Groups | Search | Server Info | Keyboard shortcuts | Login | Register [http] [https] [nntp] [nntps]

Groups > linux.kernel > #1514578 > unrolled thread

cpu hotplug: convert more drivers (batch #4)

Back to article view | Back to linux.kernel

Contents

  cpu hotplug: convert more drivers (batch #4) Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-11-03 16:00 +0100
    [PATCH 04/25] lib/percpu_counter: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-11-03 16:00 +0100
      [tip:smp/hotplug] lib/percpu_counter: Convert to hotplug state  machine tip-bot for Sebastian Andrzej Siewior <tipbot@zytor.com> - 2016-11-10 00:00 +0100
    [PATCH 10/25] s390/smp: Make cpu notifier symetric Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-11-03 16:00 +0100
      Re: [PATCH 10/25] s390/smp: Make cpu notifier symetric Heiko Carstens <heiko.carstens@de.ibm.com> - 2016-11-04 15:30 +0100
        [PATCH 10/25 v2] s390/smp: Make cpu notifier symetric Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-11-04 15:50 +0100
          [tip:smp/hotplug] s390/smp: Make cpu notifier symetric tip-bot for Thomas Gleixner <tipbot@zytor.com> - 2016-11-10 00:00 +0100
    [PATCH 11/25] s390/smp: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-11-03 16:00 +0100
      Re: [PATCH 11/25] s390/smp: Convert to hotplug state machine Heiko Carstens <heiko.carstens@de.ibm.com> - 2016-11-04 15:40 +0100
        [PATCH 11/25 v2] s390/smp: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-11-04 15:50 +0100
          [tip:smp/hotplug] s390/smp: Convert to hotplug state machine tip-bot for Sebastian Andrzej Siewior <tipbot@zytor.com> - 2016-11-10 00:00 +0100
    [PATCH 25/25] x86/mcheck: Move CPU_DEAD to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-11-03 16:00 +0100
    [PATCH 14/25] ia64/err-inject: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-11-03 16:00 +0100
      [tip:smp/hotplug] ia64/err-inject: Convert to hotplug state machine tip-bot for Sebastian Andrzej Siewior <tipbot@zytor.com> - 2016-11-10 00:00 +0100
    [PATCH 05/25] lib/radix-tree: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-11-03 16:00 +0100
      [tip:smp/hotplug] lib/radix-tree: Convert to hotplug state machine tip-bot for Sebastian Andrzej Siewior <tipbot@zytor.com> - 2016-11-10 00:00 +0100
    [PATCH 23/25] x86/mcheck: Make CPU_DOWN_PREPARE the counter part of CPU_STARTING Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-11-03 16:00 +0100
    [PATCH 02/25] kernel/printk: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-11-03 16:00 +0100
      [tip:smp/hotplug] kernel/printk: Convert to hotplug state machine tip-bot for Sebastian Andrzej Siewior <tipbot@zytor.com> - 2016-11-10 00:00 +0100
    [PATCH 06/25] mm/page_alloc: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-11-03 16:00 +0100
      [tip:smp/hotplug] mm/page_alloc: Convert to hotplug state machine tip-bot for Sebastian Andrzej Siewior <tipbot@zytor.com> - 2016-11-10 00:00 +0100
    [PATCH 18/25] x86/mcheck: Move threshold_create_device() Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-11-03 16:00 +0100
      Re: [PATCH 18/25] x86/mcheck: Move threshold_create_device() Borislav Petkov <bp@alien8.de> - 2016-11-07 11:40 +0100
    [PATCH 09/25] net/flowcache: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-11-03 16:00 +0100
      [tip:smp/hotplug] net/flowcache: Convert to hotplug state machine tip-bot for Sebastian Andrzej Siewior <tipbot@zytor.com> - 2016-11-10 00:00 +0100
    [PATCH 21/25] x86/mcheck: Split threshold_cpu_callback into two callbacks Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-11-03 16:00 +0100
      Re: [PATCH 21/25] x86/mcheck: Split threshold_cpu_callback into two  callbacks Borislav Petkov <bp@alien8.de> - 2016-11-07 14:30 +0100
        Re: [PATCH 21/25] x86/mcheck: Split threshold_cpu_callback into two  callbacks Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-11-07 14:30 +0100
          Re: [PATCH 21/25] x86/mcheck: Split threshold_cpu_callback into two  callbacks Borislav Petkov <bp@alien8.de> - 2016-11-07 16:10 +0100
            Re: [PATCH 21/25] x86/mcheck: Split threshold_cpu_callback into two  callbacks Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-11-07 16:20 +0100
            Re: [PATCH 21/25] x86/mcheck: Split threshold_cpu_callback into two  callbacks Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-11-07 18:30 +0100
              Re: [PATCH 21/25] x86/mcheck: Split threshold_cpu_callback into two  callbacks Borislav Petkov <bp@alien8.de> - 2016-11-07 19:20 +0100
    [PATCH 08/25] net/dev: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-11-03 16:00 +0100
      [tip:smp/hotplug] net/dev: Convert to hotplug state machine tip-bot for Sebastian Andrzej Siewior <tipbot@zytor.com> - 2016-11-10 00:00 +0100
    [PATCH 17/25] ia64/topology: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-11-03 16:00 +0100
      [PATCH 17/25 v2] ia64/topology: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-11-03 18:40 +0100
        [tip:smp/hotplug] ia64/topology: Convert to hotplug state machine tip-bot for Sebastian Andrzej Siewior <tipbot@zytor.com> - 2016-11-10 00:10 +0100

Page 1 of 2  [1] 2  Next page →

#1514578 — cpu hotplug: convert more drivers (batch #4)

Another small batch of drivers converted to the new hotplug state engine.
The whole series is also available at
  git://git.kernel.org/pub/scm/linux/kernel/git/bigeasy/hotplug-staging.git smp/hotplug

on top of v4.9-rc3.

Sebastian

[toc] | [next] | [standalone]

#1514580 — [PATCH 04/25] lib/percpu_counter: Convert to hotplug state machine

Install the callbacks via the state machine and let the core invoke
the callbacks on the already online CPUs.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
 include/linux/cpuhotplug.h |  1 +
 lib/percpu_counter.c       | 25 ++++++++++++++-----------
 2 files changed, 15 insertions(+), 11 deletions(-)

diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h
index c622ab349af3..04e5f99ffc70 100644
--- a/include/linux/cpuhotplug.h
+++ b/include/linux/cpuhotplug.h
@@ -33,6 +33,7 @@ enum cpuhp_state {
 	CPUHP_FS_BUFF_DEAD,
 	CPUHP_PRINTK_DEAD,
 	CPUHP_MM_MEMCQ_DEAD,
+	CPUHP_PERCPU_CNT_DEAD,
 	CPUHP_WORKQUEUE_PREP,
 	CPUHP_POWER_NUMA_PREPARE,
 	CPUHP_HRTIMERS_PREPARE,
diff --git a/lib/percpu_counter.c b/lib/percpu_counter.c
index 72d36113ccaa..c8cebb137076 100644
--- a/lib/percpu_counter.c
+++ b/lib/percpu_counter.c
@@ -158,25 +158,21 @@ EXPORT_SYMBOL(percpu_counter_destroy);
 int percpu_counter_batch __read_mostly = 32;
 EXPORT_SYMBOL(percpu_counter_batch);
 
-static void compute_batch_value(void)
+static int compute_batch_value(unsigned int cpu)
 {
 	int nr = num_online_cpus();
 
 	percpu_counter_batch = max(32, nr*2);
+	return 0;
 }
 
-static int percpu_counter_hotcpu_callback(struct notifier_block *nb,
-					unsigned long action, void *hcpu)
+static int percpu_counter_cpu_dead(unsigned int cpu)
 {
 #ifdef CONFIG_HOTPLUG_CPU
-	unsigned int cpu;
 	struct percpu_counter *fbc;
 
-	compute_batch_value();
-	if (action != CPU_DEAD && action != CPU_DEAD_FROZEN)
-		return NOTIFY_OK;
+	compute_batch_value(cpu);
 
-	cpu = (unsigned long)hcpu;
 	spin_lock_irq(&percpu_counters_lock);
 	list_for_each_entry(fbc, &percpu_counters, list) {
 		s32 *pcount;
@@ -190,7 +186,7 @@ static int percpu_counter_hotcpu_callback(struct notifier_block *nb,
 	}
 	spin_unlock_irq(&percpu_counters_lock);
 #endif
-	return NOTIFY_OK;
+	return 0;
 }
 
