linux-acpi.vger.kernel.org archive mirror
 help / color / mirror / Atom feed
* [PATCH 1/2] acpi : cpu hot-remove returns error when cpu_down() fails
@ 2012-09-28 10:36 Yasuaki Ishimatsu
  2012-09-28 10:39 ` [PATCH 2/2] acpi : prevent cpu from becoming online Yasuaki Ishimatsu
  2012-10-19  1:06 ` [PATCH 1/2] acpi : cpu hot-remove returns error when cpu_down() fails Rafael J. Wysocki
  0 siblings, 2 replies; 5+ messages in thread
From: Yasuaki Ishimatsu @ 2012-09-28 10:36 UTC (permalink / raw)
  To: srivatsa.bhat, toshi.kani, lenb, linux-acpi, linux-kernel

Even if cpu_down() fails, acpi_processor_remove() continues to remove the cpu.
But in this case, it should return error number since some process may run on
the cpu. If the cpu has a running process and the cpu is turned the power off,
the system may not work well.

Reviewed-by: Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
Reviewed-by: Toshi Kani <toshi.kani@hp.com>
Signed-off-by: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>

---
 drivers/acpi/processor_driver.c |   18 ++++++++++++------
 1 file changed, 12 insertions(+), 6 deletions(-)

Index: linux-3.6-rc7/drivers/acpi/processor_driver.c
===================================================================
--- linux-3.6-rc7.orig/drivers/acpi/processor_driver.c	2012-09-24 10:10:57.000000000 +0900
+++ linux-3.6-rc7/drivers/acpi/processor_driver.c	2012-09-28 19:16:33.207858261 +0900
@@ -605,7 +605,7 @@ err_free_pr:
 static int acpi_processor_remove(struct acpi_device *device, int type)
 {
 	struct acpi_processor *pr = NULL;
-
+	int ret;
 
 	if (!device || !acpi_driver_data(device))
 		return -EINVAL;
@@ -616,8 +616,9 @@ static int acpi_processor_remove(struct 
 		goto free;
 
 	if (type == ACPI_BUS_REMOVAL_EJECT) {
-		if (acpi_processor_handle_eject(pr))
-			return -EINVAL;
+		ret = acpi_processor_handle_eject(pr);
+		if (ret)
+			return ret;
 	}
 
 	acpi_processor_power_exit(pr, device);
@@ -848,12 +849,17 @@ static acpi_status acpi_processor_hotadd
 
 static int acpi_processor_handle_eject(struct acpi_processor *pr)
 {
-	if (cpu_online(pr->id))
-		cpu_down(pr->id);
+	int ret = 0;
+
+	if (cpu_online(pr->id)) {
+		ret = cpu_down(pr->id);
+		if (ret)
+			return ret;
+	}
 
 	arch_unregister_cpu(pr->id);
 	acpi_unmap_lsapic(pr->id);
-	return (0);
+	return ret;
 }
 #else
 static acpi_status acpi_processor_hotadd_init(struct acpi_processor *pr)


^ permalink raw reply	[flat|nested] 5+ messages in thread

* [PATCH 2/2] acpi : prevent cpu from becoming online
  2012-09-28 10:36 [PATCH 1/2] acpi : cpu hot-remove returns error when cpu_down() fails Yasuaki Ishimatsu
@ 2012-09-28 10:39 ` Yasuaki Ishimatsu
  2012-10-19  1:06 ` [PATCH 1/2] acpi : cpu hot-remove returns error when cpu_down() fails Rafael J. Wysocki
  1 sibling, 0 replies; 5+ messages in thread
From: Yasuaki Ishimatsu @ 2012-09-28 10:39 UTC (permalink / raw)
  To: srivatsa.bhat, toshi.kani, lenb, linux-acpi, linux-kernel

Even if acpi_processor_handle_eject() offlines cpu, there is a chance
to online the cpu after that. So the patch closes the window by using
get/put_online_cpus().

Why does the patch change _cpu_up() logic?

The patch cares the race of hot-remove cpu and _cpu_up(). If the patch
does not change it, there is the following race.

hot-remove cpu                         |  _cpu_up()
------------------------------------- ------------------------------------
call acpi_processor_handle_eject()     |
     call cpu_down()                   |
     call get_online_cpus()            |
                                       | call cpu_hotplug_begin() and stop here
     call arch_unregister_cpu()        |
     call acpi_unmap_lsapic()          |
     call put_online_cpus()            |
                                       | start and continue _cpu_up()
     return acpi_processor_remove()    |
continue hot-remove the cpu            |

