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

Groups > linux.kernel > #1477648 > unrolled thread

cpu hotplug: convert more drivers (batch #3)

Back to article view | Back to linux.kernel

Contents

  cpu hotplug: convert more drivers (batch #3) Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-06 19:10 +0200
    [PATCH 20/21] blk: mq: cpu-notif: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-06 19:10 +0200
    [PATCH 11/21] ACPI: processor: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-06 19:10 +0200
      Re: [PATCH 11/21] ACPI: processor: Convert to hotplug state machine "Rafael J. Wysocki" <rafael@kernel.org> - 2016-09-06 23:10 +0200
        Re: [PATCH 11/21] ACPI: processor: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-07 16:10 +0200
          Re: [PATCH 11/21] ACPI: processor: Convert to hotplug state machine "Rafael J. Wysocki" <rjw@rjwysocki.net> - 2016-09-07 18:00 +0200
    [PATCH 21/21] blk: mq: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-06 19:10 +0200
    [PATCH 03/21] ARM: OMAP: wakeupgen: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-06 19:10 +0200
    [PATCH 15/21] mips: octeon: smp: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-06 19:10 +0200
      Re: [PATCH 15/21] mips: octeon: smp: Convert to hotplug state machine Matt Redfearn <matt.redfearn@imgtec.com> - 2016-09-07 10:30 +0200
        Re: [PATCH 15/21] mips: octeon: smp: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-07 16:30 +0200
          Re: [PATCH 15/21] mips: octeon: smp: Convert to hotplug state machine Matt Redfearn <matt.redfearn@imgtec.com> - 2016-09-08 10:40 +0200
    [PATCH 04/21] ia64: mca: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-06 19:10 +0200
    [PATCH 08/21] block: softirq: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-06 19:10 +0200
    [PATCH 06/21] x86: microcode: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-06 19:10 +0200
      Re: [PATCH 06/21] x86: microcode: Convert to hotplug state machine Borislav Petkov <bp@alien8.de> - 2016-09-07 13:40 +0200
        Re: [PATCH 06/21] x86: microcode: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-07 17:00 +0200
          Re: [PATCH 06/21] x86: microcode: Convert to hotplug state machine Borislav Petkov <bp@alien8.de> - 2016-09-07 18:10 +0200
        [PATCH 06/21 v2] x86: microcode: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-07 18:50 +0200
    [PATCH 14/21] fault-injection: cpu: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-06 19:10 +0200
    [PATCH 19/21] blk: mq: reserve hotplug ID states for block Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-06 19:10 +0200
    [PATCH 13/21] padata: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-06 19:10 +0200
    [PATCH 17/21] s390: mm: pfault: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-06 19:10 +0200
    [PATCH 02/21] ARM: shmobile: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-06 19:10 +0200
      Re: [PATCH 02/21] ARM: shmobile: Convert to hotplug state machine Geert Uytterhoeven <geert@linux-m68k.org> - 2016-09-06 20:10 +0200
        Re: [PATCH 02/21] ARM: shmobile: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-07 16:00 +0200
    [PATCH 16/21] mips: loongson: smp: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-06 19:10 +0200
    [PATCH 12/21] cpufreq: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-06 19:10 +0200
      Re: [PATCH 12/21] cpufreq: Convert to hotplug state machine "Rafael J. Wysocki" <rafael@kernel.org> - 2016-09-06 23:30 +0200
        Re: [PATCH 12/21] cpufreq: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-07 16:20 +0200
          Re: [PATCH 12/21] cpufreq: Convert to hotplug state machine "Rafael J. Wysocki" <rjw@rjwysocki.net> - 2016-09-07 18:00 +0200
    [PATCH 07/21] lib: irq_poll: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-06 19:10 +0200
    [PATCH 01/21] arm64: FP/SIMD: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-06 19:10 +0200
      Re: [PATCH 01/21] arm64: FP/SIMD: Convert to hotplug state machine Will Deacon <will.deacon@arm.com> - 2016-09-06 20:20 +0200
    [PATCH 05/21] sh: SH-X3 SMP: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-06 19:10 +0200
    [PATCH 10/21] virtio scsi: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-06 19:10 +0200
    [PATCH 09/21] oprofile: timer: Convert to hotplug state machine Sebastian Andrzej Siewior <bigeasy@linutronix.de> - 2016-09-06 19:20 +0200

Page 2 of 2 — ← Prev page 1 [2]

#1477658 — [PATCH 19/21] blk: mq: reserve hotplug ID states for block

This patch only reseves two CPU hotplug ids for block/mq to avoid clashes
during merge.

Cc: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
 include/linux/cpuhotplug.h | 2 ++
 1 file changed, 2 insertions(+)

diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h
index eb0298440206..fceebbe77dd5 100644
--- a/include/linux/cpuhotplug.h
+++ b/include/linux/cpuhotplug.h
@@ -27,6 +27,7 @@ enum cpuhp_state {
 	CPUHP_VIRT_SCSI_DEAD,
 	CPUHP_ACPI_CPUDRV_DEAD,
 	CPUHP_S390_PFAULT_DEAD,
+	CPUHP_BLKMQ_DEAD,
 	CPUHP_WORKQUEUE_PREP,
 	CPUHP_POWER_NUMA_PREPARE,
 	CPUHP_HRTIMERS_PREPARE,
@@ -47,6 +48,7 @@ enum cpuhp_state {
 	CPUHP_NOTF_ERR_INJ_PREPARE,
 	CPUHP_MIPS_CAVIUM_PREPARE,
 	CPUHP_MIPS_LOONGSON_PREPARE,
+	CPUHP_BLK_MQ_PREPARE,
 	CPUHP_TIMERS_DEAD,
 	CPUHP_BRINGUP_CPU,
 	CPUHP_AP_IDLE_DEAD,
-- 
2.9.3

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

#1477663 — [PATCH 13/21] padata: Convert to hotplug state machine

Install the callbacks via the state machine. CPU-hotplug multinstance support
is used with the nocalls() version. Maybe parts of padata_alloc() could be
moved into the online callback so that we could invoke ->startup callback for
instance and drop get_online_cpus().

Cc: Steffen Klassert <steffen.klassert@secunet.com>
Cc: linux-crypto@vger.kernel.org
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
 include/linux/padata.h |  2 +-
 kernel/padata.c        | 92 ++++++++++++++++++++++++++++----------------------
 2 files changed, 53 insertions(+), 41 deletions(-)

diff --git a/include/linux/padata.h b/include/linux/padata.h
index 113ee626a4dc..0f9e567d5e15 100644
--- a/include/linux/padata.h
+++ b/include/linux/padata.h
@@ -151,7 +151,7 @@ struct parallel_data {
  * @flags: padata flags.
  */
 struct padata_instance {
-	struct notifier_block		 cpu_notifier;
+	struct hlist_node		 node;
 	struct workqueue_struct		*wq;
 	struct parallel_data		*pd;
 	struct padata_cpumask		cpumask;
diff --git a/kernel/padata.c b/kernel/padata.c
index 993278895ccc..7848f0566403 100644
--- a/kernel/padata.c
+++ b/kernel/padata.c
@@ -30,6 +30,7 @@
 #include <linux/slab.h>
 #include <linux/sysfs.h>
 #include <linux/rcupdate.h>
+#include <linux/module.h>
 
 #define MAX_OBJ_NUM 1000
 
@@ -769,52 +770,43 @@ static inline int pinst_has_cpu(struct padata_instance *pinst, int cpu)
 		cpumask_test_cpu(cpu, pinst->cpumask.cbcpu);
 }
 
-
-static int padata_cpu_callback(struct notifier_block *nfb,
-			       unsigned long action, void *hcpu)
+static int padata_cpu_online(unsigned int cpu, struct hlist_node *node)
 {
-	int err;
 	struct padata_instance *pinst;
-	int cpu = (unsigned long)hcpu;
+	int ret;
 
