[PATCH v2 0/2] new cpumask for hotpluggable CPUs

[PATCH v2 0/2] new cpumask for hotpluggable CPUs

[Mike Turquette]

This patch series introduces a new cpumask which tracks CPUs that
support hotplugging.  The purpose of this patch series is to provide a
simple method for kernel code to know which CPUs can be hotplugged and
which ones cannot.  Potential users of this code might be a thermal
mitigation technique which uses hotplug to lower temperature, or a power
capping mechanism which uses hotplug to lower power consumption.

All the of usual cpumask helper functions are created for this new mask.
The second patch in this series simply sets the bit for elligible CPUs
while they are being registered.  The cpumask itself is static after
boot and should not change (like the possbile mask).

Mike Turquette (2):
  cpumask: introduce cpumask for hotpluggable CPUs
  cpu: update cpu_hotpluggable_mask in register_cpu

 drivers/base/cpu.c      |    4 +++-
 include/linux/cpumask.h |   27 ++++++++++++++++++++++-----
 kernel/cpu.c            |   18 ++++++++++++++++++
 3 files changed, 43 insertions(+), 6 deletions(-)

-- 
1.7.4.1

[PATCH v2 1/2] cpumask: introduce cpumask for hotpluggable CPUs

[Mike Turquette]

On some platforms it is possible to have some CPUs which support CPU
hotplug and some which do not.  Currently the prescence of an 'online'
sysfs entry in userspace is adequate for applications to know that a CPU
supports hotplug, but there is no convenient way to make the same
determination in the kernel.

To better model this relationship this patch introduces a new cpumask to
track CPUs that support CPU hotplug operations.

This new cpumask is populated at boot-time and remains static for the
life of the machine.  Bits set in the mask indicate a CPU which supports
hotplug, but make no guarantees about whether that CPU is currently
online or not.  Likewise a cleared bit in the mask indicates either a
CPU which cannot hotplug or a lack of a populated CPU.

The purpose of this new cpumask is to aid kernel code which uses CPU to
take CPUs online and offline.  Possible uses are as a thermal event
mitigation technique or as a power capping mechanism.

Signed-off-by: Mike Turquette <mturquette@ti.com>
---
Change log:
v2: fixed missing parentheses in cpumask_test_cpu and improved grammar
    in comments

 include/linux/cpumask.h |   27 ++++++++++++++++++++++-----
 kernel/cpu.c            |   18 ++++++++++++++++++
 2 files changed, 40 insertions(+), 5 deletions(-)

diff --git a/include/linux/cpumask.h b/include/linux/cpumask.h
index b24ac56..52e64a7 100644
--- a/include/linux/cpumask.h
+++ b/include/linux/cpumask.h
@@ -39,10 +39,11 @@ extern int nr_cpu_ids;
  * The following particular system cpumasks and operations manage
  * possible, present, active and online cpus.
  *
- *     cpu_possible_mask- has bit 'cpu' set iff cpu is populatable
- *     cpu_present_mask - has bit 'cpu' set iff cpu is populated
- *     cpu_online_mask  - has bit 'cpu' set iff cpu available to scheduler
- *     cpu_active_mask  - has bit 'cpu' set iff cpu available to migration
+ *     cpu_possible_mask     - has bit 'cpu' set iff cpu is populatable
+ *     cpu_hotpluggable_mask - has bit 'cpu' set iff cpu is hotpluggable
+ *     cpu_present_mask      - has bit 'cpu' set iff cpu is populated
+ *     cpu_online_mask       - has bit 'cpu' set iff cpu available to scheduler
+ *     cpu_active_mask       - has bit 'cpu' set iff cpu available to migration
  *
  *  If !CONFIG_HOTPLUG_CPU, present == possible, and active == online.
  *
@@ -51,7 +52,11 @@ extern int nr_cpu_ids;
  *  life of that system boot.  The cpu_present_mask is dynamic(*),
  *  representing which CPUs are currently plugged in.  And
  *  cpu_online_mask is the dynamic subset of cpu_present_mask,
- *  indicating those CPUs available for scheduling.
+ *  indicating those CPUs available for scheduling.  The
+ *  cpu_hotpluggable_mask is also fixed at boot time as the set of CPU
+ *  id's which are possible AND can hotplug.  Cleared bits in this mask
+ *  mean that either the CPU is not possible, or it is possible but does
+ *  not support CPU hotplug operations.
  *
  *  If HOTPLUG is enabled, then cpu_possible_mask is forced to have
  *  all NR_CPUS bits set, otherwise it is just the set of CPUs that
@@ -61,6 +66,9 @@ extern int nr_cpu_ids;
  *  depending on what ACPI reports as currently plugged in, otherwise
  *  cpu_present_mask is just a copy of cpu_possible_mask.
  *
+ *  If HOTPLUG is not enabled then cpu_hotpluggable_mask is the empty
+ *  set.
+ *
  *  (*) Well, cpu_present_mask is dynamic in the hotplug case.  If not
  *      hotplug, it's a copy of cpu_possible_mask, hence fixed at boot.
  *
@@ -76,6 +84,7 @@ extern int nr_cpu_ids;
  */
 
 extern const struct cpumask *const cpu_possible_mask;
+extern const struct cpumask *const cpu_hotpluggable_mask;
 extern const struct cpumask *const cpu_online_mask;
 extern const struct cpumask *const cpu_present_mask;
 extern const struct cpumask *const cpu_active_mask;
@@ -85,19 +94,23 @@ extern const struct cpumask *const cpu_active_mask;
 #define num_possible_cpus()	cpumask_weight(cpu_possible_mask)
 #define num_present_cpus()	cpumask_weight(cpu_present_mask)
 #define num_active_cpus()	cpumask_weight(cpu_active_mask)
+#define num_hotpluggable_cpus()	cpumask_weight(cpu_hotpluggable_mask)
 #define cpu_online(cpu)		cpumask_test_cpu((cpu), cpu_online_mask)
 #define cpu_possible(cpu)	cpumask_test_cpu((cpu), cpu_possible_mask)
 #define cpu_present(cpu)	cpumask_test_cpu((cpu), cpu_present_mask)
 #define cpu_active(cpu)		cpumask_test_cpu((cpu), cpu_active_mask)
+#define cpu_hotpluggable(cpu)	cpumask_test_cpu((cpu, cpu_hotpluggable_mask))
 #else
 #define num_online_cpus()	1U
 #define num_possible_cpus()	1U
 #define num_present_cpus()	1U
 #define num_active_cpus()	1U
+#define num_hotpluggable_cpus()	0
 #define cpu_online(cpu)		((cpu) == 0)
 #define cpu_possible(cpu)	((cpu) == 0)
 #define cpu_present(cpu)	((cpu) == 0)
 #define cpu_active(cpu)		((cpu) == 0)
+#define cpu_hotpluggable(cpu)	0
 #endif
 
