unfortunately, per-cpu workqueues tend to + be more power hungry than unbound workqueues. + + Enabling this makes the per-cpu workqueues which + were observed to contribute significantly to power + consumption unbound, leading to measurably lower + power usage at the cost of small performance + overhead. + + The default value of this parameter is determined by + the config option CONFIG_WQ_POWER_EFFICIENT_DEFAULT. + x2apic_phys [X86-64,APIC] Use x2apic physical mode instead of default x2apic cluster mode on platforms supporting x2apic. diff --git a/include/linux/workqueue.h b/include/linux/workqueue.h index 623488f..fc0136b 100644 --- a/include/linux/workqueue.h +++ b/include/linux/workqueue.h @@ -303,6 +303,33 @@ enum { WQ_CPU_INTENSIVE = 1 << 5, /* cpu instensive workqueue */ WQ_SYSFS = 1 << 6, /* visible in sysfs, see wq_sysfs_register() */

+ /* + * Per-cpu workqueues are generally preferred because they tend to + * show better performance thanks to cache locality. Per-cpu + * workqueues exclude the scheduler from choosing the CPU to + * execute the worker threads, which has an unfortunate side effect + * of increasing power consumption. + * + * The scheduler considers a CPU idle if it doesn't have any task + * to execute and tries to keep idle cores idle to conserve power; + * however, for example, a per-cpu work item scheduled from an + * interrupt handler on an idle CPU will force the scheduler to + * excute the work item on that CPU breaking the idleness, which in + * turn may lead to more scheduling choices which are sub-optimal + * in terms of power consumption. + * + * Workqueues marked with WQ_POWER_EFFICIENT are per-cpu by default + * but become unbound if workqueue.power_efficient kernel param is + * specified. Per-cpu workqueues which are identified to + * contribute significantly to power-consumption are identified and + * marked with this flag and enabling the power_efficient mode + * leads to noticeable power saving at the cost of small + * performance disadvantage. + * + * http://thread.gmane.org/gmane.linux.kernel/1480396 + */ + WQ_POWER_EFFICIENT = 1 << 7, + __WQ_DRAINING = 1 << 16, /* internal: workqueue is draining */ __WQ_ORDERED = 1 << 17, /* internal: workqueue is ordered */

diff --git a/kernel/power/Kconfig b/kernel/power/Kconfig index 5dfdc9e..4645596 100644 --- a/kernel/power/Kconfig +++ b/kernel/power/Kconfig @@ -263,6 +263,26 @@ config PM_GENERIC_DOMAINS bool depends on PM

+config WQ_POWER_EFFICIENT_DEFAULT + bool "Enable workqueue power-efficient mode by default" + depends on PM + default n + help + Per-cpu workqueues are generally preferred because they show + better performance thanks to cache locality; unfortunately, + per-cpu workqueues tend to be more power hungry than unbound + workqueues. + + Enabling workqueue.power_efficient kernel parameter makes the + per-cpu workqueues which were observed to contribute + significantly to power consumption unbound, leading to measurably + lower power usage at the cost of small performance overhead. + + This config option determines whether workqueue.power_efficient + is enabled by default. + + If in doubt, say N. + config PM_GENERIC_DOMAINS_SLEEP def_bool y depends on PM_SLEEP && PM_GENERIC_DOMAINS diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 4aa9f5b..8068d97 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -272,6 +272,15 @@ static cpumask_var_t *wq_numa_possible_cpumask; static bool wq_disable_numa; module_param_named(disable_numa, wq_disable_numa, bool, 0444);

+/* see the comment above the definition of WQ_POWER_EFFICIENT */ +#ifdef CONFIG_WQ_POWER_EFFICIENT_DEFAULT +static bool wq_power_efficient = true; +#else +static bool wq_power_efficient; +#endif + +module_param_named(power_efficient, wq_power_efficient, bool, 0444); + static bool wq_numa_enabled; /* unbound NUMA affinity enabled */

/* buf for wq_update_unbound_numa_attrs(), protected by CPU hotplug exclusion */ @@ -4085,6 +4094,10 @@ struct workqueue_struct *__alloc_workqueue_key(const char *fmt, struct workqueue_struct *wq; struct pool_workqueue *pwq;

+ /* see the comment above the definition of WQ_POWER_EFFICIENT */ + if ((flags & WQ_POWER_EFFICIENT) && wq_power_efficient) + flags |= WQ_UNBOUND; + /* allocate wq and format name */ if (flags & WQ_UNBOUND) tbl_size = wq_numa_tbl_len * sizeof(wq->numa_pwq_tbl[0]);

This one is slightly updated too. The rest remain the same.

