1 files changed, 134 insertions, 85 deletions

diff --git a/arch/x86/oprofile/nmi_int.c b/arch/x86/oprofile/nmi_int.c
index cb88b1a0bd5f..b28d2f1253bb 100644
--- a/arch/x86/oprofile/nmi_int.c
+++ b/arch/x86/oprofile/nmi_int.c
@@ -31,8 +31,9 @@ static struct op_x86_model_spec *model;
 static DEFINE_PER_CPU(struct op_msrs, cpu_msrs);
 static DEFINE_PER_CPU(unsigned long, saved_lvtpc);
 
-/* 0 == registered but off, 1 == registered and on */
-static int nmi_enabled = 0;
+/* must be protected with get_online_cpus()/put_online_cpus(): */
+static int nmi_enabled;
+static int ctr_running;
 
 struct op_counter_config counter_config[OP_MAX_COUNTER];
 
@@ -61,12 +62,16 @@ static int profile_exceptions_notify(struct notifier_block *self,
 {
 	struct die_args *args = (struct die_args *)data;
 	int ret = NOTIFY_DONE;
-	int cpu = smp_processor_id();
 
 	switch (val) {
 	case DIE_NMI:
 	case DIE_NMI_IPI:
-		model->check_ctrs(args->regs, &per_cpu(cpu_msrs, cpu));
+		if (ctr_running)
+			model->check_ctrs(args->regs, &__get_cpu_var(cpu_msrs));
+		else if (!nmi_enabled)
+			break;
+		else
+			model->stop(&__get_cpu_var(cpu_msrs));
 		ret = NOTIFY_STOP;
 		break;
 	default:
@@ -95,24 +100,36 @@ static void nmi_cpu_save_registers(struct op_msrs *msrs)
 static void nmi_cpu_start(void *dummy)
 {
 	struct op_msrs const *msrs = &__get_cpu_var(cpu_msrs);
-	model->start(msrs);
+	if (!msrs->controls)
+		WARN_ON_ONCE(1);
+	else
+		model->start(msrs);
 }
 
 static int nmi_start(void)
 {
+	get_online_cpus();
 	on_each_cpu(nmi_cpu_start, NULL, 1);
+	ctr_running = 1;
+	put_online_cpus();
 	return 0;
 }
 
 static void nmi_cpu_stop(void *dummy)
 {
 	struct op_msrs const *msrs = &__get_cpu_var(cpu_msrs);
-	model->stop(msrs);
+	if (!msrs->controls)
+		WARN_ON_ONCE(1);
+	else
+		model->stop(msrs);
 }
 
 static void nmi_stop(void)
 {
+	get_online_cpus();
 	on_each_cpu(nmi_cpu_stop, NULL, 1);
+	ctr_running = 0;
+	put_online_cpus();
 }
 
 #ifdef CONFIG_OPROFILE_EVENT_MULTIPLEX
@@ -159,7 +176,7 @@ static int nmi_setup_mux(void)
 
 	for_each_possible_cpu(i) {
 		per_cpu(cpu_msrs, i).multiplex =
-			kmalloc(multiplex_size, GFP_KERNEL);
+			kzalloc(multiplex_size, GFP_KERNEL);
 		if (!per_cpu(cpu_msrs, i).multiplex)
 			return 0;
 	}
@@ -179,7 +196,6 @@ static void nmi_cpu_setup_mux(int cpu, struct op_msrs const * const msrs)
 		if (counter_config[i].enabled) {
 			multiplex[i].saved = -(u64)counter_config[i].count;
 		} else {
-			multiplex[i].addr  = 0;
 			multiplex[i].saved = 0;
 		}
 	}
@@ -189,25 +205,27 @@ static void nmi_cpu_setup_mux(int cpu, struct op_msrs const * const msrs)
 
 static void nmi_cpu_save_mpx_registers(struct op_msrs *msrs)
 {
+	struct op_msr *counters = msrs->counters;
 	struct op_msr *multiplex = msrs->multiplex;
 	int i;
 