 /* verify cpu argument to cpumask_* operators */
@@ -678,16 +691,20 @@ extern const DECLARE_BITMAP(cpu_all_bits, NR_CPUS);
 #define cpu_none_mask to_cpumask(cpu_bit_bitmap[0])
 
 #define for_each_possible_cpu(cpu) for_each_cpu((cpu), cpu_possible_mask)
+#define for_each_hotpluggable_cpu(cpu) \
+				   for_each_cpu((cpu), cpu_hotpluggable_mask)
 #define for_each_online_cpu(cpu)   for_each_cpu((cpu), cpu_online_mask)
 #define for_each_present_cpu(cpu)  for_each_cpu((cpu), cpu_present_mask)
 
 /* Wrappers for arch boot code to manipulate normally-constant masks */
 void set_cpu_possible(unsigned int cpu, bool possible);
+void set_cpu_hotpluggable(unsigned int cpu, bool hotpluggable);
 void set_cpu_present(unsigned int cpu, bool present);
 void set_cpu_online(unsigned int cpu, bool online);
 void set_cpu_active(unsigned int cpu, bool active);
 void init_cpu_present(const struct cpumask *src);
 void init_cpu_possible(const struct cpumask *src);
+void init_cpu_hotpluggable(const struct cpumask *src);
 void init_cpu_online(const struct cpumask *src);
 
 /**
diff --git a/kernel/cpu.c b/kernel/cpu.c
index 12b7458..8c397c9 100644
--- a/kernel/cpu.c
+++ b/kernel/cpu.c
@@ -536,6 +536,11 @@ static DECLARE_BITMAP(cpu_possible_bits, CONFIG_NR_CPUS) __read_mostly;
 const struct cpumask *const cpu_possible_mask = to_cpumask(cpu_possible_bits);
 EXPORT_SYMBOL(cpu_possible_mask);
 
+static DECLARE_BITMAP(cpu_hotpluggable_bits, CONFIG_NR_CPUS) __read_mostly;
+const struct cpumask *const cpu_hotpluggable_mask =
+					to_cpumask(cpu_hotpluggable_bits);
+EXPORT_SYMBOL(cpu_hotpluggable_mask);
+
 static DECLARE_BITMAP(cpu_online_bits, CONFIG_NR_CPUS) __read_mostly;
 const struct cpumask *const cpu_online_mask = to_cpumask(cpu_online_bits);
 EXPORT_SYMBOL(cpu_online_mask);
@@ -556,6 +561,14 @@ void set_cpu_possible(unsigned int cpu, bool possible)
 		cpumask_clear_cpu(cpu, to_cpumask(cpu_possible_bits));
 }
 
+void set_cpu_hotpluggable(unsigned int cpu, bool hotpluggable)
+{
+	if (hotpluggable)
+		cpumask_set_cpu(cpu, to_cpumask(cpu_hotpluggable_bits));
+	else
+		cpumask_clear_cpu(cpu, to_cpumask(cpu_hotpluggable_bits));
+}
+
 void set_cpu_present(unsigned int cpu, bool present)
 {
 	if (present)
@@ -590,6 +603,11 @@ void init_cpu_possible(const struct cpumask *src)
 	cpumask_copy(to_cpumask(cpu_possible_bits), src);
 }
 
+void init_cpu_hotpluggable(const struct cpumask *src)
+{
+	cpumask_copy(to_cpumask(cpu_hotpluggable_bits), src);
+}
+
 void init_cpu_online(const struct cpumask *src)
 {
 	cpumask_copy(to_cpumask(cpu_online_bits), src);
-- 
1.7.4.1

Re: [PATCH v2 1/2] cpumask: introduce cpumask for hotpluggable CPUs

[Daniel Lezcano]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 08/10/2011 10:03 PM, Mike Turquette wrote:
> On some platforms it is possible to have some CPUs which support CPU
> hotplug and some which do not.  Currently the prescence of an 'online'
> sysfs entry in userspace is adequate for applications to know that a CPU
> supports hotplug, but there is no convenient way to make the same
> determination in the kernel.
> 
> To better model this relationship this patch introduces a new cpumask to
> track CPUs that support CPU hotplug operations.
> 
> This new cpumask is populated at boot-time and remains static for the
> life of the machine.  Bits set in the mask indicate a CPU which supports
> hotplug, but make no guarantees about whether that CPU is currently
> online or not.  Likewise a cleared bit in the mask indicates either a
> CPU which cannot hotplug or a lack of a populated CPU.
> 
> The purpose of this new cpumask is to aid kernel code which uses CPU to
> take CPUs online and offline.  Possible uses are as a thermal event
> mitigation technique or as a power capping mechanism.
> 
> Signed-off-by: Mike Turquette <mturquette@ti.com>

Reviewed-by: Daniel Lezcano <daniel.lezcano@linaro.org>

- -- 
 <http://www.linaro.org/> Linaro.org ? Open source software for ARM SoCs

Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iQEcBAEBAgAGBQJOQu/KAAoJEAKBbMCpUGYAxnoH/0+WtrRGq0oOjkU8jIxzAEUP
OW6S7EQT13v97YapaEwUZpHOzrdeEa9bV5sVI3uc3H9n6dnQbh25eMuhiH/dTwF6
LHV1Tdv7/+ghiSc4NJunpXjAsObezUgTV9n2Zoip3kAfhMdV5DntJu0Izkdb5lkA
AD5CqV6oHqC0spa2/nMMeOYsdp3hLP3hA5wCyE8a2rxvbytxl2jRVommXOpF3AnI
s3OS7oyVFLpdvxVdAio1cXvM6vILtJFArAqw88AJUBsUsuehRxeOYVvTR3/z4RZD
LNEH72IkGY/8OVYp9/VoixWgWTUhXUR5BZwAjCP7a+xQk/1qK1LAM+/CXimwND0=
=90gJ
-----END PGP SIGNATURE-----

Re: [PATCH v2 1/2] cpumask: introduce cpumask for hotpluggable CPUs

[Amit Kucheria]

See comments inline.