So _cpu_up() can continue to itself. And hot-remove cpu can also continue
itself. If the patch changes _cpu_up() logic, the race disappears as below:

hot-remove cpu                         | _cpu_up()
-----------------------------------------------------------------------
call acpi_processor_handle_eject()     |
     call cpu_down()                   |
     call get_online_cpus()            |
                                       | call cpu_hotplug_begin() and stop here
     call arch_unregister_cpu()        |
     call acpi_unmap_lsapic()          |
          cpu's cpu_present is set     |
          to false by set_cpu_present()|
     call put_online_cpus()            |
                                       | start _cpu_up()
                                       | check cpu_present() and return -EINVAL
     return acpi_processor_remove()    |
continue hot-remove the cpu            |

Reviewed-by: Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
Reviewed-by: Toshi Kani <toshi.kani@hp.com>
Signed-off-by: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>

---
 drivers/acpi/processor_driver.c |   14 ++++++++++++++
 kernel/cpu.c                    |    8 +++++---
 2 files changed, 19 insertions(+), 3 deletions(-)

Index: linux-3.6-rc7/drivers/acpi/processor_driver.c
===================================================================
--- linux-3.6-rc7.orig/drivers/acpi/processor_driver.c	2012-09-28 19:16:33.207858261 +0900
+++ linux-3.6-rc7/drivers/acpi/processor_driver.c	2012-09-28 19:23:05.129858476 +0900
@@ -857,8 +857,22 @@ static int acpi_processor_handle_eject(s
 			return ret;
 	}
 
+	get_online_cpus();
+	/*
+	 * The cpu might become online again at this point. So we check whether
+	 * the cpu has been onlined or not. If the cpu became online, it means
+	 * that someone wants to use the cpu. So acpi_processor_handle_eject()
+	 * returns -EAGAIN.
+	 */
+	if (unlikely(cpu_online(pr->id))) {
+		put_online_cpus();
+		pr_warn("Failed to remove CPU %d, because other task "
+			"brought the CPU back online\n", pr->id);
+		return -EAGAIN;
+	}
 	arch_unregister_cpu(pr->id);
 	acpi_unmap_lsapic(pr->id);
+	put_online_cpus();
 	return ret;
 }
 #else
Index: linux-3.6-rc7/kernel/cpu.c
===================================================================
--- linux-3.6-rc7.orig/kernel/cpu.c	2012-09-24 10:10:57.000000000 +0900
+++ linux-3.6-rc7/kernel/cpu.c	2012-09-28 19:19:32.321858402 +0900
@@ -343,11 +343,13 @@ static int __cpuinit _cpu_up(unsigned in
 	unsigned long mod = tasks_frozen ? CPU_TASKS_FROZEN : 0;
 	struct task_struct *idle;
 
-	if (cpu_online(cpu) || !cpu_present(cpu))
-		return -EINVAL;
-
 	cpu_hotplug_begin();
 
+	if (cpu_online(cpu) || !cpu_present(cpu)) {
+		ret = -EINVAL;
+		goto out;
+	}
+
 	idle = idle_thread_get(cpu);
 	if (IS_ERR(idle)) {
 		ret = PTR_ERR(idle);



^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH 1/2] acpi : cpu hot-remove returns error when cpu_down() fails
  2012-09-28 10:36 [PATCH 1/2] acpi : cpu hot-remove returns error when cpu_down() fails Yasuaki Ishimatsu
  2012-09-28 10:39 ` [PATCH 2/2] acpi : prevent cpu from becoming online Yasuaki Ishimatsu
@ 2012-10-19  1:06 ` Rafael J. Wysocki
  2012-10-19  4:40   ` Yasuaki Ishimatsu
  1 sibling, 1 reply; 5+ messages in thread
From: Rafael J. Wysocki @ 2012-10-19  1:06 UTC (permalink / raw)
  To: Yasuaki Ishimatsu
  Cc: srivatsa.bhat, toshi.kani, lenb, linux-acpi, linux-kernel