-	pinst = container_of(nfb, struct padata_instance, cpu_notifier);
+	pinst = hlist_entry_safe(node, struct padata_instance, node);
+	if (!pinst_has_cpu(pinst, cpu))
+		return 0;
 
-	switch (action) {
-	case CPU_ONLINE:
-	case CPU_ONLINE_FROZEN:
-	case CPU_DOWN_FAILED:
-	case CPU_DOWN_FAILED_FROZEN:
-		if (!pinst_has_cpu(pinst, cpu))
-			break;
-		mutex_lock(&pinst->lock);
-		err = __padata_add_cpu(pinst, cpu);
-		mutex_unlock(&pinst->lock);
-		if (err)
-			return notifier_from_errno(err);
-		break;
-
-	case CPU_DOWN_PREPARE:
-	case CPU_DOWN_PREPARE_FROZEN:
-	case CPU_UP_CANCELED:
-	case CPU_UP_CANCELED_FROZEN:
-		if (!pinst_has_cpu(pinst, cpu))
-			break;
-		mutex_lock(&pinst->lock);
-		err = __padata_remove_cpu(pinst, cpu);
-		mutex_unlock(&pinst->lock);
-		if (err)
-			return notifier_from_errno(err);
-		break;
-	}
-
-	return NOTIFY_OK;
+	mutex_lock(&pinst->lock);
+	ret = __padata_add_cpu(pinst, cpu);
+	mutex_unlock(&pinst->lock);
+	return ret;
 }
+
+static int padata_cpu_prep_down(unsigned int cpu, struct hlist_node *node)
+{
+	struct padata_instance *pinst;
+	int ret;
+
+	pinst = hlist_entry_safe(node, struct padata_instance, node);
+	if (!pinst_has_cpu(pinst, cpu))
+		return 0;
+
+	mutex_lock(&pinst->lock);
+	ret = __padata_remove_cpu(pinst, cpu);
+	mutex_unlock(&pinst->lock);
+	return ret;
+}
+
+static enum cpuhp_state hp_online;
 #endif
 
 static void __padata_free(struct padata_instance *pinst)
 {
 #ifdef CONFIG_HOTPLUG_CPU
-	unregister_hotcpu_notifier(&pinst->cpu_notifier);
+	cpuhp_state_remove_instance_nocalls(hp_online, &pinst->node);
 #endif
 
 	padata_stop(pinst);
@@ -1012,11 +1004,8 @@ struct padata_instance *padata_alloc(struct workqueue_struct *wq,
 	mutex_init(&pinst->lock);
 
 #ifdef CONFIG_HOTPLUG_CPU
-	pinst->cpu_notifier.notifier_call = padata_cpu_callback;
-	pinst->cpu_notifier.priority = 0;
-	register_hotcpu_notifier(&pinst->cpu_notifier);
+	cpuhp_state_add_instance_nocalls(hp_online, &pinst->node);
 #endif
-
 	return pinst;
 
 err_free_masks:
@@ -1039,3 +1028,26 @@ void padata_free(struct padata_instance *pinst)
 	kobject_put(&pinst->kobj);
 }
 EXPORT_SYMBOL(padata_free);
+
+#ifdef CONFIG_HOTPLUG_CPU
+
+static __init int padata_driver_init(void)
+{
+	int ret;
+
+	ret = cpuhp_setup_state_multi(CPUHP_AP_ONLINE_DYN, "padata:online",
+				      padata_cpu_online,
+				      padata_cpu_prep_down);
+	if (ret < 0)
+		return ret;
+	hp_online = ret;
+	return 0;
+}
+module_init(padata_driver_init);
+
+static __exit void padata_driver_exit(void)
+{
+	cpuhp_remove_multi_state(hp_online);
+}
+module_exit(padata_driver_exit);
+#endif
-- 
2.9.3

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

#1477664 — [PATCH 17/21] s390: mm: pfault: Convert to hotplug state machine

Install the callbacks via the state machine.

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>
---
 arch/s390/mm/fault.c       | 30 ++++++++++++------------------
 include/linux/cpuhotplug.h |  1 +
 2 files changed, 13 insertions(+), 18 deletions(-)

diff --git a/arch/s390/mm/fault.c b/arch/s390/mm/fault.c
index a58bca62a93b..cbb73fabc91e 100644
--- a/arch/s390/mm/fault.c
+++ b/arch/s390/mm/fault.c
@@ -740,28 +740,21 @@ static void pfault_interrupt(struct ext_code ext_code,
 	put_task_struct(tsk);
 }
 
-static int pfault_cpu_notify(struct notifier_block *self, unsigned long action,
-			     void *hcpu)
+static int pfault_cpu_dead(unsigned int cpu)
 {
 	struct thread_struct *thread, *next;
 	struct task_struct *tsk;
 
-	switch (action & ~CPU_TASKS_FROZEN) {
-	case CPU_DEAD:
-		spin_lock_irq(&pfault_lock);
-		list_for_each_entry_safe(thread, next, &pfault_list, list) {
-			thread->pfault_wait = 0;
-			list_del(&thread->list);
-			tsk = container_of(thread, struct task_struct, thread);
-			wake_up_process(tsk);
-			put_task_struct(tsk);
-		}
-		spin_unlock_irq(&pfault_lock);
-		break;
-	default:
-		break;
+	spin_lock_irq(&pfault_lock);
+	list_for_each_entry_safe(thread, next, &pfault_list, list) {
+		thread->pfault_wait = 0;
+		list_del(&thread->list);
+		tsk = container_of(thread, struct task_struct, thread);
+		wake_up_process(tsk);
+		put_task_struct(tsk);
 	}
-	return NOTIFY_OK;
+	spin_unlock_irq(&pfault_lock);
+	return 0;
 }
 
 static int __init pfault_irq_init(void)
@@ -775,7 +768,8 @@ static int __init pfault_irq_init(void)
 	if (rc)
 		goto out_pfault;
 	irq_subclass_register(IRQ_SUBCLASS_SERVICE_SIGNAL);
-	hotcpu_notifier(pfault_cpu_notify, 0);
+	cpuhp_setup_state_nocalls(CPUHP_S390_PFAULT_DEAD, "s390/pfault:dead",
+				  NULL, pfault_cpu_dead);
 	return 0;
 
 out_pfault:
diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h
index 9e56be3eaf3f..eb0298440206 100644
--- a/include/linux/cpuhotplug.h
+++ b/include/linux/cpuhotplug.h
@@ -26,6 +26,7 @@ enum cpuhp_state {
 	CPUHP_BLOCK_SOFTIRQ_DEAD,
 	CPUHP_VIRT_SCSI_DEAD,
 	CPUHP_ACPI_CPUDRV_DEAD,
+	CPUHP_S390_PFAULT_DEAD,
 	CPUHP_WORKQUEUE_PREP,
 	CPUHP_POWER_NUMA_PREPARE,
 	CPUHP_HRTIMERS_PREPARE,
-- 
2.9.3

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

#1477665 — [PATCH 02/21] ARM: shmobile: Convert to hotplug state machine

Install the callbacks via the state machine.