On 11 Aug 10, Mike Turquette wrote:
> On some platforms it is possible to have some CPUs which support CPU
> hotplug and some which do not.  Currently the prescence of an 'online'
> sysfs entry in userspace is adequate for applications to know that a CPU
> supports hotplug, but there is no convenient way to make the same
> determination in the kernel.
> 
> To better model this relationship this patch introduces a new cpumask to
> track CPUs that support CPU hotplug operations.
> 
> This new cpumask is populated at boot-time and remains static for the
> life of the machine.  Bits set in the mask indicate a CPU which supports
> hotplug, but make no guarantees about whether that CPU is currently
> online or not.  Likewise a cleared bit in the mask indicates either a
> CPU which cannot hotplug or a lack of a populated CPU.
> 
> The purpose of this new cpumask is to aid kernel code which uses CPU to
> take CPUs online and offline.  Possible uses are as a thermal event
> mitigation technique or as a power capping mechanism.
> 
> Signed-off-by: Mike Turquette <mturquette@ti.com>
> ---
> Change log:
> v2: fixed missing parentheses in cpumask_test_cpu and improved grammar
>     in comments
> 
>  include/linux/cpumask.h |   27 ++++++++++++++++++++++-----
>  kernel/cpu.c            |   18 ++++++++++++++++++
>  2 files changed, 40 insertions(+), 5 deletions(-)
> 
> diff --git a/include/linux/cpumask.h b/include/linux/cpumask.h
> index b24ac56..52e64a7 100644
> --- a/include/linux/cpumask.h
> +++ b/include/linux/cpumask.h
> @@ -39,10 +39,11 @@ extern int nr_cpu_ids;
>   * The following particular system cpumasks and operations manage
>   * possible, present, active and online cpus.
>   *
> - *     cpu_possible_mask- has bit 'cpu' set iff cpu is populatable
> - *     cpu_present_mask - has bit 'cpu' set iff cpu is populated
> - *     cpu_online_mask  - has bit 'cpu' set iff cpu available to scheduler
> - *     cpu_active_mask  - has bit 'cpu' set iff cpu available to migration
> + *     cpu_possible_mask     - has bit 'cpu' set iff cpu is populatable
> + *     cpu_hotpluggable_mask - has bit 'cpu' set iff cpu is hotpluggable
> + *     cpu_present_mask      - has bit 'cpu' set iff cpu is populated
> + *     cpu_online_mask       - has bit 'cpu' set iff cpu available to scheduler
> + *     cpu_active_mask       - has bit 'cpu' set iff cpu available to migration
>   *
>   *  If !CONFIG_HOTPLUG_CPU, present == possible, and active == online.
>   *
> @@ -51,7 +52,11 @@ extern int nr_cpu_ids;
>   *  life of that system boot.  The cpu_present_mask is dynamic(*),
>   *  representing which CPUs are currently plugged in.  And
>   *  cpu_online_mask is the dynamic subset of cpu_present_mask,
> - *  indicating those CPUs available for scheduling.
> + *  indicating those CPUs available for scheduling.  The
> + *  cpu_hotpluggable_mask is also fixed at boot time as the set of CPU
> + *  id's which are possible AND can hotplug.  Cleared bits in this mask
> + *  mean that either the CPU is not possible, or it is possible but does
> + *  not support CPU hotplug operations.
>   *
>   *  If HOTPLUG is enabled, then cpu_possible_mask is forced to have
>   *  all NR_CPUS bits set, otherwise it is just the set of CPUs that
> @@ -61,6 +66,9 @@ extern int nr_cpu_ids;
>   *  depending on what ACPI reports as currently plugged in, otherwise
>   *  cpu_present_mask is just a copy of cpu_possible_mask.
>   *
> + *  If HOTPLUG is not enabled then cpu_hotpluggable_mask is the empty
> + *  set.
> + *
>   *  (*) Well, cpu_present_mask is dynamic in the hotplug case.  If not
>   *      hotplug, it's a copy of cpu_possible_mask, hence fixed at boot.
>   *
> @@ -76,6 +84,7 @@ extern int nr_cpu_ids;
>   */
>  
>  extern const struct cpumask *const cpu_possible_mask;
> +extern const struct cpumask *const cpu_hotpluggable_mask;
>  extern const struct cpumask *const cpu_online_mask;
>  extern const struct cpumask *const cpu_present_mask;
>  extern const struct cpumask *const cpu_active_mask;
> @@ -85,19 +94,23 @@ extern const struct cpumask *const cpu_active_mask;
>  #define num_possible_cpus()	cpumask_weight(cpu_possible_mask)
>  #define num_present_cpus()	cpumask_weight(cpu_present_mask)
>  #define num_active_cpus()	cpumask_weight(cpu_active_mask)
> +#define num_hotpluggable_cpus()	cpumask_weight(cpu_hotpluggable_mask)
>  #define cpu_online(cpu)		cpumask_test_cpu((cpu), cpu_online_mask)
>  #define cpu_possible(cpu)	cpumask_test_cpu((cpu), cpu_possible_mask)
>  #define cpu_present(cpu)	cpumask_test_cpu((cpu), cpu_present_mask)
>  #define cpu_active(cpu)		cpumask_test_cpu((cpu), cpu_active_mask)
> +#define cpu_hotpluggable(cpu)	cpumask_test_cpu((cpu, cpu_hotpluggable_mask))

The bracket should be around cpu, like this (cpu) so that there are no
side-effects when passing anything to the macro. See the other #defines above.