On Friday 28 of September 2012 19:36:02 Yasuaki Ishimatsu wrote:
> Even if cpu_down() fails, acpi_processor_remove() continues to remove the cpu.
> But in this case, it should return error number since some process may run on
> the cpu. If the cpu has a running process and the cpu is turned the power off,
> the system may not work well.
> 
> Reviewed-by: Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
> Reviewed-by: Toshi Kani <toshi.kani@hp.com>
> Signed-off-by: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
> 
> ---
>  drivers/acpi/processor_driver.c |   18 ++++++++++++------
>  1 file changed, 12 insertions(+), 6 deletions(-)
> 
> Index: linux-3.6-rc7/drivers/acpi/processor_driver.c
> ===================================================================
> --- linux-3.6-rc7.orig/drivers/acpi/processor_driver.c	2012-09-24 10:10:57.000000000 +0900
> +++ linux-3.6-rc7/drivers/acpi/processor_driver.c	2012-09-28 19:16:33.207858261 +0900
> @@ -605,7 +605,7 @@ err_free_pr:
>  static int acpi_processor_remove(struct acpi_device *device, int type)
>  {
>  	struct acpi_processor *pr = NULL;
> -
> +	int ret;
>  
>  	if (!device || !acpi_driver_data(device))
>  		return -EINVAL;
> @@ -616,8 +616,9 @@ static int acpi_processor_remove(struct 
>  		goto free;
>  
>  	if (type == ACPI_BUS_REMOVAL_EJECT) {
> -		if (acpi_processor_handle_eject(pr))
> -			return -EINVAL;
> +		ret = acpi_processor_handle_eject(pr);
> +		if (ret)
> +			return ret;
>  	}
>  
>  	acpi_processor_power_exit(pr, device);
> @@ -848,12 +849,17 @@ static acpi_status acpi_processor_hotadd
>  
>  static int acpi_processor_handle_eject(struct acpi_processor *pr)
>  {
> -	if (cpu_online(pr->id))
> -		cpu_down(pr->id);
> +	int ret = 0;
> +
> +	if (cpu_online(pr->id)) {
> +		ret = cpu_down(pr->id);

If you defined ret here ...

> +		if (ret)
> +			return ret;
> +	}
>  
>  	arch_unregister_cpu(pr->id);
>  	acpi_unmap_lsapic(pr->id);
> -	return (0);
> +	return ret;

... this line wouldn't need to be changed.

>  }
>  #else
>  static acpi_status acpi_processor_hotadd_init(struct acpi_processor *pr)

Thanks,
Rafael


-- 
I speak only for myself.
Rafael J. Wysocki, Intel Open Source Technology Center.

^ permalink raw reply	[flat|nested] 5+ messages in thread

* Re: [PATCH 1/2] acpi : cpu hot-remove returns error when cpu_down() fails
  2012-10-19  1:06 ` [PATCH 1/2] acpi : cpu hot-remove returns error when cpu_down() fails Rafael J. Wysocki
@ 2012-10-19  4:40   ` Yasuaki Ishimatsu
  2012-10-26  9:25     ` [PATCH v2 " Yasuaki Ishimatsu
  0 siblings, 1 reply; 5+ messages in thread
From: Yasuaki Ishimatsu @ 2012-10-19  4:40 UTC (permalink / raw)
  To: Rafael J. Wysocki
  Cc: srivatsa.bhat, toshi.kani, lenb, linux-acpi, linux-kernel

Hi Rafael,

2012/10/19 10:06, Rafael J. Wysocki wrote:
> On Friday 28 of September 2012 19:36:02 Yasuaki Ishimatsu wrote:
>> Even if cpu_down() fails, acpi_processor_remove() continues to remove the cpu.
>> But in this case, it should return error number since some process may run on
>> the cpu. If the cpu has a running process and the cpu is turned the power off,
>> the system may not work well.
>>
>> Reviewed-by: Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
>> Reviewed-by: Toshi Kani <toshi.kani@hp.com>
>> Signed-off-by: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>
>>
>> ---
>>   drivers/acpi/processor_driver.c |   18 ++++++++++++------
>>   1 file changed, 12 insertions(+), 6 deletions(-)
>>
>> Index: linux-3.6-rc7/drivers/acpi/processor_driver.c
>> ===================================================================
>> --- linux-3.6-rc7.orig/drivers/acpi/processor_driver.c	2012-09-24 10:10:57.000000000 +0900
>> +++ linux-3.6-rc7/drivers/acpi/processor_driver.c	2012-09-28 19:16:33.207858261 +0900
>> @@ -605,7 +605,7 @@ err_free_pr:
>>   static int acpi_processor_remove(struct acpi_device *device, int type)
>>   {
>>   	struct acpi_processor *pr = NULL;
>> -
>> +	int ret;
>>
>>   	if (!device || !acpi_driver_data(device))
>>   		return -EINVAL;
>> @@ -616,8 +616,9 @@ static int acpi_processor_remove(struct
>>   		goto free;
>>
>>   	if (type == ACPI_BUS_REMOVAL_EJECT) {
>> -		if (acpi_processor_handle_eject(pr))
>> -			return -EINVAL;
>> +		ret = acpi_processor_handle_eject(pr);
>> +		if (ret)
>> +			return ret;
>>   	}
>>
>>   	acpi_processor_power_exit(pr, device);
>> @@ -848,12 +849,17 @@ static acpi_status acpi_processor_hotadd
>>
>>   static int acpi_processor_handle_eject(struct acpi_processor *pr)
>>   {
>> -	if (cpu_online(pr->id))
>> -		cpu_down(pr->id);
>> +	int ret = 0;
>> +
>> +	if (cpu_online(pr->id)) {
>> +		ret = cpu_down(pr->id);
>
> If you defined ret here ...
>
>> +		if (ret)
>> +			return ret;
>> +	}
>>

>>   	arch_unregister_cpu(pr->id);
>>   	acpi_unmap_lsapic(pr->id);
>> -	return (0);
>> +	return ret;
>
> ... this line wouldn't need to be changed.

