[PATCH] Disable CPU hotplug during suspend

The current suspend code has to be run on one CPU, so we use the CPU
hotplug to take the non-boot CPUs offline on SMP machines.  However, we
should also make sure that these CPUs will not be enabled by someone else
after we have disabled them.

The functions disable_nonboot_cpus() and enable_nonboot_cpus() are moved to
kernel/cpu.c, because they now refer to some stuff in there that should
better be static.  Also it's better if disable_nonboot_cpus() returns an
error instead of panicking if something goes wrong, and
enable_nonboot_cpus() has no reason to panic(), because the CPUs may have
been enabled by the userland before it tries to take them online.

Signed-off-by: Rafael J. Wysocki <rjw@sisk.pl>
Acked-by: Pavel Machek <pavel@ucw.cz>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

include/linux/cpu.h

@@ -89,4 +89,12 @@ int cpu_down(unsigned int cpu);
 static inline int cpu_is_offline(int cpu) { return 0; }
 #endif
 
+#ifdef CONFIG_SUSPEND_SMP
+extern int disable_nonboot_cpus(void);
+extern void enable_nonboot_cpus(void);
+#else
+static inline int disable_nonboot_cpus(void) { return 0; }
+static inline void enable_nonboot_cpus(void) {}
+#endif
+
 #endif /* _LINUX_CPU_H_ */

include/linux/suspend.h

@@ -57,14 +57,6 @@ static inline int software_suspend(void)
 }
 #endif /* CONFIG_PM */
 
-#ifdef CONFIG_SUSPEND_SMP
-extern void disable_nonboot_cpus(void);
-extern void enable_nonboot_cpus(void);
-#else
-static inline void disable_nonboot_cpus(void) {}
-static inline void enable_nonboot_cpus(void) {}
-#endif
-
 void save_processor_state(void);
 void restore_processor_state(void);
 struct saved_context;

kernel/cpu.c

@@ -21,6 +21,11 @@ static DEFINE_MUTEX(cpu_bitmask_lock);
 
 static __cpuinitdata BLOCKING_NOTIFIER_HEAD(cpu_chain);
 
+/* If set, cpu_up and cpu_down will return -EBUSY and do nothing.
+ * Should always be manipulated under cpu_add_remove_lock
+ */
+static int cpu_hotplug_disabled;
+
 #ifdef CONFIG_HOTPLUG_CPU
 
 /* Crappy recursive lock-takers in cpufreq! Complain loudly about idiots */
@@ -108,30 +113,25 @@ static int take_cpu_down(void *unused)
 	return 0;
 }
 
-int cpu_down(unsigned int cpu)
+/* Requires cpu_add_remove_lock to be held */
+static int _cpu_down(unsigned int cpu)
 {
 	int err;
 	struct task_struct *p;
 	cpumask_t old_allowed, tmp;
 
-	mutex_lock(&cpu_add_remove_lock);
-	if (num_online_cpus() == 1) {
-		err = -EBUSY;
-		goto out;
-	}
+	if (num_online_cpus() == 1)
+		return -EBUSY;
 
-	if (!cpu_online(cpu)) {
-		err = -EINVAL;
-		goto out;
-	}
+	if (!cpu_online(cpu))
+		return -EINVAL;
 
 	err = blocking_notifier_call_chain(&cpu_chain, CPU_DOWN_PREPARE,
 						(void *)(long)cpu);
 	if (err == NOTIFY_BAD) {
 		printk("%s: attempt to take down CPU %u failed\n",
 				__FUNCTION__, cpu);
-		err = -EINVAL;
-		goto out;
+		return -EINVAL;
 	}
 
 	/* Ensure that we are not runnable on dying cpu */
@@ -179,22 +179,32 @@ int cpu_down(unsigned int cpu)
 	err = kthread_stop(p);
 out_allowed:
 	set_cpus_allowed(current, old_allowed);
-out:
+	return err;
+}
+
+int cpu_down(unsigned int cpu)
+{
+	int err = 0;
+
+	mutex_lock(&cpu_add_remove_lock);
+	if (cpu_hotplug_disabled)
+		err = -EBUSY;
+	else
+		err = _cpu_down(cpu);
+
 	mutex_unlock(&cpu_add_remove_lock);
 	return err;
 }
 #endif /*CONFIG_HOTPLUG_CPU*/
 
-int __devinit cpu_up(unsigned int cpu)
+/* Requires cpu_add_remove_lock to be held */
+static int __devinit _cpu_up(unsigned int cpu)
 {
 	int ret;
 	void *hcpu = (void *)(long)cpu;
 
-	mutex_lock(&cpu_add_remove_lock);
-	if (cpu_online(cpu) || !cpu_present(cpu)) {
-		ret = -EINVAL;
-		goto out;
-	}
+	if (cpu_online(cpu) || !cpu_present(cpu))
+		return -EINVAL;
 