>  #else
>  #define num_online_cpus()	1U
>  #define num_possible_cpus()	1U
>  #define num_present_cpus()	1U
>  #define num_active_cpus()	1U
> +#define num_hotpluggable_cpus()	0
>  #define cpu_online(cpu)		((cpu) == 0)
>  #define cpu_possible(cpu)	((cpu) == 0)
>  #define cpu_present(cpu)	((cpu) == 0)
>  #define cpu_active(cpu)		((cpu) == 0)
> +#define cpu_hotpluggable(cpu)	0
>  #endif
>  
>  /* verify cpu argument to cpumask_* operators */
> @@ -678,16 +691,20 @@ extern const DECLARE_BITMAP(cpu_all_bits, NR_CPUS);
>  #define cpu_none_mask to_cpumask(cpu_bit_bitmap[0])
>  
>  #define for_each_possible_cpu(cpu) for_each_cpu((cpu), cpu_possible_mask)
> +#define for_each_hotpluggable_cpu(cpu) \
> +				   for_each_cpu((cpu), cpu_hotpluggable_mask)
>  #define for_each_online_cpu(cpu)   for_each_cpu((cpu), cpu_online_mask)
>  #define for_each_present_cpu(cpu)  for_each_cpu((cpu), cpu_present_mask)
>  
>  /* Wrappers for arch boot code to manipulate normally-constant masks */
>  void set_cpu_possible(unsigned int cpu, bool possible);
> +void set_cpu_hotpluggable(unsigned int cpu, bool hotpluggable);
>  void set_cpu_present(unsigned int cpu, bool present);
>  void set_cpu_online(unsigned int cpu, bool online);
>  void set_cpu_active(unsigned int cpu, bool active);
>  void init_cpu_present(const struct cpumask *src);
>  void init_cpu_possible(const struct cpumask *src);
> +void init_cpu_hotpluggable(const struct cpumask *src);
>  void init_cpu_online(const struct cpumask *src);
>  
>  /**
> diff --git a/kernel/cpu.c b/kernel/cpu.c
> index 12b7458..8c397c9 100644
> --- a/kernel/cpu.c
> +++ b/kernel/cpu.c
> @@ -536,6 +536,11 @@ static DECLARE_BITMAP(cpu_possible_bits, CONFIG_NR_CPUS) __read_mostly;
>  const struct cpumask *const cpu_possible_mask = to_cpumask(cpu_possible_bits);
>  EXPORT_SYMBOL(cpu_possible_mask);
>  
> +static DECLARE_BITMAP(cpu_hotpluggable_bits, CONFIG_NR_CPUS) __read_mostly;
> +const struct cpumask *const cpu_hotpluggable_mask =
> +					to_cpumask(cpu_hotpluggable_bits);
> +EXPORT_SYMBOL(cpu_hotpluggable_mask);
> +
>  static DECLARE_BITMAP(cpu_online_bits, CONFIG_NR_CPUS) __read_mostly;
>  const struct cpumask *const cpu_online_mask = to_cpumask(cpu_online_bits);
>  EXPORT_SYMBOL(cpu_online_mask);
> @@ -556,6 +561,14 @@ void set_cpu_possible(unsigned int cpu, bool possible)
>  		cpumask_clear_cpu(cpu, to_cpumask(cpu_possible_bits));
>  }
>  
> +void set_cpu_hotpluggable(unsigned int cpu, bool hotpluggable)
> +{
> +	if (hotpluggable)
> +		cpumask_set_cpu(cpu, to_cpumask(cpu_hotpluggable_bits));
> +	else
> +		cpumask_clear_cpu(cpu, to_cpumask(cpu_hotpluggable_bits));
> +}
> +
>  void set_cpu_present(unsigned int cpu, bool present)
>  {
>  	if (present)
> @@ -590,6 +603,11 @@ void init_cpu_possible(const struct cpumask *src)
>  	cpumask_copy(to_cpumask(cpu_possible_bits), src);
>  }
>  
> +void init_cpu_hotpluggable(const struct cpumask *src)
> +{
> +	cpumask_copy(to_cpumask(cpu_hotpluggable_bits), src);
> +}
> +

Just to be clear I understand this code, init_cpu_hotpluggable will be called
in arch code, right? e.g. arch/arm/kernel/smp.c or do you intend for it to be
called in platform-specific code?

>  void init_cpu_online(const struct cpumask *src)
>  {
>  	cpumask_copy(to_cpumask(cpu_online_bits), src);
> -- 
> 1.7.4.1

Re: [PATCH v2 1/2] cpumask: introduce cpumask for hotpluggable CPUs

[Turquette, Mike]

On Wed, Aug 10, 2011 at 11:06 PM, Amit Kucheria
<amit.kucheria@linaro.org> wrote:
> See comments inline.
>
> On 11 Aug 10, Mike Turquette wrote:
>> On some platforms it is possible to have some CPUs which support CPU
>> hotplug and some which do not.  Currently the prescence of an 'online'
>> sysfs entry in userspace is adequate for applications to know that a CPU
>> supports hotplug, but there is no convenient way to make the same
>> determination in the kernel.
>>
>> To better model this relationship this patch introduces a new cpumask to
>> track CPUs that support CPU hotplug operations.
>>
>> This new cpumask is populated at boot-time and remains static for the
>> life of the machine.  Bits set in the mask indicate a CPU which supports
>> hotplug, but make no guarantees about whether that CPU is currently
>> online or not.  Likewise a cleared bit in the mask indicates either a
>> CPU which cannot hotplug or a lack of a populated CPU.
>>
>> The purpose of this new cpumask is to aid kernel code which uses CPU to
>> take CPUs online and offline.  Possible uses are as a thermal event
>> mitigation technique or as a power capping mechanism.
>>
>> Signed-off-by: Mike Turquette <mturquette@ti.com>
>> ---
>> Change log:
>> v2: fixed missing parentheses in cpumask_test_cpu and improved grammar
>>     in comments
>>
>>  include/linux/cpumask.h |   27 ++++++++++++++++++++++-----
>>  kernel/cpu.c            |   18 ++++++++++++++++++
>>  2 files changed, 40 insertions(+), 5 deletions(-)
>>
>> diff --git a/include/linux/cpumask.h b/include/linux/cpumask.h
>> index b24ac56..52e64a7 100644
>> --- a/include/linux/cpumask.h
>> +++ b/include/linux/cpumask.h
>> @@ -39,10 +39,11 @@ extern int nr_cpu_ids;
>>   * The following particular system cpumasks and operations manage
>>   * possible, present, active and online cpus.
>>   *
>> - *     cpu_possible_mask- has bit 'cpu' set iff cpu is populatable
>> - *     cpu_present_mask - has bit 'cpu' set iff cpu is populated
>> - *     cpu_online_mask  - has bit 'cpu' set iff cpu available to scheduler
>> - *     cpu_active_mask  - has bit 'cpu' set iff cpu available to migration
>> + *     cpu_possible_mask     - has bit 'cpu' set iff cpu is populatable
>> + *     cpu_hotpluggable_mask - has bit 'cpu' set iff cpu is hotpluggable
>> + *     cpu_present_mask      - has bit 'cpu' set iff cpu is populated
>> + *     cpu_online_mask       - has bit 'cpu' set iff cpu available to scheduler
>> + *     cpu_active_mask       - has bit 'cpu' set iff cpu available to migration
>>   *
>>   *  If !CONFIG_HOTPLUG_CPU, present == possible, and active == online.
>>   *
>> @@ -51,7 +52,11 @@ extern int nr_cpu_ids;
>>   *  life of that system boot.  The cpu_present_mask is dynamic(*),
>>   *  representing which CPUs are currently plugged in.  And
>>   *  cpu_online_mask is the dynamic subset of cpu_present_mask,
>> - *  indicating those CPUs available for scheduling.
>> + *  indicating those CPUs available for scheduling.  The
>> + *  cpu_hotpluggable_mask is also fixed at boot time as the set of CPU
>> + *  id's which are possible AND can hotplug.  Cleared bits in this mask
>> + *  mean that either the CPU is not possible, or it is possible but does
>> + *  not support CPU hotplug operations.
>>   *
>>   *  If HOTPLUG is enabled, then cpu_possible_mask is forced to have
>>   *  all NR_CPUS bits set, otherwise it is just the set of CPUs that
>> @@ -61,6 +66,9 @@ extern int nr_cpu_ids;
>>   *  depending on what ACPI reports as currently plugged in, otherwise
>>   *  cpu_present_mask is just a copy of cpu_possible_mask.
>>   *
>> + *  If HOTPLUG is not enabled then cpu_hotpluggable_mask is the empty
>> + *  set.
>> + *
>>   *  (*) Well, cpu_present_mask is dynamic in the hotplug case.  If not
>>   *      hotplug, it's a copy of cpu_possible_mask, hence fixed at boot.
>>   *
>> @@ -76,6 +84,7 @@ extern int nr_cpu_ids;
>>   */
>>
>>  extern const struct cpumask *const cpu_possible_mask;
>> +extern const struct cpumask *const cpu_hotpluggable_mask;
>>  extern const struct cpumask *const cpu_online_mask;
>>  extern const struct cpumask *const cpu_present_mask;
>>  extern const struct cpumask *const cpu_active_mask;
>> @@ -85,19 +94,23 @@ extern const struct cpumask *const cpu_active_mask;
>>  #define num_possible_cpus()  cpumask_weight(cpu_possible_mask)
>>  #define num_present_cpus()   cpumask_weight(cpu_present_mask)
>>  #define num_active_cpus()    cpumask_weight(cpu_active_mask)
>> +#define num_hotpluggable_cpus()      cpumask_weight(cpu_hotpluggable_mask)
>>  #define cpu_online(cpu)              cpumask_test_cpu((cpu), cpu_online_mask)
>>  #define cpu_possible(cpu)    cpumask_test_cpu((cpu), cpu_possible_mask)
>>  #define cpu_present(cpu)     cpumask_test_cpu((cpu), cpu_present_mask)
>>  #define cpu_active(cpu)              cpumask_test_cpu((cpu), cpu_active_mask)
>> +#define cpu_hotpluggable(cpu)        cpumask_test_cpu((cpu, cpu_hotpluggable_mask))
>
> The bracket should be around cpu, like this (cpu) so that there are no
> side-effects when passing anything to the macro. See the other #defines above.