 /*
@@ -222,8 +218,15 @@ EXPORT_SYMBOL(__percpu_counter_compare);
 
 static int __init percpu_counter_startup(void)
 {
-	compute_batch_value();
-	hotcpu_notifier(percpu_counter_hotcpu_callback, 0);
+	int ret;
+
+	ret = cpuhp_setup_state(CPUHP_AP_ONLINE_DYN, "lib/percpu_cnt:online",
+				compute_batch_value, NULL);
+	WARN_ON(ret < 0);
+	ret = cpuhp_setup_state_nocalls(CPUHP_PERCPU_CNT_DEAD,
+					"lib/percpu_cnt:dead", NULL,
+					percpu_counter_cpu_dead);
+	WARN_ON(ret < 0);
 	return 0;
 }
 module_init(percpu_counter_startup);
-- 
2.10.2

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

#1518550 — [tip:smp/hotplug] lib/percpu_counter: Convert to hotplug state machine

Commit-ID:  5588f5afb4cfc33eb377b751ba4b97184373e8d6
Gitweb:     http://git.kernel.org/tip/5588f5afb4cfc33eb377b751ba4b97184373e8d6
Author:     Sebastian Andrzej Siewior <bigeasy@linutronix.de>
AuthorDate: Thu, 3 Nov 2016 15:50:00 +0100
Committer:  Thomas Gleixner <tglx@linutronix.de>
CommitDate: Wed, 9 Nov 2016 23:45:26 +0100

lib/percpu_counter: Convert to hotplug state machine

Install the callbacks via the state machine and let the core invoke
the callbacks on the already online CPUs.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: rt@linutronix.de
Link: http://lkml.kernel.org/r/20161103145021.28528-5-bigeasy@linutronix.de
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

---
 include/linux/cpuhotplug.h |  1 +
 lib/percpu_counter.c       | 25 ++++++++++++++-----------
 2 files changed, 15 insertions(+), 11 deletions(-)

diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h
index c622ab3..04e5f99 100644
--- a/include/linux/cpuhotplug.h
+++ b/include/linux/cpuhotplug.h
@@ -33,6 +33,7 @@ enum cpuhp_state {
 	CPUHP_FS_BUFF_DEAD,
 	CPUHP_PRINTK_DEAD,
 	CPUHP_MM_MEMCQ_DEAD,
+	CPUHP_PERCPU_CNT_DEAD,
 	CPUHP_WORKQUEUE_PREP,
 	CPUHP_POWER_NUMA_PREPARE,
 	CPUHP_HRTIMERS_PREPARE,
diff --git a/lib/percpu_counter.c b/lib/percpu_counter.c
index 72d3611..c8cebb1 100644
--- a/lib/percpu_counter.c
+++ b/lib/percpu_counter.c
@@ -158,25 +158,21 @@ EXPORT_SYMBOL(percpu_counter_destroy);
 int percpu_counter_batch __read_mostly = 32;
 EXPORT_SYMBOL(percpu_counter_batch);
 
-static void compute_batch_value(void)
+static int compute_batch_value(unsigned int cpu)
 {
 	int nr = num_online_cpus();
 
 	percpu_counter_batch = max(32, nr*2);
+	return 0;
 }
 
-static int percpu_counter_hotcpu_callback(struct notifier_block *nb,
-					unsigned long action, void *hcpu)
+static int percpu_counter_cpu_dead(unsigned int cpu)
 {
 #ifdef CONFIG_HOTPLUG_CPU
-	unsigned int cpu;
 	struct percpu_counter *fbc;
 
-	compute_batch_value();
-	if (action != CPU_DEAD && action != CPU_DEAD_FROZEN)
-		return NOTIFY_OK;
+	compute_batch_value(cpu);
 
-	cpu = (unsigned long)hcpu;
 	spin_lock_irq(&percpu_counters_lock);
 	list_for_each_entry(fbc, &percpu_counters, list) {
 		s32 *pcount;
@@ -190,7 +186,7 @@ static int percpu_counter_hotcpu_callback(struct notifier_block *nb,
 	}
 	spin_unlock_irq(&percpu_counters_lock);
 #endif
-	return NOTIFY_OK;
+	return 0;
 }
 
 /*
@@ -222,8 +218,15 @@ EXPORT_SYMBOL(__percpu_counter_compare);
 
 static int __init percpu_counter_startup(void)
 {
-	compute_batch_value();
-	hotcpu_notifier(percpu_counter_hotcpu_callback, 0);
+	int ret;
+
+	ret = cpuhp_setup_state(CPUHP_AP_ONLINE_DYN, "lib/percpu_cnt:online",
+				compute_batch_value, NULL);
+	WARN_ON(ret < 0);
+	ret = cpuhp_setup_state_nocalls(CPUHP_PERCPU_CNT_DEAD,
+					"lib/percpu_cnt:dead", NULL,
+					percpu_counter_cpu_dead);
+	WARN_ON(ret < 0);
 	return 0;
 }
 module_init(percpu_counter_startup);

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

#1514581 — [PATCH 10/25] s390/smp: Make cpu notifier symetric

From: Thomas Gleixner <tglx@linutronix.de>

There is no reason to remove the sysfs cpu files when the CPU is dead, they
can be removed when the cpu is prepared to go down. Doing it at
DOWN_PREPARE allows us to convert it to a symetric hotplug state in the
next step.

Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: linux-s390@vger.kernel.org
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
 arch/s390/kernel/smp.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/arch/s390/kernel/smp.c b/arch/s390/kernel/smp.c
index 35531fe1c5ea..1a21e66c484a 100644
--- a/arch/s390/kernel/smp.c
+++ b/arch/s390/kernel/smp.c
@@ -1056,9 +1056,10 @@ static int smp_cpu_notify(struct notifier_block *self, unsigned long action,
 
 	switch (action & ~CPU_TASKS_FROZEN) {
 	case CPU_ONLINE:
+	case CPU_DOWN_FAILED:
 		err = sysfs_create_group(&s->kobj, &cpu_online_attr_group);
 		break;
-	case CPU_DEAD:
+	case CPU_DOWN_PREPARE::
 		sysfs_remove_group(&s->kobj, &cpu_online_attr_group);
 		break;
 	}
-- 
2.10.2

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

#1515215 — Re: [PATCH 10/25] s390/smp: Make cpu notifier symetric

On Thu, Nov 03, 2016 at 03:50:06PM +0100, Sebastian Andrzej Siewior wrote:
> From: Thomas Gleixner <tglx@linutronix.de>
> 
> There is no reason to remove the sysfs cpu files when the CPU is dead, they
> can be removed when the cpu is prepared to go down. Doing it at
> DOWN_PREPARE allows us to convert it to a symetric hotplug state in the
> next step.
> 
> Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
> Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
> Cc: linux-s390@vger.kernel.org
> Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
> Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
> ---
>  arch/s390/kernel/smp.c | 3 ++-
>  1 file changed, 2 insertions(+), 1 deletion(-)
> 
> diff --git a/arch/s390/kernel/smp.c b/arch/s390/kernel/smp.c
> index 35531fe1c5ea..1a21e66c484a 100644
> --- a/arch/s390/kernel/smp.c
> +++ b/arch/s390/kernel/smp.c
> @@ -1056,9 +1056,10 @@ static int smp_cpu_notify(struct notifier_block *self, unsigned long action,
>  
>  	switch (action & ~CPU_TASKS_FROZEN) {
>  	case CPU_ONLINE:
> +	case CPU_DOWN_FAILED:
>  		err = sysfs_create_group(&s->kobj, &cpu_online_attr_group);
>  		break;
> -	case CPU_DEAD:
> +	case CPU_DOWN_PREPARE::

This won't compile... even though it will be removed with the following
patch it would be good to fix this to keep this bisectable.

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

#1515222 — [PATCH 10/25 v2] s390/smp: Make cpu notifier symetric

From: Thomas Gleixner <tglx@linutronix.de>

There is no reason to remove the sysfs cpu files when the CPU is dead, they
can be removed when the cpu is prepared to go down. Doing it at
DOWN_PREPARE allows us to convert it to a symetric hotplug state in the
next step.

Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: linux-s390@vger.kernel.org
Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
v1…v2: s/::/

 arch/s390/kernel/smp.c |    3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

--- a/arch/s390/kernel/smp.c
+++ b/arch/s390/kernel/smp.c
@@ -1056,9 +1056,10 @@ static int smp_cpu_notify(struct notifie
 
 	switch (action & ~CPU_TASKS_FROZEN) {
 	case CPU_ONLINE:
+	case CPU_DOWN_FAILED:
 		err = sysfs_create_group(&s->kobj, &cpu_online_attr_group);
 		break;
-	case CPU_DEAD:
+	case CPU_DOWN_PREPARE:
 		sysfs_remove_group(&s->kobj, &cpu_online_attr_group);
 		break;
 	}

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

#1518549 — [tip:smp/hotplug] s390/smp: Make cpu notifier symetric

Commit-ID:  ef65d45cbfbb438a5fbe8ef3fc424314ff1e8b7c
Gitweb:     http://git.kernel.org/tip/ef65d45cbfbb438a5fbe8ef3fc424314ff1e8b7c
Author:     Thomas Gleixner <tglx@linutronix.de>
AuthorDate: Fri, 4 Nov 2016 15:41:41 +0100
Committer:  Thomas Gleixner <tglx@linutronix.de>
CommitDate: Wed, 9 Nov 2016 23:45:28 +0100

s390/smp: Make cpu notifier symetric

There is no reason to remove the sysfs cpu files when the CPU is dead, they
can be removed when the cpu is prepared to go down. Doing it at
DOWN_PREPARE allows us to convert it to a symetric hotplug state in the
next step.

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: linux-s390@vger.kernel.org
Cc: rt@linutronix.de
Link: http://lkml.kernel.org/r/20161104144140.lcee6kwmwlx37m7g@linutronix.de
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

---
 arch/s390/kernel/smp.c | 3 ++-
 1 file changed, 2 insertions(+), 1 deletion(-)

diff --git a/arch/s390/kernel/smp.c b/arch/s390/kernel/smp.c
index 35531fe..2a9c03d 100644
--- a/arch/s390/kernel/smp.c
+++ b/arch/s390/kernel/smp.c
@@ -1056,9 +1056,10 @@ static int smp_cpu_notify(struct notifier_block *self, unsigned long action,
 
 	switch (action & ~CPU_TASKS_FROZEN) {
 	case CPU_ONLINE:
+	case CPU_DOWN_FAILED:
 		err = sysfs_create_group(&s->kobj, &cpu_online_attr_group);
 		break;
-	case CPU_DEAD:
+	case CPU_DOWN_PREPARE:
 		sysfs_remove_group(&s->kobj, &cpu_online_attr_group);
 		break;
 	}

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

#1514582 — [PATCH 11/25] s390/smp: Convert to hotplug state machine

cpuhp_setup_state() invokes the startup callback on all online cpus with
the proper protection, so we can remove the cpu hotplug protection from the
init function and the creation of the per cpu files for online cpus in
smp_add_present_cpu(). smp_add_present_cpu() is called also called from
__smp_rescan_cpus(), but this callpath never adds an online cpu, it merily
adds newly present cpus, so the creation of the cpu files is not required.

Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: linux-s390@vger.kernel.org
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
 arch/s390/kernel/smp.c | 38 ++++++++++++--------------------------
 1 file changed, 12 insertions(+), 26 deletions(-)

diff --git a/arch/s390/kernel/smp.c b/arch/s390/kernel/smp.c
index 1a21e66c484a..b613aa989e40 100644
--- a/arch/s390/kernel/smp.c
+++ b/arch/s390/kernel/smp.c
@@ -1047,23 +1047,18 @@ static struct attribute_group cpu_online_attr_group = {
 	.attrs = cpu_online_attrs,
 };
 
-static int smp_cpu_notify(struct notifier_block *self, unsigned long action,
-			  void *hcpu)
+static int smp_cpu_online(unsigned int cpu)
 {
-	unsigned int cpu = (unsigned int)(long)hcpu;
 	struct device *s = &per_cpu(cpu_device, cpu)->dev;
-	int err = 0;
 
-	switch (action & ~CPU_TASKS_FROZEN) {
-	case CPU_ONLINE:
-	case CPU_DOWN_FAILED:
-		err = sysfs_create_group(&s->kobj, &cpu_online_attr_group);
-		break;
-	case CPU_DOWN_PREPARE::
-		sysfs_remove_group(&s->kobj, &cpu_online_attr_group);
-		break;
-	}
-	return notifier_from_errno(err);
+	return sysfs_create_group(&s->kobj, &cpu_online_attr_group);
+}
+static int smp_cpu_pre_down(unsigned int cpu)
+{
+	struct device *s = &per_cpu(cpu_device, cpu)->dev;
+
+	sysfs_remove_group(&s->kobj, &cpu_online_attr_group);
+	return 0;
 }
 
 static int smp_add_present_cpu(int cpu)
@@ -1084,20 +1079,13 @@ static int smp_add_present_cpu(int cpu)
 	rc = sysfs_create_group(&s->kobj, &cpu_common_attr_group);
 	if (rc)
 		goto out_cpu;
-	if (cpu_online(cpu)) {
-		rc = sysfs_create_group(&s->kobj, &cpu_online_attr_group);
-		if (rc)
-			goto out_online;
-	}
+
 	rc = topology_cpu_init(c);
 	if (rc)
 		goto out_topology;
 	return 0;
 
 out_topology:
-	if (cpu_online(cpu))
-		sysfs_remove_group(&s->kobj, &cpu_online_attr_group);
-out_online:
 	sysfs_remove_group(&s->kobj, &cpu_common_attr_group);
 out_cpu:
 #ifdef CONFIG_HOTPLUG_CPU
@@ -1150,17 +1138,15 @@ static int __init s390_smp_init(void)
 	if (rc)
 		return rc;
 #endif
-	cpu_notifier_register_begin();
 	for_each_present_cpu(cpu) {
 		rc = smp_add_present_cpu(cpu);
 		if (rc)
 			goto out;
 	}
 
-	__hotcpu_notifier(smp_cpu_notify, 0);
-
+	rc = cpuhp_setup_state(CPUHP_AP_ONLINE_DYN, "s390/smp:online",
+			       smp_cpu_online, smp_cpu_pre_down);
 out:
-	cpu_notifier_register_done();
 	return rc;
 }
 subsys_initcall(s390_smp_init);
-- 
2.10.2

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

#1515217 — Re: [PATCH 11/25] s390/smp: Convert to hotplug state machine

On Thu, Nov 03, 2016 at 03:50:07PM +0100, Sebastian Andrzej Siewior wrote:
> @@ -1084,20 +1079,13 @@ static int smp_add_present_cpu(int cpu)
>  	rc = sysfs_create_group(&s->kobj, &cpu_common_attr_group);
>  	if (rc)
>  		goto out_cpu;
> -	if (cpu_online(cpu)) {
> -		rc = sysfs_create_group(&s->kobj, &cpu_online_attr_group);
> -		if (rc)
> -			goto out_online;
> -	}
> +
>  	rc = topology_cpu_init(c);

Please don't add whitespace :)

If you fix the minor nits both s390 related patches are
Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com>

Thanks!

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

#1515230 — [PATCH 11/25 v2] s390/smp: Convert to hotplug state machine

cpuhp_setup_state() invokes the startup callback on all online cpus with
the proper protection, so we can remove the cpu hotplug protection from the
init function and the creation of the per cpu files for online cpus in
smp_add_present_cpu(). smp_add_present_cpu() is called also called from
__smp_rescan_cpus(), but this callpath never adds an online cpu, it merily
adds newly present cpus, so the creation of the cpu files is not required.

Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: linux-s390@vger.kernel.org
Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
v1…v2:
	- removed (or not adding) one empty line
	- rebase on top of the previous patch

 arch/s390/kernel/smp.c |   37 +++++++++++--------------------------
 1 file changed, 11 insertions(+), 26 deletions(-)

--- a/arch/s390/kernel/smp.c
+++ b/arch/s390/kernel/smp.c
@@ -1047,23 +1047,18 @@ static struct attribute_group cpu_online
 	.attrs = cpu_online_attrs,
 };
 
-static int smp_cpu_notify(struct notifier_block *self, unsigned long action,
-			  void *hcpu)
+static int smp_cpu_online(unsigned int cpu)
 {
-	unsigned int cpu = (unsigned int)(long)hcpu;
 	struct device *s = &per_cpu(cpu_device, cpu)->dev;
-	int err = 0;
 
-	switch (action & ~CPU_TASKS_FROZEN) {
-	case CPU_ONLINE:
-	case CPU_DOWN_FAILED:
-		err = sysfs_create_group(&s->kobj, &cpu_online_attr_group);
-		break;
-	case CPU_DOWN_PREPARE:
-		sysfs_remove_group(&s->kobj, &cpu_online_attr_group);
-		break;
-	}
-	return notifier_from_errno(err);
+	return sysfs_create_group(&s->kobj, &cpu_online_attr_group);
+}
+static int smp_cpu_pre_down(unsigned int cpu)
+{
+	struct device *s = &per_cpu(cpu_device, cpu)->dev;
+
+	sysfs_remove_group(&s->kobj, &cpu_online_attr_group);
+	return 0;
 }
 
 static int smp_add_present_cpu(int cpu)
@@ -1084,20 +1079,12 @@ static int smp_add_present_cpu(int cpu)
 	rc = sysfs_create_group(&s->kobj, &cpu_common_attr_group);
 	if (rc)
 		goto out_cpu;
-	if (cpu_online(cpu)) {
-		rc = sysfs_create_group(&s->kobj, &cpu_online_attr_group);
-		if (rc)
-			goto out_online;
-	}
 	rc = topology_cpu_init(c);
 	if (rc)
 		goto out_topology;
 	return 0;
 
 out_topology:
-	if (cpu_online(cpu))
-		sysfs_remove_group(&s->kobj, &cpu_online_attr_group);
-out_online:
 	sysfs_remove_group(&s->kobj, &cpu_common_attr_group);
 out_cpu:
 #ifdef CONFIG_HOTPLUG_CPU
@@ -1150,17 +1137,15 @@ static int __init s390_smp_init(void)
 	if (rc)
 		return rc;
 #endif
-	cpu_notifier_register_begin();
 	for_each_present_cpu(cpu) {
 		rc = smp_add_present_cpu(cpu);
 		if (rc)
 			goto out;
 	}
 
-	__hotcpu_notifier(smp_cpu_notify, 0);
-
+	rc = cpuhp_setup_state(CPUHP_AP_ONLINE_DYN, "s390/smp:online",
+			       smp_cpu_online, smp_cpu_pre_down);
 out:
-	cpu_notifier_register_done();
 	return rc;
 }
 subsys_initcall(s390_smp_init);

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

#1518539 — [tip:smp/hotplug] s390/smp: Convert to hotplug state machine

Commit-ID:  dfbbd86a0f1c3ceec15b64c8f2149a903806ed8c
Gitweb:     http://git.kernel.org/tip/dfbbd86a0f1c3ceec15b64c8f2149a903806ed8c
Author:     Sebastian Andrzej Siewior <bigeasy@linutronix.de>
AuthorDate: Fri, 4 Nov 2016 15:45:03 +0100
Committer:  Thomas Gleixner <tglx@linutronix.de>
CommitDate: Wed, 9 Nov 2016 23:45:28 +0100

s390/smp: Convert to hotplug state machine

cpuhp_setup_state() invokes the startup callback on all online cpus with
the proper protection, so we can remove the cpu hotplug protection from the
init function and the creation of the per cpu files for online cpus in
smp_add_present_cpu(). smp_add_present_cpu() is called also called from
__smp_rescan_cpus(), but this callpath never adds an online cpu, it merily
adds newly present cpus, so the creation of the cpu files is not required.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Acked-by: Heiko Carstens <heiko.carstens@de.ibm.com>
Cc: Martin Schwidefsky <schwidefsky@de.ibm.com>
Cc: linux-s390@vger.kernel.org
Cc: rt@linutronix.de
Link: http://lkml.kernel.org/r/20161104144502.7kd4bxz2rxqvtack@linutronix.de
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

---
 arch/s390/kernel/smp.c | 37 +++++++++++--------------------------
 1 file changed, 11 insertions(+), 26 deletions(-)

diff --git a/arch/s390/kernel/smp.c b/arch/s390/kernel/smp.c
index 2a9c03d..870fb5a 100644
--- a/arch/s390/kernel/smp.c
+++ b/arch/s390/kernel/smp.c
@@ -1047,23 +1047,18 @@ static struct attribute_group cpu_online_attr_group = {
 	.attrs = cpu_online_attrs,
 };
 
-static int smp_cpu_notify(struct notifier_block *self, unsigned long action,
-			  void *hcpu)
+static int smp_cpu_online(unsigned int cpu)
 {
-	unsigned int cpu = (unsigned int)(long)hcpu;
 	struct device *s = &per_cpu(cpu_device, cpu)->dev;
-	int err = 0;
 
-	switch (action & ~CPU_TASKS_FROZEN) {
-	case CPU_ONLINE:
-	case CPU_DOWN_FAILED:
-		err = sysfs_create_group(&s->kobj, &cpu_online_attr_group);
-		break;
-	case CPU_DOWN_PREPARE:
-		sysfs_remove_group(&s->kobj, &cpu_online_attr_group);
-		break;
-	}
-	return notifier_from_errno(err);
+	return sysfs_create_group(&s->kobj, &cpu_online_attr_group);
+}
+static int smp_cpu_pre_down(unsigned int cpu)
+{
+	struct device *s = &per_cpu(cpu_device, cpu)->dev;
+
+	sysfs_remove_group(&s->kobj, &cpu_online_attr_group);
+	return 0;
 }
 
 static int smp_add_present_cpu(int cpu)
@@ -1084,20 +1079,12 @@ static int smp_add_present_cpu(int cpu)
 	rc = sysfs_create_group(&s->kobj, &cpu_common_attr_group);
 	if (rc)
 		goto out_cpu;
-	if (cpu_online(cpu)) {
-		rc = sysfs_create_group(&s->kobj, &cpu_online_attr_group);
-		if (rc)
-			goto out_online;
-	}
 	rc = topology_cpu_init(c);
 	if (rc)
 		goto out_topology;
 	return 0;
 
 out_topology:
-	if (cpu_online(cpu))
-		sysfs_remove_group(&s->kobj, &cpu_online_attr_group);
-out_online:
 	sysfs_remove_group(&s->kobj, &cpu_common_attr_group);
 out_cpu:
 #ifdef CONFIG_HOTPLUG_CPU
@@ -1150,17 +1137,15 @@ static int __init s390_smp_init(void)
 	if (rc)
 		return rc;
 #endif
-	cpu_notifier_register_begin();
 	for_each_present_cpu(cpu) {
 		rc = smp_add_present_cpu(cpu);
 		if (rc)
 			goto out;
 	}
 
-	__hotcpu_notifier(smp_cpu_notify, 0);
-
+	rc = cpuhp_setup_state(CPUHP_AP_ONLINE_DYN, "s390/smp:online",
+			       smp_cpu_online, smp_cpu_pre_down);
 out:
-	cpu_notifier_register_done();
 	return rc;
 }
 subsys_initcall(s390_smp_init);

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

#1514583 — [PATCH 25/25] x86/mcheck: Move CPU_DEAD to hotplug state machine

This moves the last piece of the old hotplug notifier code in MCE to the
new hotplug state machine.

Cc: Tony Luck <tony.luck@intel.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: linux-edac@vger.kernel.org
Cc: x86@kernel.org
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
 arch/x86/kernel/cpu/mcheck/mce.c | 53 +++++++++++++++++-----------------------
 include/linux/cpuhotplug.h       |  1 +
 2 files changed, 23 insertions(+), 31 deletions(-)

diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c
index b1770ebcb8de..1e2036e012e0 100644
--- a/arch/x86/kernel/cpu/mcheck/mce.c
+++ b/arch/x86/kernel/cpu/mcheck/mce.c
@@ -2480,28 +2480,6 @@ static void mce_disable_cpu(void)
 	vendor_disable_error_reporting();
 }
 
-/* Get notified when a cpu comes on/off. Be hotplug friendly. */
-static int
-mce_cpu_callback(struct notifier_block *nfb, unsigned long action, void *hcpu)
-{
-	unsigned int cpu = (unsigned long)hcpu;
-
-	switch (action & ~CPU_TASKS_FROZEN) {
-	case CPU_DEAD:
-		if (threshold_cpu_callback_dead)
-			threshold_cpu_callback_dead(cpu);
-		mce_device_remove(cpu);
-		mce_intel_hcpu_update(cpu);
-
-		/* intentionally ignoring frozen here */
-		if (!(action & CPU_TASKS_FROZEN))
-			cmci_rediscover();
-		break;
-	}
-
-	return NOTIFY_OK;
-}
-
 static int mce_cpu_online(unsigned int cpu)
 {
 	int ret;
@@ -2518,6 +2496,19 @@ static int mce_cpu_online(unsigned int cpu)
 	return 0;
 }
 