 	for (i = 0; i < model->num_counters; ++i) {
 		int virt = op_x86_phys_to_virt(i);
-		if (multiplex[virt].addr)
-			rdmsrl(multiplex[virt].addr, multiplex[virt].saved);
+		if (counters[i].addr)
+			rdmsrl(counters[i].addr, multiplex[virt].saved);
 	}
 }
 
 static void nmi_cpu_restore_mpx_registers(struct op_msrs *msrs)
 {
+	struct op_msr *counters = msrs->counters;
 	struct op_msr *multiplex = msrs->multiplex;
 	int i;
 
 	for (i = 0; i < model->num_counters; ++i) {
 		int virt = op_x86_phys_to_virt(i);
-		if (multiplex[virt].addr)
-			wrmsrl(multiplex[virt].addr, multiplex[virt].saved);
+		if (counters[i].addr)
+			wrmsrl(counters[i].addr, multiplex[virt].saved);
 	}
 }
 
@@ -222,7 +240,7 @@ static void nmi_cpu_switch(void *dummy)
 
 	/* move to next set */
 	si += model->num_counters;
-	if ((si > model->num_virt_counters) || (counter_config[si].count == 0))
+	if ((si >= model->num_virt_counters) || (counter_config[si].count == 0))
 		per_cpu(switch_index, cpu) = 0;
 	else
 		per_cpu(switch_index, cpu) = si;
@@ -251,7 +269,10 @@ static int nmi_switch_event(void)
 	if (nmi_multiplex_on() < 0)
 		return -EINVAL;		/* not necessary */
 
-	on_each_cpu(nmi_cpu_switch, NULL, 1);
+	get_online_cpus();
+	if (ctr_running)
+		on_each_cpu(nmi_cpu_switch, NULL, 1);
+	put_online_cpus();
 
 	return 0;
 }
@@ -294,6 +315,7 @@ static void free_msrs(void)
 		kfree(per_cpu(cpu_msrs, i).controls);
 		per_cpu(cpu_msrs, i).controls = NULL;
 	}
+	nmi_shutdown_mux();
 }
 
 static int allocate_msrs(void)
@@ -303,17 +325,24 @@ static int allocate_msrs(void)
 
 	int i;
 	for_each_possible_cpu(i) {
-		per_cpu(cpu_msrs, i).counters = kmalloc(counters_size,
+		per_cpu(cpu_msrs, i).counters = kzalloc(counters_size,
 							GFP_KERNEL);
 		if (!per_cpu(cpu_msrs, i).counters)
-			return 0;
-		per_cpu(cpu_msrs, i).controls = kmalloc(controls_size,
+			goto fail;
+		per_cpu(cpu_msrs, i).controls = kzalloc(controls_size,
 							GFP_KERNEL);
 		if (!per_cpu(cpu_msrs, i).controls)
-			return 0;
+			goto fail;
 	}
 
+	if (!nmi_setup_mux())
+		goto fail;
+
 	return 1;
+
+fail:
+	free_msrs();
+	return 0;
 }
 
 static void nmi_cpu_setup(void *dummy)
@@ -335,49 +364,6 @@ static struct notifier_block profile_exceptions_nb = {
 	.priority = 2
 };
 