Wow, not sure how that happened...  will send in V3...

>>  #else
>>  #define num_online_cpus()    1U
>>  #define num_possible_cpus()  1U
>>  #define num_present_cpus()   1U
>>  #define num_active_cpus()    1U
>> +#define num_hotpluggable_cpus()      0
>>  #define cpu_online(cpu)              ((cpu) == 0)
>>  #define cpu_possible(cpu)    ((cpu) == 0)
>>  #define cpu_present(cpu)     ((cpu) == 0)
>>  #define cpu_active(cpu)              ((cpu) == 0)
>> +#define cpu_hotpluggable(cpu)        0
>>  #endif
>>
>>  /* verify cpu argument to cpumask_* operators */
>> @@ -678,16 +691,20 @@ extern const DECLARE_BITMAP(cpu_all_bits, NR_CPUS);
>>  #define cpu_none_mask to_cpumask(cpu_bit_bitmap[0])
>>
>>  #define for_each_possible_cpu(cpu) for_each_cpu((cpu), cpu_possible_mask)
>> +#define for_each_hotpluggable_cpu(cpu) \
>> +                                for_each_cpu((cpu), cpu_hotpluggable_mask)
>>  #define for_each_online_cpu(cpu)   for_each_cpu((cpu), cpu_online_mask)
>>  #define for_each_present_cpu(cpu)  for_each_cpu((cpu), cpu_present_mask)
>>
>>  /* Wrappers for arch boot code to manipulate normally-constant masks */
>>  void set_cpu_possible(unsigned int cpu, bool possible);
>> +void set_cpu_hotpluggable(unsigned int cpu, bool hotpluggable);
>>  void set_cpu_present(unsigned int cpu, bool present);
>>  void set_cpu_online(unsigned int cpu, bool online);
>>  void set_cpu_active(unsigned int cpu, bool active);
>>  void init_cpu_present(const struct cpumask *src);
>>  void init_cpu_possible(const struct cpumask *src);
>> +void init_cpu_hotpluggable(const struct cpumask *src);
>>  void init_cpu_online(const struct cpumask *src);
>>
>>  /**
>> diff --git a/kernel/cpu.c b/kernel/cpu.c
>> index 12b7458..8c397c9 100644
>> --- a/kernel/cpu.c
>> +++ b/kernel/cpu.c
>> @@ -536,6 +536,11 @@ static DECLARE_BITMAP(cpu_possible_bits, CONFIG_NR_CPUS) __read_mostly;
>>  const struct cpumask *const cpu_possible_mask = to_cpumask(cpu_possible_bits);
>>  EXPORT_SYMBOL(cpu_possible_mask);
>>
>> +static DECLARE_BITMAP(cpu_hotpluggable_bits, CONFIG_NR_CPUS) __read_mostly;
>> +const struct cpumask *const cpu_hotpluggable_mask =
>> +                                     to_cpumask(cpu_hotpluggable_bits);
>> +EXPORT_SYMBOL(cpu_hotpluggable_mask);
>> +
>>  static DECLARE_BITMAP(cpu_online_bits, CONFIG_NR_CPUS) __read_mostly;
>>  const struct cpumask *const cpu_online_mask = to_cpumask(cpu_online_bits);
>>  EXPORT_SYMBOL(cpu_online_mask);
>> @@ -556,6 +561,14 @@ void set_cpu_possible(unsigned int cpu, bool possible)
>>               cpumask_clear_cpu(cpu, to_cpumask(cpu_possible_bits));
>>  }
>>
>> +void set_cpu_hotpluggable(unsigned int cpu, bool hotpluggable)
>> +{
>> +     if (hotpluggable)
>> +             cpumask_set_cpu(cpu, to_cpumask(cpu_hotpluggable_bits));
>> +     else
>> +             cpumask_clear_cpu(cpu, to_cpumask(cpu_hotpluggable_bits));
>> +}
>> +
>>  void set_cpu_present(unsigned int cpu, bool present)
>>  {
>>       if (present)
>> @@ -590,6 +603,11 @@ void init_cpu_possible(const struct cpumask *src)
>>       cpumask_copy(to_cpumask(cpu_possible_bits), src);
>>  }
>>
>> +void init_cpu_hotpluggable(const struct cpumask *src)
>> +{
>> +     cpumask_copy(to_cpumask(cpu_hotpluggable_bits), src);
>> +}
>> +
>
> Just to be clear I understand this code, init_cpu_hotpluggable will be called
> in arch code, right? e.g. arch/arm/kernel/smp.c or do you intend for it to be
> called in platform-specific code?

It could be called in either, though I assume arch code is a better
place to call this.