+static int mce_cpu_dead(unsigned int cpu)
+{
+	if (threshold_cpu_callback_dead)
+		threshold_cpu_callback_dead(cpu);
+	mce_device_remove(cpu);
+	mce_intel_hcpu_update(cpu);
+
+	/* intentionally ignoring frozen here */
+	if (!cpuhp_tasks_frozen)
+		cmci_rediscover();
+	return 0;
+}
+
 static int mce_cpu_down_dying(unsigned int cpu)
 {
 	struct timer_list *t = this_cpu_ptr(&mce_timer);
@@ -2527,10 +2518,6 @@ static int mce_cpu_down_dying(unsigned int cpu)
 	return 0;
 }
 
-static struct notifier_block mce_cpu_notifier = {
-	.notifier_call = mce_cpu_callback,
-};
-
 static __init void mce_init_banks(void)
 {
 	int i;
@@ -2580,20 +2567,21 @@ static __init int mcheck_init_device(void)
 	if (err)
 		goto err_init_pool;
 
+	err = cpuhp_setup_state(CPUHP_X86_MCE_DEAD, "x86/mce:dead", NULL,
+				mce_cpu_dead);
+	if (err)
+		goto err_init_pool;
+
 	err = cpuhp_setup_state(CPUHP_AP_X86_MCE_STARTING, "x86/mce:starting",
 				mcheck_cpu_starting, mce_cpu_down_dying);
 	if (err)
-		goto err_init_pool;
+		goto err_hp_starting;
 	err = cpuhp_setup_state(CPUHP_AP_ONLINE_DYN, "x86/mce:online",
 				mce_cpu_online, NULL);
 	if (err < 0)
 		goto err_hp_online;
 	hp_online = err;
 
-	cpu_notifier_register_begin();
-	__register_hotcpu_notifier(&mce_cpu_notifier);
-	cpu_notifier_register_done();
-
 	register_syscore_ops(&mce_syscore_ops);
 
 	/* register character device /dev/mcelog */
@@ -2610,6 +2598,9 @@ static __init int mcheck_init_device(void)
 err_hp_online:
 	cpuhp_remove_state(CPUHP_AP_X86_MCE_STARTING);
 
+err_hp_starting:
+	cpuhp_remove_state(CPUHP_X86_MCE_DEAD);
+
 err_init_pool:
 	mca_cfg.banks = 0;
 	kfree(mce_banks);
diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h
index 4fe2ba418471..acb37a7f7831 100644
--- a/include/linux/cpuhotplug.h
+++ b/include/linux/cpuhotplug.h
@@ -16,6 +16,7 @@ enum cpuhp_state {
 	CPUHP_PERF_SUPERH,
 	CPUHP_X86_HPET_DEAD,
 	CPUHP_X86_APB_DEAD,
+	CPUHP_X86_MCE_DEAD,
 	CPUHP_VIRT_NET_DEAD,
 	CPUHP_SLUB_DEAD,
 	CPUHP_MM_WRITEBACK_DEAD,
-- 
2.10.2

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

#1514584 — [PATCH 14/25] ia64/err-inject: Convert to hotplug state machine

Install the callbacks via the state machine and let the core invoke
the callbacks on the already online CPUs.

The removal of the files happens now in the prepare down stage as there is
no reason to keep them around until the cpu has actually died.

Cc: Tony Luck <tony.luck@intel.com>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: linux-ia64@vger.kernel.org
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
 arch/ia64/kernel/err_inject.c | 78 +++++++++++--------------------------------
 1 file changed, 19 insertions(+), 59 deletions(-)

diff --git a/arch/ia64/kernel/err_inject.c b/arch/ia64/kernel/err_inject.c
index 5ed0ea92c5bf..85bba43e7d5d 100644
--- a/arch/ia64/kernel/err_inject.c
+++ b/arch/ia64/kernel/err_inject.c
@@ -224,85 +224,45 @@ static struct attribute_group err_inject_attr_group = {
 	.name = "err_inject"
 };
 /* Add/Remove err_inject interface for CPU device */
-static int err_inject_add_dev(struct device *sys_dev)
+static int err_inject_add_dev(unsigned int cpu)
 {
+	struct device *sys_dev = get_cpu_device(cpu);
+
 	return sysfs_create_group(&sys_dev->kobj, &err_inject_attr_group);
 }
 
-static int err_inject_remove_dev(struct device *sys_dev)
+static int err_inject_remove_dev(unsigned int cpu)
 {
+	struct device *sys_dev = get_cpu_device(cpu);
+
 	sysfs_remove_group(&sys_dev->kobj, &err_inject_attr_group);
 	return 0;
 }
-static int err_inject_cpu_callback(struct notifier_block *nfb,
-		unsigned long action, void *hcpu)
+
+static enum cpuhp_state hp_online;
+
+static int __init err_inject_init(void)
 {
-	unsigned int cpu = (unsigned long)hcpu;
-	struct device *sys_dev;
-
-	sys_dev = get_cpu_device(cpu);
-	switch (action) {
-	case CPU_ONLINE:
-	case CPU_ONLINE_FROZEN:
-		err_inject_add_dev(sys_dev);
-		break;
-	case CPU_DEAD:
-	case CPU_DEAD_FROZEN:
-		err_inject_remove_dev(sys_dev);
-		break;
-	}
-
-	return NOTIFY_OK;
-}
-
-static struct notifier_block err_inject_cpu_notifier =
-{
-	.notifier_call = err_inject_cpu_callback,
-};
-
-static int __init
-err_inject_init(void)
-{
-	int i;
-
+	int ret;
 #ifdef ERR_INJ_DEBUG
 	printk(KERN_INFO "Enter error injection driver.\n");
 #endif
 
-	cpu_notifier_register_begin();
-
-	for_each_online_cpu(i) {
-		err_inject_cpu_callback(&err_inject_cpu_notifier, CPU_ONLINE,
-				(void *)(long)i);
+	ret = cpuhp_setup_state(CPUHP_AP_ONLINE_DYN, "ia64/err_inj:online",
+				err_inject_add_dev, err_inject_remove_dev);
+	if (ret >= 0) {
+		hp_online = ret;
+		ret = 0;
 	}
-
-	__register_hotcpu_notifier(&err_inject_cpu_notifier);
-
-	cpu_notifier_register_done();
-
-	return 0;
+	return ret;
 }
 
-static void __exit
-err_inject_exit(void)
+static void __exit err_inject_exit(void)
 {
-	int i;
-	struct device *sys_dev;
-
 #ifdef ERR_INJ_DEBUG
 	printk(KERN_INFO "Exit error injection driver.\n");
 #endif
-
-	cpu_notifier_register_begin();
-
-	for_each_online_cpu(i) {
-		sys_dev = get_cpu_device(i);
-		sysfs_remove_group(&sys_dev->kobj, &err_inject_attr_group);
-	}
-
-	__unregister_hotcpu_notifier(&err_inject_cpu_notifier);
-
-	cpu_notifier_register_done();
+	cpuhp_remove_state(hp_online);
 }
 
 module_init(err_inject_init);
-- 
2.10.2

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

#1518546 — [tip:smp/hotplug] ia64/err-inject: Convert to hotplug state machine

Commit-ID:  122231445f2517de1d3dd30acd28801b66d72cab
Gitweb:     http://git.kernel.org/tip/122231445f2517de1d3dd30acd28801b66d72cab
Author:     Sebastian Andrzej Siewior <bigeasy@linutronix.de>
AuthorDate: Thu, 3 Nov 2016 15:50:10 +0100
Committer:  Thomas Gleixner <tglx@linutronix.de>
CommitDate: Wed, 9 Nov 2016 23:45:29 +0100

ia64/err-inject: Convert to hotplug state machine

Install the callbacks via the state machine and let the core invoke
the callbacks on the already online CPUs.