Cc: Simon Horman <horms@verge.net.au>
Cc: Magnus Damm <magnus.damm@gmail.com>
Cc: linux-arm-kernel@lists.infradead.org
Cc: linux-sh@vger.kernel.org
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
 arch/arm/mach-shmobile/platsmp-scu.c | 26 ++++++++------------------
 include/linux/cpuhotplug.h           |  1 +
 2 files changed, 9 insertions(+), 18 deletions(-)

diff --git a/arch/arm/mach-shmobile/platsmp-scu.c b/arch/arm/mach-shmobile/platsmp-scu.c
index 8d478f1da265..d1ecaf37d142 100644
--- a/arch/arm/mach-shmobile/platsmp-scu.c
+++ b/arch/arm/mach-shmobile/platsmp-scu.c
@@ -21,26 +21,14 @@
 static phys_addr_t shmobile_scu_base_phys;
 static void __iomem *shmobile_scu_base;
 
-static int shmobile_smp_scu_notifier_call(struct notifier_block *nfb,
-					  unsigned long action, void *hcpu)
+static int shmobile_scu_cpu_prepare(unsigned int cpu)
 {
-	unsigned int cpu = (long)hcpu;
-
-	switch (action) {
-	case CPU_UP_PREPARE:
-		/* For this particular CPU register SCU SMP boot vector */
-		shmobile_smp_hook(cpu, virt_to_phys(shmobile_boot_scu),
-				  shmobile_scu_base_phys);
-		break;
-	};
-
-	return NOTIFY_OK;
+	/* For this particular CPU register SCU SMP boot vector */
+	shmobile_smp_hook(cpu, virt_to_phys(shmobile_boot_scu),
+			  shmobile_scu_base_phys);
+	return 0;
 }
 
-static struct notifier_block shmobile_smp_scu_notifier = {
-	.notifier_call = shmobile_smp_scu_notifier_call,
-};
-
 void __init shmobile_smp_scu_prepare_cpus(phys_addr_t scu_base_phys,
 					  unsigned int max_cpus)
 {
@@ -54,7 +42,9 @@ void __init shmobile_smp_scu_prepare_cpus(phys_addr_t scu_base_phys,
 	scu_power_mode(shmobile_scu_base, SCU_PM_NORMAL);
 
 	/* Use CPU notifier for reset vector control */
-	register_cpu_notifier(&shmobile_smp_scu_notifier);
+	cpuhp_setup_state_nocalls(CPUHP_ARM_SHMOBILE_SCU_PREPARE,
+				  "arm/shmobile-scu:prepare",
+				  shmobile_scu_cpu_prepare, NULL);
 }
 
 #ifdef CONFIG_HOTPLUG_CPU
diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h
index 0da071ff36d2..008eed0c0787 100644
--- a/include/linux/cpuhotplug.h
+++ b/include/linux/cpuhotplug.h
@@ -35,6 +35,7 @@ enum cpuhp_state {
 	CPUHP_POWERPC_PMAC_PREPARE,
 	CPUHP_POWERPC_MMU_CTX_PREPARE,
 	CPUHP_NOTIFY_PREPARE,
+	CPUHP_ARM_SHMOBILE_SCU_PREPARE,
 	CPUHP_TIMERS_DEAD,
 	CPUHP_BRINGUP_CPU,
 	CPUHP_AP_IDLE_DEAD,
-- 
2.9.3

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

#1477715 — Re: [PATCH 02/21] ARM: shmobile: Convert to hotplug state machine

Hi Sebastian,

CC linux-renesas-soc

On Tue, Sep 6, 2016 at 7:04 PM, Sebastian Andrzej Siewior
<bigeasy@linutronix.de> wrote:
> Install the callbacks via the state machine.

Please describe why this is desirable.

> Cc: Simon Horman <horms@verge.net.au>
> Cc: Magnus Damm <magnus.damm@gmail.com>
> Cc: linux-arm-kernel@lists.infradead.org
> Cc: linux-sh@vger.kernel.org
> Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
> ---
>  arch/arm/mach-shmobile/platsmp-scu.c | 26 ++++++++------------------
>  include/linux/cpuhotplug.h           |  1 +
>  2 files changed, 9 insertions(+), 18 deletions(-)
>
> diff --git a/arch/arm/mach-shmobile/platsmp-scu.c b/arch/arm/mach-shmobile/platsmp-scu.c
> index 8d478f1da265..d1ecaf37d142 100644
> --- a/arch/arm/mach-shmobile/platsmp-scu.c
> +++ b/arch/arm/mach-shmobile/platsmp-scu.c
> @@ -21,26 +21,14 @@
>  static phys_addr_t shmobile_scu_base_phys;
>  static void __iomem *shmobile_scu_base;
>
> -static int shmobile_smp_scu_notifier_call(struct notifier_block *nfb,
> -                                         unsigned long action, void *hcpu)
> +static int shmobile_scu_cpu_prepare(unsigned int cpu)
>  {
> -       unsigned int cpu = (long)hcpu;
> -
> -       switch (action) {
> -       case CPU_UP_PREPARE:
> -               /* For this particular CPU register SCU SMP boot vector */
> -               shmobile_smp_hook(cpu, virt_to_phys(shmobile_boot_scu),
> -                                 shmobile_scu_base_phys);
> -               break;
> -       };
> -
> -       return NOTIFY_OK;
> +       /* For this particular CPU register SCU SMP boot vector */
> +       shmobile_smp_hook(cpu, virt_to_phys(shmobile_boot_scu),
> +                         shmobile_scu_base_phys);
> +       return 0;
>  }
>
> -static struct notifier_block shmobile_smp_scu_notifier = {
> -       .notifier_call = shmobile_smp_scu_notifier_call,
> -};
> -
>  void __init shmobile_smp_scu_prepare_cpus(phys_addr_t scu_base_phys,
>                                           unsigned int max_cpus)
>  {
> @@ -54,7 +42,9 @@ void __init shmobile_smp_scu_prepare_cpus(phys_addr_t scu_base_phys,
>         scu_power_mode(shmobile_scu_base, SCU_PM_NORMAL);
>
>         /* Use CPU notifier for reset vector control */
> -       register_cpu_notifier(&shmobile_smp_scu_notifier);
> +       cpuhp_setup_state_nocalls(CPUHP_ARM_SHMOBILE_SCU_PREPARE,
> +                                 "arm/shmobile-scu:prepare",
> +                                 shmobile_scu_cpu_prepare, NULL);
>  }
>
>  #ifdef CONFIG_HOTPLUG_CPU
> diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h
> index 0da071ff36d2..008eed0c0787 100644
> --- a/include/linux/cpuhotplug.h
> +++ b/include/linux/cpuhotplug.h
> @@ -35,6 +35,7 @@ enum cpuhp_state {
>         CPUHP_POWERPC_PMAC_PREPARE,
>         CPUHP_POWERPC_MMU_CTX_PREPARE,
>         CPUHP_NOTIFY_PREPARE,
> +       CPUHP_ARM_SHMOBILE_SCU_PREPARE,
>         CPUHP_TIMERS_DEAD,
>         CPUHP_BRINGUP_CPU,
>         CPUHP_AP_IDLE_DEAD,

What's the rationale behind adding all these numbers and always
iterating over all
of them, even though most/all of them cannot be used at the same time
(e.g. CPUHP_SH_SH3X_PREPARE is for SuperH, while
 CPUHP_ARM_SHMOBILE_SCU_PREPARE is for ARM)?

Gr{oetje,eeting}s,

                        Geert

--
Geert Uytterhoeven -- There's lots of Linux beyond ia32 -- geert@linux-m68k.org

In personal conversations with technical people, I call myself a hacker. But
when I'm talking to journalists I just say "programmer" or something like that.
                                -- Linus Torvalds

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

#1478321 — Re: [PATCH 02/21] ARM: shmobile: Convert to hotplug state machine

On 2016-09-06 20:05:37 [+0200], Geert Uytterhoeven wrote:
> Hi Sebastian,
Hi Geert,

> Please describe why this is desirable.

We have now an old and new infrastructure in place to handle hotplug
notifications while a CPU goes up and/or down. We want to get rid of the
old infrastructure and are converting its users away. 

> > diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h
> > index 0da071ff36d2..008eed0c0787 100644
> > --- a/include/linux/cpuhotplug.h
> > +++ b/include/linux/cpuhotplug.h
> > @@ -35,6 +35,7 @@ enum cpuhp_state {
> >         CPUHP_POWERPC_PMAC_PREPARE,
> >         CPUHP_POWERPC_MMU_CTX_PREPARE,
> >         CPUHP_NOTIFY_PREPARE,
> > +       CPUHP_ARM_SHMOBILE_SCU_PREPARE,
> >         CPUHP_TIMERS_DEAD,
> >         CPUHP_BRINGUP_CPU,
> >         CPUHP_AP_IDLE_DEAD,
> 
> What's the rationale behind adding all these numbers and always
> iterating over all
> of them, even though most/all of them cannot be used at the same time
> (e.g. CPUHP_SH_SH3X_PREPARE is for SuperH, while
>  CPUHP_ARM_SHMOBILE_SCU_PREPARE is for ARM)?

PREPARE + STARTING states require static ids. The online callback has
dynamic allocations of ids since it was assumed that this is the most
common one.
We will most likely evaluate the situation once we done and total number
of IDS are not acceptable.

> 
> Gr{oetje,eeting}s,
> 
>                         Geert
> 

Sebastian

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

#1477666 — [PATCH 16/21] mips: loongson: smp: Convert to hotplug state machine

Install the callbacks via the state machine.

Cc: Ralf Baechle <ralf@linux-mips.org>
Cc: linux-mips@linux-mips.org
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
 arch/mips/loongson64/loongson-3/smp.c | 34 ++++++++--------------------------
 include/linux/cpuhotplug.h            |  1 +
 2 files changed, 9 insertions(+), 26 deletions(-)

diff --git a/arch/mips/loongson64/loongson-3/smp.c b/arch/mips/loongson64/loongson-3/smp.c
index 2fec6f753a35..bd7c8cf5332d 100644
--- a/arch/mips/loongson64/loongson-3/smp.c
+++ b/arch/mips/loongson64/loongson-3/smp.c
@@ -677,7 +677,7 @@ void play_dead(void)
 	play_dead_at_ckseg1(state_addr);
 }
 
-void loongson3_disable_clock(int cpu)
+static int loongson3_disable_clock(unsigned int cpu)
 {
 	uint64_t core_id = cpu_data[cpu].core;
 	uint64_t package_id = cpu_data[cpu].package;
@@ -688,9 +688,10 @@ void loongson3_disable_clock(int cpu)
 		if (!(loongson_sysconf.workarounds & WORKAROUND_CPUHOTPLUG))
 			LOONGSON_FREQCTRL(package_id) &= ~(1 << (core_id * 4 + 3));
 	}
+	return 0;
 }
 
-void loongson3_enable_clock(int cpu)
+static int loongson3_enable_clock(unsigned int cpu)
 {
 	uint64_t core_id = cpu_data[cpu].core;
 	uint64_t package_id = cpu_data[cpu].package;
@@ -701,34 +702,15 @@ void loongson3_enable_clock(int cpu)
 		if (!(loongson_sysconf.workarounds & WORKAROUND_CPUHOTPLUG))
 			LOONGSON_FREQCTRL(package_id) |= 1 << (core_id * 4 + 3);
 	}
-}
-
-#define CPU_POST_DEAD_FROZEN	(CPU_POST_DEAD | CPU_TASKS_FROZEN)
-static int loongson3_cpu_callback(struct notifier_block *nfb,
-	unsigned long action, void *hcpu)
-{
-	unsigned int cpu = (unsigned long)hcpu;
-
-	switch (action) {
-	case CPU_POST_DEAD:
-	case CPU_POST_DEAD_FROZEN:
-		pr_info("Disable clock for CPU#%d\n", cpu);
-		loongson3_disable_clock(cpu);
-		break;
-	case CPU_UP_PREPARE:
-	case CPU_UP_PREPARE_FROZEN:
-		pr_info("Enable clock for CPU#%d\n", cpu);
-		loongson3_enable_clock(cpu);
-		break;
-	}
-
-	return NOTIFY_OK;
+	return 0;
 }
 
 static int register_loongson3_notifier(void)
 {
-	hotcpu_notifier(loongson3_cpu_callback, 0);
-	return 0;
+	return cpuhp_setup_state_nocalls(CPUHP_MIPS_LOONGSON_PREPARE,
+					 "mips/loongson:prepare",
+					 loongson3_enable_clock,
+					 loongson3_disable_clock);
 }
 early_initcall(register_loongson3_notifier);
 
diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h
index 058d312fdf6f..9e56be3eaf3f 100644
--- a/include/linux/cpuhotplug.h
+++ b/include/linux/cpuhotplug.h
@@ -45,6 +45,7 @@ enum cpuhp_state {
 	CPUHP_X86_MICRCODE_PREPARE,
 	CPUHP_NOTF_ERR_INJ_PREPARE,
 	CPUHP_MIPS_CAVIUM_PREPARE,
+	CPUHP_MIPS_LOONGSON_PREPARE,
 	CPUHP_TIMERS_DEAD,
 	CPUHP_BRINGUP_CPU,
 	CPUHP_AP_IDLE_DEAD,
-- 
2.9.3

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

#1477668 — [PATCH 12/21] cpufreq: Convert to hotplug state machine

Install the callbacks via the state machine.

Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
Cc: Viresh Kumar <viresh.kumar@linaro.org>
Cc: linux-pm@vger.kernel.org
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
 drivers/cpufreq/cpufreq.c | 38 ++++++++++++--------------------------
 1 file changed, 12 insertions(+), 26 deletions(-)

diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
index 3dd4884c6f9e..e0bc632a259e 100644
--- a/drivers/cpufreq/cpufreq.c
+++ b/drivers/cpufreq/cpufreq.c
@@ -1358,7 +1358,7 @@ static int cpufreq_add_dev(struct device *dev, struct subsys_interface *sif)
 	return add_cpu_dev_symlink(policy, cpu);
 }
 
-static void cpufreq_offline(unsigned int cpu)
+static int cpufreq_offline(unsigned int cpu)
 {
 	struct cpufreq_policy *policy;
 	int ret;
@@ -1368,7 +1368,7 @@ static void cpufreq_offline(unsigned int cpu)
 	policy = cpufreq_cpu_get_raw(cpu);
 	if (!policy) {
 		pr_debug("%s: No cpu_data found\n", __func__);
-		return;
+		return 0;
 	}
 
 	down_write(&policy->rwsem);
@@ -1417,6 +1417,7 @@ static void cpufreq_offline(unsigned int cpu)
 
 unlock:
 	up_write(&policy->rwsem);
+	return 0;
 }
 
 /**
@@ -2332,28 +2333,6 @@ int cpufreq_update_policy(unsigned int cpu)
 }
 EXPORT_SYMBOL(cpufreq_update_policy);
 
-static int cpufreq_cpu_callback(struct notifier_block *nfb,
-					unsigned long action, void *hcpu)
-{
-	unsigned int cpu = (unsigned long)hcpu;
-
-	switch (action & ~CPU_TASKS_FROZEN) {
-	case CPU_ONLINE:
-	case CPU_DOWN_FAILED:
-		cpufreq_online(cpu);
-		break;
-
-	case CPU_DOWN_PREPARE:
-		cpufreq_offline(cpu);
-		break;
-	}
-	return NOTIFY_OK;
-}
-
-static struct notifier_block __refdata cpufreq_cpu_notifier = {
-	.notifier_call = cpufreq_cpu_callback,
-};
-
 /*********************************************************************
  *               BOOST						     *
  *********************************************************************/
@@ -2455,6 +2434,7 @@ EXPORT_SYMBOL_GPL(cpufreq_boost_enabled);
 /*********************************************************************
  *               REGISTER / UNREGISTER CPUFREQ DRIVER                *
  *********************************************************************/
+static enum cpuhp_state hp_online;
 