This call is made redundant by patch 2 which populates the
hotpluggable bits in register_cpu, but some platforms might still need
it so it is included for completeness sake.  x86 and ARM will not call
this, for example.  In fact I could not find any users of the
init_cpu_xxx cpumask helpers on any architectures.

Thanks,
Mike

>>  void init_cpu_online(const struct cpumask *src)
>>  {
>>       cpumask_copy(to_cpumask(cpu_online_bits), src);
>> --
>> 1.7.4.1
>

[PATCH v2 2/2] cpu: update cpu_hotpluggable_mask in register_cpu

[Mike Turquette]

Update the cpu_hotpluggable_mask for each registered CPU which supports
hotplug.  This makes it trivial for kernel code to know which CPUs
support hotplug operations.

Signed-off-by: Mike Turquette <mturquette@ti.com>
---
Change log:
v2: no change

 drivers/base/cpu.c |    4 +++-
 1 files changed, 3 insertions(+), 1 deletions(-)

diff --git a/drivers/base/cpu.c b/drivers/base/cpu.c
index 251acea..91ddcf8 100644
--- a/drivers/base/cpu.c
+++ b/drivers/base/cpu.c
@@ -224,8 +224,10 @@ int __cpuinit register_cpu(struct cpu *cpu, int num)
 
 	error = sysdev_register(&cpu->sysdev);
 
-	if (!error && cpu->hotpluggable)
+	if (!error && cpu->hotpluggable) {
 		register_cpu_control(cpu);
+		set_cpu_hotpluggable(num, true);
+	}
 	if (!error)
 		per_cpu(cpu_sys_devices, num) = &cpu->sysdev;
 	if (!error)
-- 
1.7.4.1

Re: [PATCH v2 2/2] cpu: update cpu_hotpluggable_mask in register_cpu

[Daniel Lezcano]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

On 08/10/2011 10:03 PM, Mike Turquette wrote:
> Update the cpu_hotpluggable_mask for each registered CPU which supports
> hotplug.  This makes it trivial for kernel code to know which CPUs
> support hotplug operations.
> 
> Signed-off-by: Mike Turquette <mturquette@ti.com>

Reviewed-by: <Daniel Lezcano> daniel.lezcano@linaro.org

- -- 
 <http://www.linaro.org/> Linaro.org ? Open source software for ARM SoCs

Follow Linaro:  <http://www.facebook.com/pages/Linaro> Facebook |
<http://twitter.com/#!/linaroorg> Twitter |
<http://www.linaro.org/linaro-blog/> Blog

-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org/

iQEcBAEBAgAGBQJOQvAUAAoJEAKBbMCpUGYADTYH/3jGeK5pryiVKaxQ7wUBwsG6
7ugSLCCgBVUL6ueuVmlphhfW8Vv9Uz72LlJnaJUmJ0deK/Rf5x1ccFulBTtIsyjK
Zbw4Wg0MYxfNo0OAi7OSjWMUw31PC8kKVBSUZd3zHeqi6TanKUtxR5im4EWGhQbc
+Njhx2jAqZBEEcPnODiRx+CbYJX17N2Cesmk8Z+Tk00LIpM30iGGX15ZWXnV6KWM
zD3vG7ATLa/rv1/xPa67SDXdZIo99wDodjtjTFPClygM/XTSNBiIxsm9LnnUkmyw
gw1Q6Ng+wo2XyhxJ//b+kM0natnc7Tl6RtDIL7T+j9rIAdf8Sz9DPgmafWQRxYY=
=hwkF
-----END PGP SIGNATURE-----

Re: [PATCH v2 0/2] new cpumask for hotpluggable CPUs

[Peter Zijlstra]

On Wed, 2011-08-10 at 13:03 -0700, Mike Turquette wrote:
> This patch series introduces a new cpumask which tracks CPUs that
> support hotplugging.  The purpose of this patch series is to provide a
> simple method for kernel code to know which CPUs can be hotplugged and
> which ones cannot.  Potential users of this code might be a thermal
> mitigation technique which uses hotplug to lower temperature, or a power
> capping mechanism which uses hotplug to lower power consumption.
> 
> All the of usual cpumask helper functions are created for this new mask.
> The second patch in this series simply sets the bit for elligible CPUs
> while they are being registered.  The cpumask itself is static after
> boot and should not change (like the possbile mask).

I still most strongly object to people using hotplug for these goals.

Why do you need to go through the entire dance of hotplug just to idle a
cpu? Hotplug not only idles the cpu but tears down (and rebuilds) an
insane amount of resources associated with the cpu.

Nacked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>

Re: [PATCH v2 0/2] new cpumask for hotpluggable CPUs

[Turquette, Mike]

On Thu, Aug 11, 2011 at 11:30 AM, Peter Zijlstra <peterz@infradead.org> wrote:
> On Wed, 2011-08-10 at 13:03 -0700, Mike Turquette wrote:
>> This patch series introduces a new cpumask which tracks CPUs that
>> support hotplugging.  The purpose of this patch series is to provide a
>> simple method for kernel code to know which CPUs can be hotplugged and
>> which ones cannot.  Potential users of this code might be a thermal
>> mitigation technique which uses hotplug to lower temperature, or a power
>> capping mechanism which uses hotplug to lower power consumption.
>>
>> All the of usual cpumask helper functions are created for this new mask.
>> The second patch in this series simply sets the bit for elligible CPUs
>> while they are being registered.  The cpumask itself is static after
>> boot and should not change (like the possbile mask).
>
> I still most strongly object to people using hotplug for these goals.
>
> Why do you need to go through the entire dance of hotplug just to idle a
> cpu? Hotplug not only idles the cpu but tears down (and rebuilds) an
> insane amount of resources associated with the cpu.

I think you're nacking the wrong series.  This patchset simply allows
kernel space to know which CPUs can go offline and which one can't,
which seems pretty innocuous.  Are you fundamentally opposed to the
kernel having better accessor functions to this data?

I'll soon be posting some code which does implement hotplug as a
power-capping feature.  I think *that* is the patch that you'll want
to nack.