 	ret = blocking_notifier_call_chain(&cpu_chain, CPU_UP_PREPARE, hcpu);
 	if (ret == NOTIFY_BAD) {
@@ -219,7 +229,95 @@ int __devinit cpu_up(unsigned int cpu)
 	if (ret != 0)
 		blocking_notifier_call_chain(&cpu_chain,
 				CPU_UP_CANCELED, hcpu);
+
+	return ret;
+}
+
+int __devinit cpu_up(unsigned int cpu)
+{
+	int err = 0;
+
+	mutex_lock(&cpu_add_remove_lock);
+	if (cpu_hotplug_disabled)
+		err = -EBUSY;
+	else
+		err = _cpu_up(cpu);
+
+	mutex_unlock(&cpu_add_remove_lock);
+	return err;
+}
+
+#ifdef CONFIG_SUSPEND_SMP
+static cpumask_t frozen_cpus;
+
+int disable_nonboot_cpus(void)
+{
+	int cpu, first_cpu, error;
+
+	mutex_lock(&cpu_add_remove_lock);
+	first_cpu = first_cpu(cpu_present_map);
+	if (!cpu_online(first_cpu)) {
+		error = _cpu_up(first_cpu);
+		if (error) {
+			printk(KERN_ERR "Could not bring CPU%d up.\n",
+				first_cpu);
+			goto out;
+		}
+	}
+	error = set_cpus_allowed(current, cpumask_of_cpu(first_cpu));
+	if (error) {
+		printk(KERN_ERR "Could not run on CPU%d\n", first_cpu);
+		goto out;
+	}
+	/* We take down all of the non-boot CPUs in one shot to avoid races
+	 * with the userspace trying to use the CPU hotplug at the same time
+	 */
+	cpus_clear(frozen_cpus);
+	printk("Disabling non-boot CPUs ...\n");
+	for_each_online_cpu(cpu) {
+		if (cpu == first_cpu)
+			continue;
+		error = _cpu_down(cpu);
+		if (!error) {
+			cpu_set(cpu, frozen_cpus);
+			printk("CPU%d is down\n", cpu);
+		} else {
+			printk(KERN_ERR "Error taking CPU%d down: %d\n",
+				cpu, error);
+			break;
+		}
+	}
+	if (!error) {
+		BUG_ON(num_online_cpus() > 1);
+		/* Make sure the CPUs won't be enabled by someone else */
+		cpu_hotplug_disabled = 1;
+	} else {
+		printk(KERN_ERR "Non-boot CPUs are not disabled");
+	}
 out:
 	mutex_unlock(&cpu_add_remove_lock);
-	return ret;
+	return error;
+}
+
+void enable_nonboot_cpus(void)
+{
+	int cpu, error;
+
+	/* Allow everyone to use the CPU hotplug again */
+	mutex_lock(&cpu_add_remove_lock);
+	cpu_hotplug_disabled = 0;
+	mutex_unlock(&cpu_add_remove_lock);
+
+	printk("Enabling non-boot CPUs ...\n");
+	for_each_cpu_mask(cpu, frozen_cpus) {
+		error = cpu_up(cpu);
+		if (!error) {
+			printk("CPU%d is up\n", cpu);
+			continue;
+		}
+		printk(KERN_WARNING "Error taking CPU%d up: %d\n",
+			cpu, error);
+	}
+	cpus_clear(frozen_cpus);
 }
+#endif

kernel/power/Makefile

@@ -7,6 +7,4 @@ obj-y				:= main.o process.o console.o
 obj-$(CONFIG_PM_LEGACY)		+= pm.o
 obj-$(CONFIG_SOFTWARE_SUSPEND)	+= swsusp.o disk.o snapshot.o swap.o user.o
 
-obj-$(CONFIG_SUSPEND_SMP)	+= smp.o
-
 obj-$(CONFIG_MAGIC_SYSRQ)	+= poweroff.o

kernel/power/disk.c

@@ -18,6 +18,7 @@
 #include <linux/fs.h>
 #include <linux/mount.h>
 #include <linux/pm.h>
+#include <linux/cpu.h>
 
 #include "power.h"
 
@@ -72,7 +73,10 @@ static int prepare_processes(void)
 	int error;
 
 	pm_prepare_console();
-	disable_nonboot_cpus();
+
+	error = disable_nonboot_cpus();
+	if (error)
+		goto enable_cpus;
 
 	if (freeze_processes()) {
 		error = -EBUSY;
@@ -84,6 +88,7 @@ static int prepare_processes(void)
 		return 0;
 thaw:
 	thaw_processes();
+enable_cpus:
 	enable_nonboot_cpus();
 	pm_restore_console();
 	return error;

kernel/power/main.c

@@ -16,6 +16,7 @@
 #include <linux/init.h>
 #include <linux/pm.h>
 #include <linux/console.h>
+#include <linux/cpu.h>
 
 #include "power.h"
 
@@ -51,20 +52,17 @@ void pm_set_ops(struct pm_ops * ops)
 
 static int suspend_prepare(suspend_state_t state)
 {
-	int error = 0;
+	int error;
 	unsigned int free_pages;
 
 	if (!pm_ops || !pm_ops->enter)
 		return -EPERM;
 
 	pm_prepare_console();
 
-	disable_nonboot_cpus();
-
-	if (num_online_cpus() != 1) {
-		error = -EPERM;
+	error = disable_nonboot_cpus();
+	if (error)
 		goto Enable_cpu;
-	}
 
 	if (freeze_processes()) {
 		error = -EAGAIN;

kernel/power/user.c

@@ -19,6 +19,7 @@
 #include <linux/swapops.h>
 #include <linux/pm.h>
 #include <linux/fs.h>
+#include <linux/cpu.h>
 
 #include <asm/uaccess.h>
 
@@ -139,12 +140,15 @@ static int snapshot_ioctl(struct inode *inode, struct file *filp,
 		if (data->frozen)
 			break;
 		down(&pm_sem);
-		disable_nonboot_cpus();
-		if (freeze_processes()) {
-			thaw_processes();
-			enable_nonboot_cpus();
-			error = -EBUSY;
+		error = disable_nonboot_cpus();
+		if (!error) {
+			error = freeze_processes();
+			if (error) {
+				thaw_processes();
+				error = -EBUSY;
+			}
 		}
+		enable_nonboot_cpus();
 		up(&pm_sem);
 		if (!error)
 			data->frozen = 1;