 /**
  * cpufreq_register_driver - register a CPU Frequency driver
@@ -2517,7 +2497,13 @@ int cpufreq_register_driver(struct cpufreq_driver *driver_data)
 		goto err_if_unreg;
 	}
 
-	register_hotcpu_notifier(&cpufreq_cpu_notifier);
+	ret = cpuhp_setup_state_nocalls(CPUHP_AP_ONLINE_DYN, "cpufreq:online",
+					cpufreq_online,
+					cpufreq_offline);
+	if (ret < 0)
+		goto err_if_unreg;
+	hp_online = ret;
+
 	pr_debug("driver %s up and running\n", driver_data->name);
 	goto out;
 
@@ -2556,7 +2542,7 @@ int cpufreq_unregister_driver(struct cpufreq_driver *driver)
 	get_online_cpus();
 	subsys_interface_unregister(&cpufreq_interface);
 	remove_boost_sysfs_file();
-	unregister_hotcpu_notifier(&cpufreq_cpu_notifier);
+	cpuhp_remove_state_nocalls(hp_online);
 
 	write_lock_irqsave(&cpufreq_driver_lock, flags);
 
-- 
2.9.3

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

#1477831 — Re: [PATCH 12/21] cpufreq: Convert to hotplug state machine

On Tue, Sep 6, 2016 at 7:04 PM, Sebastian Andrzej Siewior
<bigeasy@linutronix.de> wrote:
> Install the callbacks via the state machine.
>
> Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>
> Cc: Viresh Kumar <viresh.kumar@linaro.org>
> Cc: linux-pm@vger.kernel.org
> Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
> ---
>  drivers/cpufreq/cpufreq.c | 38 ++++++++++++--------------------------
>  1 file changed, 12 insertions(+), 26 deletions(-)
>
> diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
> index 3dd4884c6f9e..e0bc632a259e 100644
> --- a/drivers/cpufreq/cpufreq.c
> +++ b/drivers/cpufreq/cpufreq.c
> @@ -1358,7 +1358,7 @@ static int cpufreq_add_dev(struct device *dev, struct subsys_interface *sif)
>         return add_cpu_dev_symlink(policy, cpu);
>  }
>
> -static void cpufreq_offline(unsigned int cpu)
> +static int cpufreq_offline(unsigned int cpu)
>  {
>         struct cpufreq_policy *policy;
>         int ret;
> @@ -1368,7 +1368,7 @@ static void cpufreq_offline(unsigned int cpu)
>         policy = cpufreq_cpu_get_raw(cpu);
>         if (!policy) {
>                 pr_debug("%s: No cpu_data found\n", __func__);
> -               return;
> +               return 0;
>         }
>
>         down_write(&policy->rwsem);
> @@ -1417,6 +1417,7 @@ static void cpufreq_offline(unsigned int cpu)
>
>  unlock:
>         up_write(&policy->rwsem);
> +       return 0;
>  }
>
>  /**
> @@ -2332,28 +2333,6 @@ int cpufreq_update_policy(unsigned int cpu)
>  }
>  EXPORT_SYMBOL(cpufreq_update_policy);
>
> -static int cpufreq_cpu_callback(struct notifier_block *nfb,
> -                                       unsigned long action, void *hcpu)
> -{
> -       unsigned int cpu = (unsigned long)hcpu;
> -
> -       switch (action & ~CPU_TASKS_FROZEN) {
> -       case CPU_ONLINE:
> -       case CPU_DOWN_FAILED:
> -               cpufreq_online(cpu);
> -               break;
> -
> -       case CPU_DOWN_PREPARE:
> -               cpufreq_offline(cpu);
> -               break;
> -       }
> -       return NOTIFY_OK;
> -}
> -
> -static struct notifier_block __refdata cpufreq_cpu_notifier = {
> -       .notifier_call = cpufreq_cpu_callback,
> -};
> -
>  /*********************************************************************
>   *               BOOST                                              *
>   *********************************************************************/
> @@ -2455,6 +2434,7 @@ EXPORT_SYMBOL_GPL(cpufreq_boost_enabled);
>  /*********************************************************************
>   *               REGISTER / UNREGISTER CPUFREQ DRIVER                *
>   *********************************************************************/
> +static enum cpuhp_state hp_online;
>
>  /**
>   * cpufreq_register_driver - register a CPU Frequency driver
> @@ -2517,7 +2497,13 @@ int cpufreq_register_driver(struct cpufreq_driver *driver_data)
>                 goto err_if_unreg;
>         }
>
> -       register_hotcpu_notifier(&cpufreq_cpu_notifier);
> +       ret = cpuhp_setup_state_nocalls(CPUHP_AP_ONLINE_DYN, "cpufreq:online",
> +                                       cpufreq_online,
> +                                       cpufreq_offline);
> +       if (ret < 0)
> +               goto err_if_unreg;
> +       hp_online = ret;

hp_online is enum cpuhp_state (and we pass it to
cpuhp_remove_state_nocalls(() later on), but
cpuhp_setup_state_nocalls() returns an int (and that should be 0 if it
is not an error code AFAICS), so is this actually correct?

Thanks,
Rafael

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

#1478333 — Re: [PATCH 12/21] cpufreq: Convert to hotplug state machine

On 2016-09-06 23:27:46 [+0200], Rafael J. Wysocki wrote:
> > +       ret = cpuhp_setup_state_nocalls(CPUHP_AP_ONLINE_DYN, "cpufreq:online",
> > +                                       cpufreq_online,
> > +                                       cpufreq_offline);
> > +       if (ret < 0)
> > +               goto err_if_unreg;
> > +       hp_online = ret;
> 
> hp_online is enum cpuhp_state (and we pass it to
> cpuhp_remove_state_nocalls(() later on), but
> cpuhp_setup_state_nocalls() returns an int (and that should be 0 if it
> is not an error code AFAICS), so is this actually correct?

Not sure what you are pointing out here. Let me try to cover it.
cpuhp_setup_state_nocalls() return <0 for errors. Those are are not
assigned to hp_online. It returns 0 for success on ID was !=
CPUHP_AP_ONLINE_DYN and >= 0 for success if ID was CPUHP_AP_ONLINE_DYN.
In the latter case the dynamic assigned ID is returned which should be
used if you plan to remove the callbacks.
Assigning an unsigned int to enum is okay because enumeration constants
itself should be an int.

> Thanks,
> Rafael

Sebastian

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

#1478447 — Re: [PATCH 12/21] cpufreq: Convert to hotplug state machine

On Wednesday, September 07, 2016 04:18:29 PM Sebastian Andrzej Siewior wrote:
> On 2016-09-06 23:27:46 [+0200], Rafael J. Wysocki wrote:
> > > +       ret = cpuhp_setup_state_nocalls(CPUHP_AP_ONLINE_DYN, "cpufreq:online",
> > > +                                       cpufreq_online,
> > > +                                       cpufreq_offline);
> > > +       if (ret < 0)
> > > +               goto err_if_unreg;
> > > +       hp_online = ret;
> > 
> > hp_online is enum cpuhp_state (and we pass it to
> > cpuhp_remove_state_nocalls(() later on), but
> > cpuhp_setup_state_nocalls() returns an int (and that should be 0 if it
> > is not an error code AFAICS), so is this actually correct?
> 
> Not sure what you are pointing out here. Let me try to cover it.
> cpuhp_setup_state_nocalls() return <0 for errors. Those are are not
> assigned to hp_online. It returns 0 for success on ID was !=
> CPUHP_AP_ONLINE_DYN and >= 0 for success if ID was CPUHP_AP_ONLINE_DYN.
> In the latter case the dynamic assigned ID is returned which should be
> used if you plan to remove the callbacks.