The removal of the files happens now in the prepare down stage as there is
no reason to keep them around until the cpu has actually died.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Fenghua Yu <fenghua.yu@intel.com>
Cc: Tony Luck <tony.luck@intel.com>
Cc: linux-ia64@vger.kernel.org
Cc: rt@linutronix.de
Link: http://lkml.kernel.org/r/20161103145021.28528-15-bigeasy@linutronix.de
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

---
 arch/ia64/kernel/err_inject.c | 74 ++++++++++---------------------------------
 1 file changed, 17 insertions(+), 57 deletions(-)

diff --git a/arch/ia64/kernel/err_inject.c b/arch/ia64/kernel/err_inject.c
index 5ed0ea9..85bba43 100644
--- a/arch/ia64/kernel/err_inject.c
+++ b/arch/ia64/kernel/err_inject.c
@@ -224,85 +224,45 @@ static struct attribute_group err_inject_attr_group = {
 	.name = "err_inject"
 };
 /* Add/Remove err_inject interface for CPU device */
-static int err_inject_add_dev(struct device *sys_dev)
+static int err_inject_add_dev(unsigned int cpu)
 {
+	struct device *sys_dev = get_cpu_device(cpu);
+
 	return sysfs_create_group(&sys_dev->kobj, &err_inject_attr_group);
 }
 
-static int err_inject_remove_dev(struct device *sys_dev)
+static int err_inject_remove_dev(unsigned int cpu)
 {
+	struct device *sys_dev = get_cpu_device(cpu);
+
 	sysfs_remove_group(&sys_dev->kobj, &err_inject_attr_group);
 	return 0;
 }
-static int err_inject_cpu_callback(struct notifier_block *nfb,
-		unsigned long action, void *hcpu)
-{
-	unsigned int cpu = (unsigned long)hcpu;
-	struct device *sys_dev;
-
-	sys_dev = get_cpu_device(cpu);
-	switch (action) {
-	case CPU_ONLINE:
-	case CPU_ONLINE_FROZEN:
-		err_inject_add_dev(sys_dev);
-		break;
-	case CPU_DEAD:
-	case CPU_DEAD_FROZEN:
-		err_inject_remove_dev(sys_dev);
-		break;
-	}
-
-	return NOTIFY_OK;
-}
 
-static struct notifier_block err_inject_cpu_notifier =
-{
-	.notifier_call = err_inject_cpu_callback,
-};
+static enum cpuhp_state hp_online;
 
-static int __init
-err_inject_init(void)
+static int __init err_inject_init(void)
 {
-	int i;
-
+	int ret;
 #ifdef ERR_INJ_DEBUG
 	printk(KERN_INFO "Enter error injection driver.\n");
 #endif
 
-	cpu_notifier_register_begin();
-
-	for_each_online_cpu(i) {
-		err_inject_cpu_callback(&err_inject_cpu_notifier, CPU_ONLINE,
-				(void *)(long)i);
+	ret = cpuhp_setup_state(CPUHP_AP_ONLINE_DYN, "ia64/err_inj:online",
+				err_inject_add_dev, err_inject_remove_dev);
+	if (ret >= 0) {
+		hp_online = ret;
+		ret = 0;
 	}
-
-	__register_hotcpu_notifier(&err_inject_cpu_notifier);
-
-	cpu_notifier_register_done();
-
-	return 0;
+	return ret;
 }
 
-static void __exit
-err_inject_exit(void)
+static void __exit err_inject_exit(void)
 {
-	int i;
-	struct device *sys_dev;
-
 #ifdef ERR_INJ_DEBUG
 	printk(KERN_INFO "Exit error injection driver.\n");
 #endif
-
-	cpu_notifier_register_begin();
-
-	for_each_online_cpu(i) {
-		sys_dev = get_cpu_device(i);
-		sysfs_remove_group(&sys_dev->kobj, &err_inject_attr_group);
-	}
-
-	__unregister_hotcpu_notifier(&err_inject_cpu_notifier);
-
-	cpu_notifier_register_done();
+	cpuhp_remove_state(hp_online);
 }
 
 module_init(err_inject_init);

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

#1514585 — [PATCH 05/25] lib/radix-tree: Convert to hotplug state machine

Install the callbacks via the state machine.

Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
 include/linux/cpuhotplug.h |  1 +
 lib/radix-tree.c           | 25 ++++++++++++-------------
 2 files changed, 13 insertions(+), 13 deletions(-)

diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h
index 04e5f99ffc70..89310fb1031d 100644
--- a/include/linux/cpuhotplug.h
+++ b/include/linux/cpuhotplug.h
@@ -34,6 +34,7 @@ enum cpuhp_state {
 	CPUHP_PRINTK_DEAD,
 	CPUHP_MM_MEMCQ_DEAD,
 	CPUHP_PERCPU_CNT_DEAD,
+	CPUHP_RADIX_DEAD,
 	CPUHP_WORKQUEUE_PREP,
 	CPUHP_POWER_NUMA_PREPARE,
 	CPUHP_HRTIMERS_PREPARE,
diff --git a/lib/radix-tree.c b/lib/radix-tree.c
index 8e6d552c40dd..4b8bb3618b83 100644
--- a/lib/radix-tree.c
+++ b/lib/radix-tree.c
@@ -1642,32 +1642,31 @@ static __init void radix_tree_init_maxnodes(void)
 	}
 }
 
-static int radix_tree_callback(struct notifier_block *nfb,
-				unsigned long action, void *hcpu)
+static int radix_tree_cpu_dead(unsigned int cpu)
 {
-	int cpu = (long)hcpu;
 	struct radix_tree_preload *rtp;
 	struct radix_tree_node *node;
 
 	/* Free per-cpu pool of preloaded nodes */
-	if (action == CPU_DEAD || action == CPU_DEAD_FROZEN) {
-		rtp = &per_cpu(radix_tree_preloads, cpu);
-		while (rtp->nr) {
-			node = rtp->nodes;
-			rtp->nodes = node->private_data;
-			kmem_cache_free(radix_tree_node_cachep, node);
-			rtp->nr--;
-		}
+	rtp = &per_cpu(radix_tree_preloads, cpu);
+	while (rtp->nr) {
+		node = rtp->nodes;
+		rtp->nodes = node->private_data;
+		kmem_cache_free(radix_tree_node_cachep, node);
+		rtp->nr--;
 	}
-	return NOTIFY_OK;
+	return 0;
 }
 
 void __init radix_tree_init(void)
 {
+	int ret;
 	radix_tree_node_cachep = kmem_cache_create("radix_tree_node",
 			sizeof(struct radix_tree_node), 0,
 			SLAB_PANIC | SLAB_RECLAIM_ACCOUNT,
 			radix_tree_node_ctor);
 	radix_tree_init_maxnodes();
-	hotcpu_notifier(radix_tree_callback, 0);
+	ret = cpuhp_setup_state_nocalls(CPUHP_RADIX_DEAD, "lib/radix:dead",
+					NULL, radix_tree_cpu_dead);
+	WARN_ON(ret < 0);
 }
-- 
2.10.2

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

#1518535 — [tip:smp/hotplug] lib/radix-tree: Convert to hotplug state machine

Commit-ID:  d544abd5ff7d8b07c0c67682a63e4939c3c82914
Gitweb:     http://git.kernel.org/tip/d544abd5ff7d8b07c0c67682a63e4939c3c82914
Author:     Sebastian Andrzej Siewior <bigeasy@linutronix.de>
AuthorDate: Thu, 3 Nov 2016 15:50:01 +0100
Committer:  Thomas Gleixner <tglx@linutronix.de>
CommitDate: Wed, 9 Nov 2016 23:45:27 +0100

lib/radix-tree: Convert to hotplug state machine

Install the callbacks via the state machine.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: rt@linutronix.de
Link: http://lkml.kernel.org/r/20161103145021.28528-6-bigeasy@linutronix.de
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

---
 include/linux/cpuhotplug.h |  1 +
 lib/radix-tree.c           | 25 ++++++++++++-------------
 2 files changed, 13 insertions(+), 13 deletions(-)

diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h
index 04e5f99..89310fb 100644
--- a/include/linux/cpuhotplug.h
+++ b/include/linux/cpuhotplug.h
@@ -34,6 +34,7 @@ enum cpuhp_state {
 	CPUHP_PRINTK_DEAD,
 	CPUHP_MM_MEMCQ_DEAD,
 	CPUHP_PERCPU_CNT_DEAD,
+	CPUHP_RADIX_DEAD,
 	CPUHP_WORKQUEUE_PREP,
 	CPUHP_POWER_NUMA_PREPARE,
 	CPUHP_HRTIMERS_PREPARE,
diff --git a/lib/radix-tree.c b/lib/radix-tree.c
index 8e6d552..4b8bb36 100644
--- a/lib/radix-tree.c
+++ b/lib/radix-tree.c
@@ -1642,32 +1642,31 @@ static __init void radix_tree_init_maxnodes(void)
 	}
 }
 