Thanks for reviewing,
Mike

> Nacked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>
>

Re: [PATCH v2 0/2] new cpumask for hotpluggable CPUs

[Peter Zijlstra]

On Thu, 2011-08-11 at 12:25 -0700, Turquette, Mike wrote:
> On Thu, Aug 11, 2011 at 11:30 AM, Peter Zijlstra <peterz@infradead.org> wrote:
> > On Wed, 2011-08-10 at 13:03 -0700, Mike Turquette wrote:
> >> This patch series introduces a new cpumask which tracks CPUs that
> >> support hotplugging.  The purpose of this patch series is to provide a
> >> simple method for kernel code to know which CPUs can be hotplugged and
> >> which ones cannot.  Potential users of this code might be a thermal
> >> mitigation technique which uses hotplug to lower temperature, or a power
> >> capping mechanism which uses hotplug to lower power consumption.
> >>
> >> All the of usual cpumask helper functions are created for this new mask.
> >> The second patch in this series simply sets the bit for elligible CPUs
> >> while they are being registered.  The cpumask itself is static after
> >> boot and should not change (like the possbile mask).
> >
> > I still most strongly object to people using hotplug for these goals.
> >
> > Why do you need to go through the entire dance of hotplug just to idle a
> > cpu? Hotplug not only idles the cpu but tears down (and rebuilds) an
> > insane amount of resources associated with the cpu.
> 
> I think you're nacking the wrong series.  This patchset simply allows
> kernel space to know which CPUs can go offline and which one can't,
> which seems pretty innocuous.  Are you fundamentally opposed to the
> kernel having better accessor functions to this data?

Yeah, people might think its sane to use it..

> I'll soon be posting some code which does implement hotplug as a
> power-capping feature.  I think *that* is the patch that you'll want
> to nack.

That too of course..

[PATCH v3 1/2] cpumask: introduce cpumask for hotpluggable CPUs

[Mike Turquette]

On some platforms it is possible to have some CPUs which support CPU
hotplug and some which do not.  Currently the prescence of an 'online'
sysfs entry in userspace is adequate for applications to know that a CPU
supports hotplug, but there is no convenient way to make the same
determination in the kernel.

To better model this relationship this patch introduces a new cpumask to
track CPUs that support CPU hotplug operations.

This new cpumask is populated at boot-time and remains static for the
life of the machine.  Bits set in the mask indicate a CPU which supports
hotplug, but make no guarantees about whether that CPU is currently
online or not.  Likewise a cleared bit in the mask indicates either a
CPU which cannot hotplug or a lack of a populated CPU.

The purpose of this new cpumask is to aid kernel code which uses CPU to
take CPUs online and offline.  Possible uses are as a thermal event
mitigation technique or as a power capping mechanism.

Signed-off-by: Mike Turquette <mturquette@ti.com>
---
Change log:
v2: fixed missing parentheses in cpumask_test_cpu and improved grammar
    in comments
v3: really fixed the parentheses issue from v2

 include/linux/cpumask.h |   27 ++++++++++++++++++++++-----
 kernel/cpu.c            |   18 ++++++++++++++++++
 2 files changed, 40 insertions(+), 5 deletions(-)

diff --git a/include/linux/cpumask.h b/include/linux/cpumask.h
index b24ac56..9584807 100644
--- a/include/linux/cpumask.h
+++ b/include/linux/cpumask.h
@@ -39,10 +39,11 @@ extern int nr_cpu_ids;
  * The following particular system cpumasks and operations manage
  * possible, present, active and online cpus.
  *
- *     cpu_possible_mask- has bit 'cpu' set iff cpu is populatable
- *     cpu_present_mask - has bit 'cpu' set iff cpu is populated
- *     cpu_online_mask  - has bit 'cpu' set iff cpu available to scheduler
- *     cpu_active_mask  - has bit 'cpu' set iff cpu available to migration
+ *     cpu_possible_mask     - has bit 'cpu' set iff cpu is populatable
+ *     cpu_hotpluggable_mask - has bit 'cpu' set iff cpu is hotpluggable
+ *     cpu_present_mask      - has bit 'cpu' set iff cpu is populated
+ *     cpu_online_mask       - has bit 'cpu' set iff cpu available to scheduler
+ *     cpu_active_mask       - has bit 'cpu' set iff cpu available to migration
  *
  *  If !CONFIG_HOTPLUG_CPU, present == possible, and active == online.
  *
@@ -51,7 +52,11 @@ extern int nr_cpu_ids;
  *  life of that system boot.  The cpu_present_mask is dynamic(*),
  *  representing which CPUs are currently plugged in.  And
  *  cpu_online_mask is the dynamic subset of cpu_present_mask,
- *  indicating those CPUs available for scheduling.
+ *  indicating those CPUs available for scheduling.  The
+ *  cpu_hotpluggable_mask is also fixed at boot time as the set of CPU
+ *  id's which are possible AND can hotplug.  Cleared bits in this mask
+ *  mean that either the CPU is not possible, or it is possible but does
+ *  not support CPU hotplug operations.
  *
  *  If HOTPLUG is enabled, then cpu_possible_mask is forced to have
  *  all NR_CPUS bits set, otherwise it is just the set of CPUs that
@@ -61,6 +66,9 @@ extern int nr_cpu_ids;
  *  depending on what ACPI reports as currently plugged in, otherwise
  *  cpu_present_mask is just a copy of cpu_possible_mask.
  *
+ *  If HOTPLUG is not enabled then cpu_hotpluggable_mask is the empty
+ *  set.
+ *
  *  (*) Well, cpu_present_mask is dynamic in the hotplug case.  If not
  *      hotplug, it's a copy of cpu_possible_mask, hence fixed at boot.
  *
@@ -76,6 +84,7 @@ extern int nr_cpu_ids;
  */
 
 extern const struct cpumask *const cpu_possible_mask;
+extern const struct cpumask *const cpu_hotpluggable_mask;
 extern const struct cpumask *const cpu_online_mask;
 extern const struct cpumask *const cpu_present_mask;
 extern const struct cpumask *const cpu_active_mask;
@@ -85,19 +94,23 @@ extern const struct cpumask *const cpu_active_mask;
 #define num_possible_cpus()	cpumask_weight(cpu_possible_mask)
 #define num_present_cpus()	cpumask_weight(cpu_present_mask)
 #define num_active_cpus()	cpumask_weight(cpu_active_mask)
+#define num_hotpluggable_cpus()	cpumask_weight(cpu_hotpluggable_mask)
 #define cpu_online(cpu)		cpumask_test_cpu((cpu), cpu_online_mask)
 #define cpu_possible(cpu)	cpumask_test_cpu((cpu), cpu_possible_mask)
 #define cpu_present(cpu)	cpumask_test_cpu((cpu), cpu_present_mask)
 #define cpu_active(cpu)		cpumask_test_cpu((cpu), cpu_active_mask)
+#define cpu_hotpluggable(cpu)	cpumask_test_cpu((cpu), cpu_hotpluggable_mask)
 #else
 #define num_online_cpus()	1U
 #define num_possible_cpus()	1U
 #define num_present_cpus()	1U
 #define num_active_cpus()	1U
+#define num_hotpluggable_cpus()	0
 #define cpu_online(cpu)		((cpu) == 0)
 #define cpu_possible(cpu)	((cpu) == 0)
 #define cpu_present(cpu)	((cpu) == 0)
 #define cpu_active(cpu)		((cpu) == 0)
+#define cpu_hotpluggable(cpu)	0
 #endif
 