OK, that last part wasn't clear to me.  The kerneldoc comment for
__cpuhp_setup_state() doesn't mention the possible non-zero return values
on success, which is a bit confusing IMHO.

> Assigning an unsigned int to enum is okay because enumeration constants
> itself should be an int.

But the unsigned int still may be out of range for the given enum, so I
wouldn't call it particularly clean. :-)

Anyway, please feel free to add

Acked-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>

to the patch.

Thanks,
Rafael

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

#1477670 — [PATCH 07/21] lib: irq_poll: Convert to hotplug state machine

Install the callbacks via the state machine.

Cc: Jens Axboe <axboe@kernel.dk>
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
 include/linux/cpuhotplug.h |  1 +
 lib/irq_poll.c             | 26 +++++++++-----------------
 2 files changed, 10 insertions(+), 17 deletions(-)

diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h
index 4ca7551b902f..be6c01b7efbe 100644
--- a/include/linux/cpuhotplug.h
+++ b/include/linux/cpuhotplug.h
@@ -22,6 +22,7 @@ enum cpuhp_state {
 	CPUHP_CPUIDLE_DEAD,
 	CPUHP_ARM64_FPSIMD_DEAD,
 	CPUHP_ARM_OMAP_WAKE_DEAD,
+	CPUHP_IRQ_POLL_DEAD,
 	CPUHP_WORKQUEUE_PREP,
 	CPUHP_POWER_NUMA_PREPARE,
 	CPUHP_HRTIMERS_PREPARE,
diff --git a/lib/irq_poll.c b/lib/irq_poll.c
index 836f7db4e548..2be55692aa43 100644
--- a/lib/irq_poll.c
+++ b/lib/irq_poll.c
@@ -184,30 +184,21 @@ void irq_poll_init(struct irq_poll *iop, int weight, irq_poll_fn *poll_fn)
 }
 EXPORT_SYMBOL(irq_poll_init);
 
-static int irq_poll_cpu_notify(struct notifier_block *self,
-				 unsigned long action, void *hcpu)
+static int irq_poll_cpu_dead(unsigned int cpu)
 {
 	/*
 	 * If a CPU goes away, splice its entries to the current CPU
 	 * and trigger a run of the softirq
 	 */
-	if (action == CPU_DEAD || action == CPU_DEAD_FROZEN) {
-		int cpu = (unsigned long) hcpu;
+	local_irq_disable();
+	list_splice_init(&per_cpu(blk_cpu_iopoll, cpu),
+			 this_cpu_ptr(&blk_cpu_iopoll));
+	__raise_softirq_irqoff(IRQ_POLL_SOFTIRQ);
+	local_irq_enable();
 
-		local_irq_disable();
-		list_splice_init(&per_cpu(blk_cpu_iopoll, cpu),
-				 this_cpu_ptr(&blk_cpu_iopoll));
-		__raise_softirq_irqoff(IRQ_POLL_SOFTIRQ);
-		local_irq_enable();
-	}
-
-	return NOTIFY_OK;
+	return 0;
 }
 
-static struct notifier_block irq_poll_cpu_notifier = {
-	.notifier_call	= irq_poll_cpu_notify,
-};
-
 static __init int irq_poll_setup(void)
 {
 	int i;
@@ -216,7 +207,8 @@ static __init int irq_poll_setup(void)
 		INIT_LIST_HEAD(&per_cpu(blk_cpu_iopoll, i));
 
 	open_softirq(IRQ_POLL_SOFTIRQ, irq_poll_softirq);
-	register_hotcpu_notifier(&irq_poll_cpu_notifier);
+	cpuhp_setup_state_nocalls(CPUHP_IRQ_POLL_DEAD, "irq_poll:dead", NULL,
+				  irq_poll_cpu_dead);
 	return 0;
 }
 subsys_initcall(irq_poll_setup);
-- 
2.9.3

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

#1477672 — [PATCH 01/21] arm64: FP/SIMD: Convert to hotplug state machine

Install the callbacks via the state machine.

Cc: Catalin Marinas <catalin.marinas@arm.com>
Cc: Will Deacon <will.deacon@arm.com>
Cc: linux-arm-kernel@lists.infradead.org
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
 arch/arm64/kernel/fpsimd.c | 22 +++++-----------------
 include/linux/cpuhotplug.h |  1 +
 2 files changed, 6 insertions(+), 17 deletions(-)

diff --git a/arch/arm64/kernel/fpsimd.c b/arch/arm64/kernel/fpsimd.c
index 975b274ee7b5..394c61db5566 100644
--- a/arch/arm64/kernel/fpsimd.c
+++ b/arch/arm64/kernel/fpsimd.c
@@ -299,28 +299,16 @@ static inline void fpsimd_pm_init(void) { }
 #endif /* CONFIG_CPU_PM */
 
 #ifdef CONFIG_HOTPLUG_CPU
-static int fpsimd_cpu_hotplug_notifier(struct notifier_block *nfb,
-				       unsigned long action,
-				       void *hcpu)
+static int fpsimd_cpu_dead(unsigned int cpu)
 {
-	unsigned int cpu = (long)hcpu;
-
-	switch (action) {
-	case CPU_DEAD:
-	case CPU_DEAD_FROZEN:
-		per_cpu(fpsimd_last_state, cpu) = NULL;
-		break;
-	}
-	return NOTIFY_OK;
+	per_cpu(fpsimd_last_state, cpu) = NULL;
+	return 0;
 }
 
-static struct notifier_block fpsimd_cpu_hotplug_notifier_block = {
-	.notifier_call = fpsimd_cpu_hotplug_notifier,
-};
-
 static inline void fpsimd_hotplug_init(void)
 {
-	register_cpu_notifier(&fpsimd_cpu_hotplug_notifier_block);
+	cpuhp_setup_state_nocalls(CPUHP_ARM64_FPSIMD_DEAD, "arm64/fpsimd:dead",
+				  NULL, fpsimd_cpu_dead);
 }
 
 #else
diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h
index afd59e2ca4b3..0da071ff36d2 100644
--- a/include/linux/cpuhotplug.h
+++ b/include/linux/cpuhotplug.h
@@ -20,6 +20,7 @@ enum cpuhp_state {
 	CPUHP_SOFTIRQ_DEAD,
 	CPUHP_NET_MVNETA_DEAD,
 	CPUHP_CPUIDLE_DEAD,
+	CPUHP_ARM64_FPSIMD_DEAD,
 	CPUHP_WORKQUEUE_PREP,
 	CPUHP_POWER_NUMA_PREPARE,
 	CPUHP_HRTIMERS_PREPARE,
-- 
2.9.3

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

#1477726 — Re: [PATCH 01/21] arm64: FP/SIMD: Convert to hotplug state machine

On Tue, Sep 06, 2016 at 07:04:37PM +0200, Sebastian Andrzej Siewior wrote:
> Install the callbacks via the state machine.
> 
> Cc: Catalin Marinas <catalin.marinas@arm.com>
> Cc: Will Deacon <will.deacon@arm.com>
> Cc: linux-arm-kernel@lists.infradead.org
> Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
> ---
>  arch/arm64/kernel/fpsimd.c | 22 +++++-----------------
>  include/linux/cpuhotplug.h |  1 +
>  2 files changed, 6 insertions(+), 17 deletions(-)

Looks fine to me:

Acked-by: Will Deacon <will.deacon@arm.com>

> diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h
> index afd59e2ca4b3..0da071ff36d2 100644
> --- a/include/linux/cpuhotplug.h
> +++ b/include/linux/cpuhotplug.h
> @@ -20,6 +20,7 @@ enum cpuhp_state {
>  	CPUHP_SOFTIRQ_DEAD,
>  	CPUHP_NET_MVNETA_DEAD,
>  	CPUHP_CPUIDLE_DEAD,
> +	CPUHP_ARM64_FPSIMD_DEAD,
>  	CPUHP_WORKQUEUE_PREP,
>  	CPUHP_POWER_NUMA_PREPARE,
>  	CPUHP_HRTIMERS_PREPARE,

Are you planning to take this via your tree? The arm64 queue already
touches cpuhotplug.h for the hw-breakpoint and debugmon changes, so
there's a simple conflict to address here.