-static int radix_tree_callback(struct notifier_block *nfb,
-				unsigned long action, void *hcpu)
+static int radix_tree_cpu_dead(unsigned int cpu)
 {
-	int cpu = (long)hcpu;
 	struct radix_tree_preload *rtp;
 	struct radix_tree_node *node;
 
 	/* Free per-cpu pool of preloaded nodes */
-	if (action == CPU_DEAD || action == CPU_DEAD_FROZEN) {
-		rtp = &per_cpu(radix_tree_preloads, cpu);
-		while (rtp->nr) {
-			node = rtp->nodes;
-			rtp->nodes = node->private_data;
-			kmem_cache_free(radix_tree_node_cachep, node);
-			rtp->nr--;
-		}
+	rtp = &per_cpu(radix_tree_preloads, cpu);
+	while (rtp->nr) {
+		node = rtp->nodes;
+		rtp->nodes = node->private_data;
+		kmem_cache_free(radix_tree_node_cachep, node);
+		rtp->nr--;
 	}
-	return NOTIFY_OK;
+	return 0;
 }
 
 void __init radix_tree_init(void)
 {
+	int ret;
 	radix_tree_node_cachep = kmem_cache_create("radix_tree_node",
 			sizeof(struct radix_tree_node), 0,
 			SLAB_PANIC | SLAB_RECLAIM_ACCOUNT,
 			radix_tree_node_ctor);
 	radix_tree_init_maxnodes();
-	hotcpu_notifier(radix_tree_callback, 0);
+	ret = cpuhp_setup_state_nocalls(CPUHP_RADIX_DEAD, "lib/radix:dead",
+					NULL, radix_tree_cpu_dead);
+	WARN_ON(ret < 0);
 }

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

#1514586 — [PATCH 23/25] x86/mcheck: Make CPU_DOWN_PREPARE the counter part of CPU_STARTING

The previous patch moved mcheck_cpu_init() out of identify_cpu() and put
it as the first CPU hotplug callback which is invoked on the target CPU
during bring up. It enables MCE and starts the MCE timer. If a CPU goes
down then those two things have to be reverted and this happens
currently in CPU_DOWN_PREPARE. This is not symmetrical because
CPU_DOWN_PREPARE is the counterpart of CPU_ONLINE.
Usually CPU_DOWN_FAILED and CPU_ONLINE can do the same thing but not in
this case since here (in CPU_DOWN_FAILED) it tries to revert what was
done in CPU_DOWN_PREPARE.

To make this simpler make CPU_DOWN_PREPARE the counterpart of
mcheck_cpu_starting() and just disable MCE and stop the timer. With this
change the callback is symmetrical again and we don't need CPU_DOWN_FAILED
including mce_reenable_cpu(). smp_call_function_single() can be dropped
because it is already invoked on the proper CPU and interrupts are disabled
at this point.

Cc: Tony Luck <tony.luck@intel.com>
Cc: Borislav Petkov <bp@alien8.de>
Cc: linux-edac@vger.kernel.org
Cc: x86@kernel.org
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
 arch/x86/kernel/cpu/mcheck/mce.c | 44 +++++++++++-----------------------------
 1 file changed, 12 insertions(+), 32 deletions(-)

diff --git a/arch/x86/kernel/cpu/mcheck/mce.c b/arch/x86/kernel/cpu/mcheck/mce.c
index 72af9db8526d..596a7128a46b 100644
--- a/arch/x86/kernel/cpu/mcheck/mce.c
+++ b/arch/x86/kernel/cpu/mcheck/mce.c
@@ -2469,43 +2469,22 @@ static void mce_device_remove(unsigned int cpu)
 }
 
 /* Make sure there are no machine checks on offlined CPUs. */
-static void mce_disable_cpu(void *h)
+static void mce_disable_cpu(void)
 {
-	unsigned long action = *(unsigned long *)h;
-
 	if (!mce_available(raw_cpu_ptr(&cpu_info)))
 		return;
 
-	if (!(action & CPU_TASKS_FROZEN))
+	if (!cpuhp_tasks_frozen)
 		cmci_clear();
 
 	vendor_disable_error_reporting();
 }
 
-static void mce_reenable_cpu(void *h)
-{
-	unsigned long action = *(unsigned long *)h;
-	int i;
-
-	if (!mce_available(raw_cpu_ptr(&cpu_info)))
-		return;
-
-	if (!(action & CPU_TASKS_FROZEN))
-		cmci_reenable();
-	for (i = 0; i < mca_cfg.banks; i++) {
-		struct mce_bank *b = &mce_banks[i];
-
-		if (b->init)
-			wrmsrl(msr_ops.ctl(i), b->ctl);
-	}
-}
-
 /* Get notified when a cpu comes on/off. Be hotplug friendly. */
 static int
 mce_cpu_callback(struct notifier_block *nfb, unsigned long action, void *hcpu)
 {
 	unsigned int cpu = (unsigned long)hcpu;
-	struct timer_list *t = &per_cpu(mce_timer, cpu);
 
 	switch (action & ~CPU_TASKS_FROZEN) {
 	case CPU_ONLINE:
@@ -2530,19 +2509,20 @@ mce_cpu_callback(struct notifier_block *nfb, unsigned long action, void *hcpu)
 		if (!(action & CPU_TASKS_FROZEN))
 			cmci_rediscover();
 		break;
-	case CPU_DOWN_PREPARE:
-		smp_call_function_single(cpu, mce_disable_cpu, &action, 1);
-		del_timer_sync(t);
-		break;
-	case CPU_DOWN_FAILED:
-		smp_call_function_single(cpu, mce_reenable_cpu, &action, 1);
-		mce_start_timer(cpu, t);
-		break;
 	}
 
 	return NOTIFY_OK;
 }
 
+static int mce_cpu_down_dying(unsigned int cpu)
+{
+	struct timer_list *t = this_cpu_ptr(&mce_timer);
+
+	mce_disable_cpu();
+	del_timer_sync(t);
+	return 0;
+}
+
 static struct notifier_block mce_cpu_notifier = {
 	.notifier_call = mce_cpu_callback,
 };
@@ -2597,7 +2577,7 @@ static __init int mcheck_init_device(void)
 		goto err_init_pool;
 
 	err = cpuhp_setup_state(CPUHP_AP_X86_MCE_STARTING, "x86/mce:starting",
-				mcheck_cpu_starting, NULL);
+				mcheck_cpu_starting, mce_cpu_down_dying);
 	if (err)
 		goto err_init_pool;
 
-- 
2.10.2

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

#1514587 — [PATCH 02/25] kernel/printk: Convert to hotplug state machine

Install the callbacks via the state machine.

Cc: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
 include/linux/cpuhotplug.h |  1 +
 kernel/printk/printk.c     | 27 ++++++++++++---------------
 2 files changed, 13 insertions(+), 15 deletions(-)

diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h
index 69b74fa0da60..4174083280d7 100644
--- a/include/linux/cpuhotplug.h
+++ b/include/linux/cpuhotplug.h
@@ -31,6 +31,7 @@ enum cpuhp_state {
 	CPUHP_S390_PFAULT_DEAD,
 	CPUHP_BLK_MQ_DEAD,
 	CPUHP_FS_BUFF_DEAD,
+	CPUHP_PRINTK_DEAD,
 	CPUHP_WORKQUEUE_PREP,
 	CPUHP_POWER_NUMA_PREPARE,
 	CPUHP_HRTIMERS_PREPARE,
diff --git a/kernel/printk/printk.c b/kernel/printk/printk.c
index de08fc90baaf..4487ffcd42d5 100644
--- a/kernel/printk/printk.c
+++ b/kernel/printk/printk.c
@@ -2185,27 +2185,18 @@ void resume_console(void)
 
 /**
  * console_cpu_notify - print deferred console messages after CPU hotplug
- * @self: notifier struct
- * @action: CPU hotplug event
- * @hcpu: unused
+ * @cpu: unused
  *
  * If printk() is called from a CPU that is not online yet, the messages
  * will be spooled but will not show up on the console.  This function is
  * called when a new CPU comes online (or fails to come up), and ensures
  * that any such output gets printed.
  */
-static int console_cpu_notify(struct notifier_block *self,
-	unsigned long action, void *hcpu)
+static int console_cpu_notify(unsigned int cpu)
 {
-	switch (action) {
-	case CPU_ONLINE:
-	case CPU_DEAD:
-	case CPU_DOWN_FAILED:
-	case CPU_UP_CANCELED:
-		console_lock();
-		console_unlock();
-	}
-	return NOTIFY_OK;
+	console_lock();
+	console_unlock();
+	return 0;
 }
 