 /* verify cpu argument to cpumask_* operators */
@@ -678,16 +691,20 @@ extern const DECLARE_BITMAP(cpu_all_bits, NR_CPUS);
 #define cpu_none_mask to_cpumask(cpu_bit_bitmap[0])
 
 #define for_each_possible_cpu(cpu) for_each_cpu((cpu), cpu_possible_mask)
+#define for_each_hotpluggable_cpu(cpu) \
+				   for_each_cpu((cpu), cpu_hotpluggable_mask)
 #define for_each_online_cpu(cpu)   for_each_cpu((cpu), cpu_online_mask)
 #define for_each_present_cpu(cpu)  for_each_cpu((cpu), cpu_present_mask)
 
 /* Wrappers for arch boot code to manipulate normally-constant masks */
 void set_cpu_possible(unsigned int cpu, bool possible);
+void set_cpu_hotpluggable(unsigned int cpu, bool hotpluggable);
 void set_cpu_present(unsigned int cpu, bool present);
 void set_cpu_online(unsigned int cpu, bool online);
 void set_cpu_active(unsigned int cpu, bool active);
 void init_cpu_present(const struct cpumask *src);
 void init_cpu_possible(const struct cpumask *src);
+void init_cpu_hotpluggable(const struct cpumask *src);
 void init_cpu_online(const struct cpumask *src);
 
 /**
diff --git a/kernel/cpu.c b/kernel/cpu.c
index 12b7458..8c397c9 100644
--- a/kernel/cpu.c
+++ b/kernel/cpu.c
@@ -536,6 +536,11 @@ static DECLARE_BITMAP(cpu_possible_bits, CONFIG_NR_CPUS) __read_mostly;
 const struct cpumask *const cpu_possible_mask = to_cpumask(cpu_possible_bits);
 EXPORT_SYMBOL(cpu_possible_mask);
 
+static DECLARE_BITMAP(cpu_hotpluggable_bits, CONFIG_NR_CPUS) __read_mostly;
+const struct cpumask *const cpu_hotpluggable_mask =
+					to_cpumask(cpu_hotpluggable_bits);
+EXPORT_SYMBOL(cpu_hotpluggable_mask);
+
 static DECLARE_BITMAP(cpu_online_bits, CONFIG_NR_CPUS) __read_mostly;
 const struct cpumask *const cpu_online_mask = to_cpumask(cpu_online_bits);
 EXPORT_SYMBOL(cpu_online_mask);
@@ -556,6 +561,14 @@ void set_cpu_possible(unsigned int cpu, bool possible)
 		cpumask_clear_cpu(cpu, to_cpumask(cpu_possible_bits));
 }
 
+void set_cpu_hotpluggable(unsigned int cpu, bool hotpluggable)
+{
+	if (hotpluggable)
+		cpumask_set_cpu(cpu, to_cpumask(cpu_hotpluggable_bits));
+	else
+		cpumask_clear_cpu(cpu, to_cpumask(cpu_hotpluggable_bits));
+}
+
 void set_cpu_present(unsigned int cpu, bool present)
 {
 	if (present)
@@ -590,6 +603,11 @@ void init_cpu_possible(const struct cpumask *src)
 	cpumask_copy(to_cpumask(cpu_possible_bits), src);
 }
 
+void init_cpu_hotpluggable(const struct cpumask *src)
+{
+	cpumask_copy(to_cpumask(cpu_hotpluggable_bits), src);
+}
+
 void init_cpu_online(const struct cpumask *src)
 {
 	cpumask_copy(to_cpumask(cpu_online_bits), src);
-- 
1.7.4.1

[PATCH v3 2/2] cpu: update cpu_hotpluggable_mask in register_cpu

[Mike Turquette]

Update the cpu_hotpluggable_mask for each registered CPU which supports
hotplug.  This makes it trivial for kernel code to know which CPUs
support hotplug operations.

Signed-off-by: Mike Turquette <mturquette@ti.com>
---
Change log:
v2: no change
v3: no change

 drivers/base/cpu.c |    4 +++-
 1 files changed, 3 insertions(+), 1 deletions(-)

diff --git a/drivers/base/cpu.c b/drivers/base/cpu.c
index 251acea..91ddcf8 100644
--- a/drivers/base/cpu.c
+++ b/drivers/base/cpu.c
@@ -224,8 +224,10 @@ int __cpuinit register_cpu(struct cpu *cpu, int num)
 
 	error = sysdev_register(&cpu->sysdev);
 
-	if (!error && cpu->hotpluggable)
+	if (!error && cpu->hotpluggable) {
 		register_cpu_control(cpu);
+		set_cpu_hotpluggable(num, true);
+	}
 	if (!error)
 		per_cpu(cpu_sys_devices, num) = &cpu->sysdev;
 	if (!error)
-- 
1.7.4.1

Re: [PATCH v3 1/2] cpumask: introduce cpumask for hotpluggable CPUs

[Peter Zijlstra]

On Thu, 2011-08-11 at 12:54 -0700, Mike Turquette wrote:
> On some platforms it is possible to have some CPUs which support CPU
> hotplug and some which do not.  Currently the prescence of an 'online'
> sysfs entry in userspace is adequate for applications to know that a CPU
> supports hotplug, but there is no convenient way to make the same
> determination in the kernel.
> 
> To better model this relationship this patch introduces a new cpumask to
> track CPUs that support CPU hotplug operations.
> 
> This new cpumask is populated at boot-time and remains static for the
> life of the machine.  Bits set in the mask indicate a CPU which supports
> hotplug, but make no guarantees about whether that CPU is currently
> online or not.  Likewise a cleared bit in the mask indicates either a
> CPU which cannot hotplug or a lack of a populated CPU.
> 
> The purpose of this new cpumask is to aid kernel code which uses CPU to
> take CPUs online and offline.  Possible uses are as a thermal event
> mitigation technique or as a power capping mechanism. 

Nacked-by: Peter Zijlstra <a.p.zijlstra@chello.nl>

the kernel really shouldn't be using hotplug for this (nor should
userspace really). hot-unplugging random cpus wrecks things like
cpusets. Furthermore hotplug does way too much work to use as a simple
means to idle a cpu.

Even the availability of this mask is wrong, since that implies the
information is useful, which per the above it is not, the kernel
shouldn't care about this full-stop.

The only reason for the OS to unplug a CPU is imminent and unavoidable
hardware failure. Thermal capping is not that (and yes ACPI-4.0 is a
broken piece of shit).