Will

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

#1477674 — [PATCH 05/21] sh: SH-X3 SMP: Convert to hotplug state machine

Install the callbacks via the state machine.

Cc: linux-sh@vger.kernel.org
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
 arch/sh/kernel/cpu/sh4a/smp-shx3.c | 26 +++++---------------------
 include/linux/cpuhotplug.h         |  1 +
 2 files changed, 6 insertions(+), 21 deletions(-)

diff --git a/arch/sh/kernel/cpu/sh4a/smp-shx3.c b/arch/sh/kernel/cpu/sh4a/smp-shx3.c
index 839612c8a0a0..0d3637c494bf 100644
--- a/arch/sh/kernel/cpu/sh4a/smp-shx3.c
+++ b/arch/sh/kernel/cpu/sh4a/smp-shx3.c
@@ -122,32 +122,16 @@ static void shx3_update_boot_vector(unsigned int cpu)
 	__raw_writel(STBCR_RESET, STBCR_REG(cpu));
 }
 
-static int
-shx3_cpu_callback(struct notifier_block *nfb, unsigned long action, void *hcpu)
+static int shx3_cpu_prepare(unsigned int cpu)
 {
-	unsigned int cpu = (unsigned int)hcpu;
-
-	switch (action) {
-	case CPU_UP_PREPARE:
-		shx3_update_boot_vector(cpu);
-		break;
-	case CPU_ONLINE:
-		pr_info("CPU %u is now online\n", cpu);
-		break;
-	case CPU_DEAD:
-		break;
-	}
-
-	return NOTIFY_OK;
+	shx3_update_boot_vector(cpu);
+	return 0;
 }
 
-static struct notifier_block shx3_cpu_notifier = {
-	.notifier_call		= shx3_cpu_callback,
-};
-
 static int register_shx3_cpu_notifier(void)
 {
-	register_hotcpu_notifier(&shx3_cpu_notifier);
+	cpuhp_setup_state_nocalls(CPUHP_SH_SH3X_PREPARE, "sh/shx3:prepare",
+				  shx3_cpu_prepare, NULL);
 	return 0;
 }
 late_initcall(register_shx3_cpu_notifier);
diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h
index 35859aafbbfc..8dec2a236af3 100644
--- a/include/linux/cpuhotplug.h
+++ b/include/linux/cpuhotplug.h
@@ -37,6 +37,7 @@ enum cpuhp_state {
 	CPUHP_POWERPC_MMU_CTX_PREPARE,
 	CPUHP_NOTIFY_PREPARE,
 	CPUHP_ARM_SHMOBILE_SCU_PREPARE,
+	CPUHP_SH_SH3X_PREPARE,
 	CPUHP_TIMERS_DEAD,
 	CPUHP_BRINGUP_CPU,
 	CPUHP_AP_IDLE_DEAD,
-- 
2.9.3

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

#1477675 — [PATCH 10/21] virtio scsi: Convert to hotplug state machine

Install the callbacks via the state machine. It uses the multi instance
infrastructure of the hotplug code to handle each interface.

virtscsi_set_affinity() is removed from virtscsi_init() because
virtscsi_cpu_notif_add() (the function which registers the instance) is invoked
right after it and the cpuhp_state_add_instance() functions invokes the startup
callback on all online CPUs.

The same thing can not be applied virtscsi_cpu_notif_remove() because
virtscsi_remove_vqs() invokes virtscsi_set_affinity() with affinity = false as
argument but the old CPU_DEAD state invoked the function with affinity = true
(which does not match the DEAD callback).

Cc: "James E.J. Bottomley" <jejb@linux.vnet.ibm.com>
Cc: "Martin K. Petersen" <martin.petersen@oracle.com>
Cc: linux-scsi@vger.kernel.org
Cc: "Michael S. Tsirkin" <mst@redhat.com>
Cc: virtualization@lists.linux-foundation.org
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
 drivers/scsi/virtio_scsi.c | 76 ++++++++++++++++++++++++++++++----------------
 include/linux/cpuhotplug.h |  1 +
 2 files changed, 50 insertions(+), 27 deletions(-)

diff --git a/drivers/scsi/virtio_scsi.c b/drivers/scsi/virtio_scsi.c
index 7dbbb29d24c6..deefab3a94d0 100644
--- a/drivers/scsi/virtio_scsi.c
+++ b/drivers/scsi/virtio_scsi.c
@@ -107,8 +107,8 @@ struct virtio_scsi {
 	/* If the affinity hint is set for virtqueues */
 	bool affinity_hint_set;
 
-	/* CPU hotplug notifier */
-	struct notifier_block nb;
+	struct hlist_node node;
+	struct hlist_node node_dead;
 
 	/* Protected by event_vq lock */
 	bool stop_events;
@@ -118,6 +118,7 @@ struct virtio_scsi {
 	struct virtio_scsi_vq req_vqs[];
 };
 
+static enum cpuhp_state virtioscsi_online;
 static struct kmem_cache *virtscsi_cmd_cache;
 static mempool_t *virtscsi_cmd_pool;
 
@@ -852,21 +853,33 @@ static void virtscsi_set_affinity(struct virtio_scsi *vscsi, bool affinity)
 	put_online_cpus();
 }
 
-static int virtscsi_cpu_callback(struct notifier_block *nfb,
-				 unsigned long action, void *hcpu)
+static int virtscsi_cpu_online(unsigned int cpu, struct hlist_node *node)
 {
-	struct virtio_scsi *vscsi = container_of(nfb, struct virtio_scsi, nb);
-	switch(action) {
-	case CPU_ONLINE:
-	case CPU_ONLINE_FROZEN:
-	case CPU_DEAD:
-	case CPU_DEAD_FROZEN:
-		__virtscsi_set_affinity(vscsi, true);
-		break;
-	default:
-		break;
-	}
-	return NOTIFY_OK;
+	struct virtio_scsi *vscsi = hlist_entry_safe(node, struct virtio_scsi,
+						     node);
+	__virtscsi_set_affinity(vscsi, true);
+	return 0;
+}
+
+static int virtscsi_cpu_notif_add(struct virtio_scsi *vi)
+{
+	int ret;
+
+	ret = cpuhp_state_add_instance(virtioscsi_online, &vi->node);
+	if (ret)
+		return ret;
+
+	ret = cpuhp_state_add_instance(CPUHP_VIRT_SCSI_DEAD, &vi->node_dead);
+	if (ret)
+		cpuhp_state_remove_instance(virtioscsi_online, &vi->node);
+	return ret;
+}
+
+static void virtscsi_cpu_notif_remove(struct virtio_scsi *vi)
+{
+	cpuhp_state_remove_instance_nocalls(virtioscsi_online, &vi->node);
+	cpuhp_state_remove_instance_nocalls(CPUHP_VIRT_SCSI_DEAD,
+					    &vi->node_dead);
 }
 
 static void virtscsi_init_vq(struct virtio_scsi_vq *virtscsi_vq,
@@ -929,8 +942,6 @@ static int virtscsi_init(struct virtio_device *vdev,
 		virtscsi_init_vq(&vscsi->req_vqs[i - VIRTIO_SCSI_VQ_BASE],
 				 vqs[i]);
 