Thank you for your review.
O.K. I'll put the return code back.

Thanks,
Yasuaki Ishimatsu

>
>>   }
>>   #else
>>   static acpi_status acpi_processor_hotadd_init(struct acpi_processor *pr)
>
> Thanks,
> Rafael
>
>



^ permalink raw reply	[flat|nested] 5+ messages in thread

* [PATCH v2 1/2] acpi : cpu hot-remove returns error when cpu_down() fails
  2012-10-19  4:40   ` Yasuaki Ishimatsu
@ 2012-10-26  9:25     ` Yasuaki Ishimatsu
  0 siblings, 0 replies; 5+ messages in thread
From: Yasuaki Ishimatsu @ 2012-10-26  9:25 UTC (permalink / raw)
  To: Rafael J. Wysocki
  Cc: srivatsa.bhat, toshi.kani, lenb, linux-acpi, linux-kernel


v1 -> v2
- update Rafael's comment.

>>   static int acpi_processor_handle_eject(struct acpi_processor *pr)
>>   {
>> -    if (cpu_online(pr->id))
>> -        cpu_down(pr->id);
>> +    int ret = 0;
>> +
>> +    if (cpu_online(pr->id)) {
>> +        ret = cpu_down(pr->id);
>
> If you defined ret here ...
>
>> +        if (ret)
>> +            return ret;
>> +    }
>>

>>       arch_unregister_cpu(pr->id);
>>       acpi_unmap_lsapic(pr->id);
>> -    return (0);
>> +    return ret;
>
> ... this line wouldn't need to be changed.

---
Even if cpu_down() fails, acpi_processor_remove() continues to remove the cpu.
But in this case, it should return error number since some process may run on
the cpu. If the cpu has a running process and the cpu is turned the power off,
the system may not work well.

Reviewed-by: Srivatsa S. Bhat <srivatsa.bhat@linux.vnet.ibm.com>
Reviewed-by: Toshi Kani <toshi.kani@hp.com>
Signed-off-by: Yasuaki Ishimatsu <isimatu.yasuaki@jp.fujitsu.com>

---
  drivers/acpi/processor_driver.c |   16 +++++++++++-----
  1 file changed, 11 insertions(+), 5 deletions(-)

Index: linux-3.7-rc2/drivers/acpi/processor_driver.c
===================================================================
--- linux-3.7-rc2.orig/drivers/acpi/processor_driver.c	2012-10-21 04:11:32.000000000 +0900
+++ linux-3.7-rc2/drivers/acpi/processor_driver.c	2012-10-26 18:15:43.721665836 +0900
@@ -605,7 +605,7 @@ err_free_pr:
  static int acpi_processor_remove(struct acpi_device *device, int type)
  {
  	struct acpi_processor *pr = NULL;
-
+	int ret;
  
  	if (!device || !acpi_driver_data(device))
  		return -EINVAL;
@@ -616,8 +616,9 @@ static int acpi_processor_remove(struct
  		goto free;
  
  	if (type == ACPI_BUS_REMOVAL_EJECT) {
-		if (acpi_processor_handle_eject(pr))
-			return -EINVAL;
+		ret = acpi_processor_handle_eject(pr);
+		if (ret)
+			return ret;
  	}
  
  	acpi_processor_power_exit(pr);
@@ -848,8 +849,13 @@ static acpi_status acpi_processor_hotadd
  
  static int acpi_processor_handle_eject(struct acpi_processor *pr)
  {
-	if (cpu_online(pr->id))
-		cpu_down(pr->id);
+	int ret = 0;
+
+	if (cpu_online(pr->id)) {
+		ret = cpu_down(pr->id);
+		if (ret)
+			return ret;
+	}
  
  	arch_unregister_cpu(pr->id);
  	acpi_unmap_lsapic(pr->id);


^ permalink raw reply	[flat|nested] 5+ messages in thread

end of thread, other threads:[~2012-10-26  9:26 UTC | newest]

Thread overview: 5+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2012-09-28 10:36 [PATCH 1/2] acpi : cpu hot-remove returns error when cpu_down() fails Yasuaki Ishimatsu
2012-09-28 10:39 ` [PATCH 2/2] acpi : prevent cpu from becoming online Yasuaki Ishimatsu
2012-10-19  1:06 ` [PATCH 1/2] acpi : cpu hot-remove returns error when cpu_down() fails Rafael J. Wysocki
2012-10-19  4:40   ` Yasuaki Ishimatsu
2012-10-26  9:25     ` [PATCH v2 " Yasuaki Ishimatsu

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).