From 0668106ca3865ba945e155097fb042bf66d364d3 Mon Sep 17 00:00:00 2001

From: Viresh Kumar viresh.kumar@linaro.org Date: Wed, 24 Apr 2013 17:12:54 +0530 Subject: workqueue: Add system wide power_efficient workqueues

This patch adds system wide workqueues aligned towards power saving. This is done by allocating them with WQ_UNBOUND flag if 'wq_power_efficient' is set to 'true'.

tj: updated comments a bit.

Signed-off-by: Viresh Kumar viresh.kumar@linaro.org Signed-off-by: Tejun Heo tj@kernel.org --- include/linux/workqueue.h | 8 ++++++++ kernel/workqueue.c | 13 ++++++++++++- 2 files changed, 20 insertions(+), 1 deletion(-)

diff --git a/include/linux/workqueue.h b/include/linux/workqueue.h index fc0136b..a9f4119 100644 --- a/include/linux/workqueue.h +++ b/include/linux/workqueue.h @@ -360,11 +360,19 @@ enum { * * system_freezable_wq is equivalent to system_wq except that it's * freezable. + * + * *_power_efficient_wq are inclined towards saving power and converted + * into WQ_UNBOUND variants if 'wq_power_efficient' is enabled; otherwise, + * they are same as their non-power-efficient counterparts - e.g. + * system_power_efficient_wq is identical to system_wq if + * 'wq_power_efficient' is disabled. See WQ_POWER_EFFICIENT for more info. */ extern struct workqueue_struct *system_wq; extern struct workqueue_struct *system_long_wq; extern struct workqueue_struct *system_unbound_wq; extern struct workqueue_struct *system_freezable_wq; +extern struct workqueue_struct *system_power_efficient_wq; +extern struct workqueue_struct *system_freezable_power_efficient_wq;

static inline struct workqueue_struct * __deprecated __system_nrt_wq(void) { diff --git a/kernel/workqueue.c b/kernel/workqueue.c index 8068d97..16ca2d3 100644 --- a/kernel/workqueue.c +++ b/kernel/workqueue.c @@ -314,6 +314,10 @@ struct workqueue_struct *system_unbound_wq __read_mostly; EXPORT_SYMBOL_GPL(system_unbound_wq); struct workqueue_struct *system_freezable_wq __read_mostly; EXPORT_SYMBOL_GPL(system_freezable_wq); +struct workqueue_struct *system_power_efficient_wq __read_mostly; +EXPORT_SYMBOL_GPL(system_power_efficient_wq); +struct workqueue_struct *system_freezable_power_efficient_wq __read_mostly; +EXPORT_SYMBOL_GPL(system_freezable_power_efficient_wq);

static int worker_thread(void *__worker); static void copy_workqueue_attrs(struct workqueue_attrs *to, @@ -4987,8 +4991,15 @@ static int __init init_workqueues(void) WQ_UNBOUND_MAX_ACTIVE); system_freezable_wq = alloc_workqueue("events_freezable", WQ_FREEZABLE, 0); + system_power_efficient_wq = alloc_workqueue("events_power_efficient", + WQ_POWER_EFFICIENT, 0); + system_freezable_power_efficient_wq = alloc_workqueue("events_freezable_power_efficient", + WQ_FREEZABLE | WQ_POWER_EFFICIENT, + 0); BUG_ON(!system_wq || !system_highpri_wq || !system_long_wq || - !system_unbound_wq || !system_freezable_wq); + !system_unbound_wq || !system_freezable_wq || + !system_power_efficient_wq || + !system_freezable_power_efficient_wq); return 0; } early_initcall(init_workqueues);