-	virtscsi_set_affinity(vscsi, true);
-
 	virtscsi_config_set(vdev, cdb_size, VIRTIO_SCSI_CDB_SIZE);
 	virtscsi_config_set(vdev, sense_size, VIRTIO_SCSI_SENSE_SIZE);
 
@@ -987,12 +998,9 @@ static int virtscsi_probe(struct virtio_device *vdev)
 	if (err)
 		goto virtscsi_init_failed;
 
-	vscsi->nb.notifier_call = &virtscsi_cpu_callback;
-	err = register_hotcpu_notifier(&vscsi->nb);
-	if (err) {
-		pr_err("registering cpu notifier failed\n");
+	err = virtscsi_cpu_notif_add(vscsi);
+	if (err)
 		goto scsi_add_host_failed;
-	}
 
 	cmd_per_lun = virtscsi_config_get(vdev, cmd_per_lun) ?: 1;
 	shost->cmd_per_lun = min_t(u32, cmd_per_lun, shost->can_queue);
@@ -1049,7 +1057,7 @@ static void virtscsi_remove(struct virtio_device *vdev)
 
 	scsi_remove_host(shost);
 
-	unregister_hotcpu_notifier(&vscsi->nb);
+	virtscsi_cpu_notif_remove(vscsi);
 
 	virtscsi_remove_vqs(vdev);
 	scsi_host_put(shost);
@@ -1061,7 +1069,7 @@ static int virtscsi_freeze(struct virtio_device *vdev)
 	struct Scsi_Host *sh = virtio_scsi_host(vdev);
 	struct virtio_scsi *vscsi = shost_priv(sh);
 
-	unregister_hotcpu_notifier(&vscsi->nb);
+	virtscsi_cpu_notif_remove(vscsi);
 	virtscsi_remove_vqs(vdev);
 	return 0;
 }
@@ -1076,12 +1084,11 @@ static int virtscsi_restore(struct virtio_device *vdev)
 	if (err)
 		return err;
 
-	err = register_hotcpu_notifier(&vscsi->nb);
+	err = virtscsi_cpu_notif_add(vscsi);
 	if (err) {
 		vdev->config->del_vqs(vdev);
 		return err;
 	}
-
 	virtio_device_ready(vdev);
 
 	if (virtio_has_feature(vdev, VIRTIO_SCSI_F_HOTPLUG))
@@ -1136,6 +1143,16 @@ static int __init init(void)
 		pr_err("mempool_create() for virtscsi_cmd_pool failed\n");
 		goto error;
 	}
+	ret = cpuhp_setup_state_multi(CPUHP_AP_ONLINE_DYN,
+				      "scsi/virtio:online",
+				      virtscsi_cpu_online, NULL);
+	if (ret < 0)
+		goto error;
+	virtioscsi_online = ret;
+	ret = cpuhp_setup_state_multi(CPUHP_VIRT_SCSI_DEAD, "scsi/virtio:dead",
+				      NULL, virtscsi_cpu_online);
+	if (ret)
+		goto error;
 	ret = register_virtio_driver(&virtio_scsi_driver);
 	if (ret < 0)
 		goto error;
@@ -1151,12 +1168,17 @@ static int __init init(void)
 		kmem_cache_destroy(virtscsi_cmd_cache);
 		virtscsi_cmd_cache = NULL;
 	}
+	if (virtioscsi_online)
+		cpuhp_remove_multi_state(virtioscsi_online);
+	cpuhp_remove_multi_state(CPUHP_VIRT_SCSI_DEAD);
 	return ret;
 }
 
 static void __exit fini(void)
 {
 	unregister_virtio_driver(&virtio_scsi_driver);
+	cpuhp_remove_multi_state(virtioscsi_online);
+	cpuhp_remove_multi_state(CPUHP_VIRT_SCSI_DEAD);
 	mempool_destroy(virtscsi_cmd_pool);
 	kmem_cache_destroy(virtscsi_cmd_cache);
 }
diff --git a/include/linux/cpuhotplug.h b/include/linux/cpuhotplug.h
index a12044828e03..73a4daf08c04 100644
--- a/include/linux/cpuhotplug.h
+++ b/include/linux/cpuhotplug.h
@@ -24,6 +24,7 @@ enum cpuhp_state {
 	CPUHP_ARM_OMAP_WAKE_DEAD,
 	CPUHP_IRQ_POLL_DEAD,
 	CPUHP_BLOCK_SOFTIRQ_DEAD,
+	CPUHP_VIRT_SCSI_DEAD,
 	CPUHP_WORKQUEUE_PREP,
 	CPUHP_POWER_NUMA_PREPARE,
 	CPUHP_HRTIMERS_PREPARE,
-- 
2.9.3

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

#1477682 — [PATCH 09/21] oprofile: timer: Convert to hotplug state machine

Install the callbacks via the state machine and let the core invoke
the callbacks on the already online CPUs.
Since the online target runs always on the target CPU we can drop
smp_call_function_single(). The functions is invoked with interrupts off to
keep the old calling convention. If the maintainer things that this function
can be called with interrupts enabled then it can be removed :)

Cc: Robert Richter <rric@kernel.org>
Cc: oprofile-list@lists.sf.net
Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de>
---
 drivers/oprofile/timer_int.c | 44 +++++++++++++++++++++++---------------------
 1 file changed, 23 insertions(+), 21 deletions(-)

diff --git a/drivers/oprofile/timer_int.c b/drivers/oprofile/timer_int.c
index bdef916e5dda..2498a6cd7c24 100644
--- a/drivers/oprofile/timer_int.c
+++ b/drivers/oprofile/timer_int.c
@@ -74,37 +74,39 @@ static void oprofile_hrtimer_stop(void)
 	put_online_cpus();
 }
 
-static int oprofile_cpu_notify(struct notifier_block *self,
-			       unsigned long action, void *hcpu)
+static int oprofile_timer_online(unsigned int cpu)
 {
-	long cpu = (long) hcpu;
-
-	switch (action) {
-	case CPU_ONLINE:
-	case CPU_ONLINE_FROZEN:
-		smp_call_function_single(cpu, __oprofile_hrtimer_start,
-					 NULL, 1);
-		break;
-	case CPU_DEAD:
-	case CPU_DEAD_FROZEN:
-		__oprofile_hrtimer_stop(cpu);
-		break;
-	}
-	return NOTIFY_OK;
+	local_irq_disable();
+	__oprofile_hrtimer_start(NULL);
+	local_irq_enable();
+	return 0;
 }
 
-static struct notifier_block __refdata oprofile_cpu_notifier = {
-	.notifier_call = oprofile_cpu_notify,
-};
+static int oprofile_timer_prep_down(unsigned int cpu)
+{
+	__oprofile_hrtimer_stop(cpu);
+	return 0;
+}
+
+static enum cpuhp_state hp_online;
 
 static int oprofile_hrtimer_setup(void)
 {
-	return register_hotcpu_notifier(&oprofile_cpu_notifier);
+	int ret;
+
+	ret = cpuhp_setup_state_nocalls(CPUHP_AP_ONLINE_DYN,
+					"oprofile/timer:online",
+					oprofile_timer_online,
+					oprofile_timer_prep_down);
+	if (ret < 0)
+		return ret;
+	hp_online = ret;
+	return 0;
 }
 
 static void oprofile_hrtimer_shutdown(void)
 {
-	unregister_hotcpu_notifier(&oprofile_cpu_notifier);
+	cpuhp_remove_state_nocalls(hp_online);
 }
 
 int oprofile_timer_init(struct oprofile_operations *ops)
-- 
2.9.3

[toc] | [prev] | [standalone]

Page 2 of 2 — ← Prev page 1 [2]

Back to top | Article view | linux.kernel

csiph-web