 /**
@@ -2843,6 +2834,7 @@ EXPORT_SYMBOL(unregister_console);
 static int __init printk_late_init(void)
 {
 	struct console *con;
+	int ret;
 
 	for_each_console(con) {
 		if (!keep_bootcon && con->flags & CON_BOOT) {
@@ -2857,7 +2849,12 @@ static int __init printk_late_init(void)
 				unregister_console(con);
 		}
 	}
-	hotcpu_notifier(console_cpu_notify, 0);
+	ret = cpuhp_setup_state_nocalls(CPUHP_PRINTK_DEAD, "printk:dead", NULL,
+					console_cpu_notify);
+	WARN_ON(ret < 0);
+	ret = cpuhp_setup_state_nocalls(CPUHP_AP_ONLINE_DYN, "printk:online",
+					console_cpu_notify, NULL);
+	WARN_ON(ret < 0);
 	return 0;
 }
 late_initcall(printk_late_init);
-- 
2.10.2

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

#1518534 — [tip:smp/hotplug] kernel/printk: Convert to hotplug state machine

Commit-ID:  90b14889d2f9b29d7e5b4b2d36251c13ce3dd13f
Gitweb:     http://git.kernel.org/tip/90b14889d2f9b29d7e5b4b2d36251c13ce3dd13f
Author:     Sebastian Andrzej Siewior <bigeasy@linutronix.de>
AuthorDate: Thu, 3 Nov 2016 15:49:58 +0100
Committer:  Thomas Gleixner <tglx@linutronix.de>
CommitDate: Wed, 9 Nov 2016 23:45:26 +0100

kernel/printk: Convert to hotplug state machine

Install the callbacks via the state machine.

Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
Cc: Andrew Morton <akpm@linux-foundation.org>
Cc: rt@linutronix.de
Link: http://lkml.kernel.org/r/20161103145021.28528-3-bigeasy@linutronix.de
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>

---
 include/linux/cpuhotplug.h |  1 +
 kernel/printk/printk.c     | 29 +++++++++++++----------------
 2 files changed, 14 insertions(+), 16 deletions(-)

diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h
index 69b74fa..4174083 100644
--- a/include/linux/cpuhotplug.h
+++ b/include/linux/cpuhotplug.h
@@ -31,6 +31,7 @@ enum cpuhp_state {
 	CPUHP_S390_PFAULT_DEAD,
 	CPUHP_BLK_MQ_DEAD,
 	CPUHP_FS_BUFF_DEAD,
+	CPUHP_PRINTK_DEAD,
 	CPUHP_WORKQUEUE_PREP,
 	CPUHP_POWER_NUMA_PREPARE,
 	CPUHP_HRTIMERS_PREPARE,
diff --git a/kernel/printk/printk.c b/kernel/printk/printk.c
index de08fc9..4487ffc 100644
--- a/kernel/printk/printk.c
+++ b/kernel/printk/printk.c
@@ -2185,27 +2185,18 @@ void resume_console(void)
 
 /**
  * console_cpu_notify - print deferred console messages after CPU hotplug
- * @self: notifier struct
- * @action: CPU hotplug event
- * @hcpu: unused
+ * @cpu: unused
  *
  * If printk() is called from a CPU that is not online yet, the messages
  * will be spooled but will not show up on the console.  This function is
  * called when a new CPU comes online (or fails to come up), and ensures
  * that any such output gets printed.
  */
-static int console_cpu_notify(struct notifier_block *self,
-	unsigned long action, void *hcpu)
-{
-	switch (action) {
-	case CPU_ONLINE:
-	case CPU_DEAD:
-	case CPU_DOWN_FAILED:
-	case CPU_UP_CANCELED:
-		console_lock();
-		console_unlock();
-	}
-	return NOTIFY_OK;
+static int console_cpu_notify(unsigned int cpu)
+{
+	console_lock();
+	console_unlock();
+	return 0;
 }
 
 /**
@@ -2843,6 +2834,7 @@ EXPORT_SYMBOL(unregister_console);
 static int __init printk_late_init(void)
 {
 	struct console *con;
+	int ret;
 
 	for_each_console(con) {
 		if (!keep_bootcon && con->flags & CON_BOOT) {
@@ -2857,7 +2849,12 @@ static int __init printk_late_init(void)
 				unregister_console(con);
 		}
 	}
-	hotcpu_notifier(console_cpu_notify, 0);
+	ret = cpuhp_setup_state_nocalls(CPUHP_PRINTK_DEAD, "printk:dead", NULL,
+					console_cpu_notify);
+	WARN_ON(ret < 0);
+	ret = cpuhp_setup_state_nocalls(CPUHP_AP_ONLINE_DYN, "printk:online",
+					console_cpu_notify, NULL);
+	WARN_ON(ret < 0);
 	return 0;
 }
 late_initcall(printk_late_init);

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

#1514588 — [PATCH 06/25] mm/page_alloc: Convert to hotplug state machine

Install the callbacks via the state machine.

Cc: linux-mm@kvack.org
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
 include/linux/cpuhotplug.h |  1 +
 mm/page_alloc.c            | 49 +++++++++++++++++++++++-----------------------
 2 files changed, 26 insertions(+), 24 deletions(-)

diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h
index 89310fb1031d..31c58f6ec3c6 100644
--- a/include/linux/cpuhotplug.h
+++ b/include/linux/cpuhotplug.h
@@ -35,6 +35,7 @@ enum cpuhp_state {
 	CPUHP_MM_MEMCQ_DEAD,
 	CPUHP_PERCPU_CNT_DEAD,
 	CPUHP_RADIX_DEAD,
+	CPUHP_PAGE_ALLOC_DEAD,
 	CPUHP_WORKQUEUE_PREP,
 	CPUHP_POWER_NUMA_PREPARE,
 	CPUHP_HRTIMERS_PREPARE,
diff --git a/mm/page_alloc.c b/mm/page_alloc.c
index 8fd42aa7c4bd..68873a164cc0 100644
--- a/mm/page_alloc.c
+++ b/mm/page_alloc.c
@@ -6491,38 +6491,39 @@ void __init free_area_init(unsigned long *zones_size)
 			__pa(PAGE_OFFSET) >> PAGE_SHIFT, NULL);
 }
 
-static int page_alloc_cpu_notify(struct notifier_block *self,
-				 unsigned long action, void *hcpu)
+static int page_alloc_cpu_dead(unsigned int cpu)
 {
-	int cpu = (unsigned long)hcpu;
 
-	if (action == CPU_DEAD || action == CPU_DEAD_FROZEN) {
-		lru_add_drain_cpu(cpu);
-		drain_pages(cpu);
+	lru_add_drain_cpu(cpu);
+	drain_pages(cpu);
 
-		/*
-		 * Spill the event counters of the dead processor
-		 * into the current processors event counters.
-		 * This artificially elevates the count of the current
-		 * processor.
-		 */
-		vm_events_fold_cpu(cpu);
+	/*
+	 * Spill the event counters of the dead processor
+	 * into the current processors event counters.
+	 * This artificially elevates the count of the current
+	 * processor.
+	 */
+	vm_events_fold_cpu(cpu);
 
-		/*
-		 * Zero the differential counters of the dead processor
-		 * so that the vm statistics are consistent.
-		 *
-		 * This is only okay since the processor is dead and cannot
-		 * race with what we are doing.
-		 */
-		cpu_vm_stats_fold(cpu);
-	}
-	return NOTIFY_OK;
+	/*
+	 * Zero the differential counters of the dead processor
+	 * so that the vm statistics are consistent.
+	 *
+	 * This is only okay since the processor is dead and cannot
+	 * race with what we are doing.
+	 */
+	cpu_vm_stats_fold(cpu);
+	return 0;
 }
 
 void __init page_alloc_init(void)
 {
-	hotcpu_notifier(page_alloc_cpu_notify, 0);
+	int ret;
+
+	ret = cpuhp_setup_state_nocalls(CPUHP_PAGE_ALLOC_DEAD,
+					"mm/page_alloc:dead", NULL,
+					page_alloc_cpu_dead);
+	WARN_ON(ret < 0);
 }
 
 /*
-- 
2.10.2

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

Page 1 of 2  [1] 2  Next page →

Back to top | Article view | linux.kernel

csiph-web