-static int nmi_setup(void)
-{
-	int err = 0;
-	int cpu;
-
-	if (!allocate_msrs())
-		err = -ENOMEM;
-	else if (!nmi_setup_mux())
-		err = -ENOMEM;
-	else
-		err = register_die_notifier(&profile_exceptions_nb);
-
-	if (err) {
-		free_msrs();
-		nmi_shutdown_mux();
-		return err;
-	}
-
-	/* We need to serialize save and setup for HT because the subset
-	 * of msrs are distinct for save and setup operations
-	 */
-
-	/* Assume saved/restored counters are the same on all CPUs */
-	model->fill_in_addresses(&per_cpu(cpu_msrs, 0));
-	for_each_possible_cpu(cpu) {
-		if (!cpu)
-			continue;
-
-		memcpy(per_cpu(cpu_msrs, cpu).counters,
-		       per_cpu(cpu_msrs, 0).counters,
-		       sizeof(struct op_msr) * model->num_counters);
-
-		memcpy(per_cpu(cpu_msrs, cpu).controls,
-		       per_cpu(cpu_msrs, 0).controls,
-		       sizeof(struct op_msr) * model->num_controls);
-
-		mux_clone(cpu);
-	}
-	on_each_cpu(nmi_cpu_setup, NULL, 1);
-	nmi_enabled = 1;
-	return 0;
-}
-
 static void nmi_cpu_restore_registers(struct op_msrs *msrs)
 {
 	struct op_msr *counters = msrs->counters;
@@ -411,20 +397,24 @@ static void nmi_cpu_shutdown(void *dummy)
 	apic_write(APIC_LVTPC, per_cpu(saved_lvtpc, cpu));
 	apic_write(APIC_LVTERR, v);
 	nmi_cpu_restore_registers(msrs);
+	if (model->cpu_down)
+		model->cpu_down();
 }
 
-static void nmi_shutdown(void)
+static void nmi_cpu_up(void *dummy)
 {
-	struct op_msrs *msrs;
+	if (nmi_enabled)
+		nmi_cpu_setup(dummy);
+	if (ctr_running)
+		nmi_cpu_start(dummy);
+}
 
-	nmi_enabled = 0;
-	on_each_cpu(nmi_cpu_shutdown, NULL, 1);
-	unregister_die_notifier(&profile_exceptions_nb);
-	nmi_shutdown_mux();
-	msrs = &get_cpu_var(cpu_msrs);
-	model->shutdown(msrs);
-	free_msrs();
-	put_cpu_var(cpu_msrs);
+static void nmi_cpu_down(void *dummy)
+{
+	if (ctr_running)
+		nmi_cpu_stop(dummy);
+	if (nmi_enabled)
+		nmi_cpu_shutdown(dummy);
 }
 
 static int nmi_create_files(struct super_block *sb, struct dentry *root)
@@ -456,7 +446,6 @@ static int nmi_create_files(struct super_block *sb, struct dentry *root)
 	return 0;
 }
 
-#ifdef CONFIG_SMP
 static int oprofile_cpu_notifier(struct notifier_block *b, unsigned long action,
 				 void *data)
 {
@@ -464,10 +453,10 @@ static int oprofile_cpu_notifier(struct notifier_block *b, unsigned long action,
 	switch (action) {
 	case CPU_DOWN_FAILED:
 	case CPU_ONLINE:
-		smp_call_function_single(cpu, nmi_cpu_start, NULL, 0);
+		smp_call_function_single(cpu, nmi_cpu_up, NULL, 0);
 		break;
 	case CPU_DOWN_PREPARE:
-		smp_call_function_single(cpu, nmi_cpu_stop, NULL, 1);
+		smp_call_function_single(cpu, nmi_cpu_down, NULL, 1);
 		break;
 	}
 	return NOTIFY_DONE;
@@ -476,7 +465,75 @@ static int oprofile_cpu_notifier(struct notifier_block *b, unsigned long action,
 static struct notifier_block oprofile_cpu_nb = {
 	.notifier_call = oprofile_cpu_notifier
 };
-#endif
+
+static int nmi_setup(void)
+{
+	int err = 0;
+	int cpu;
+
+	if (!allocate_msrs())
+		return -ENOMEM;
+
+	/* We need to serialize save and setup for HT because the subset
+	 * of msrs are distinct for save and setup operations
+	 */
+
+	/* Assume saved/restored counters are the same on all CPUs */
+	err = model->fill_in_addresses(&per_cpu(cpu_msrs, 0));
+	if (err)
+		goto fail;
+
+	for_each_possible_cpu(cpu) {
+		if (!cpu)
+			continue;
+
+		memcpy(per_cpu(cpu_msrs, cpu).counters,
+		       per_cpu(cpu_msrs, 0).counters,
+		       sizeof(struct op_msr) * model->num_counters);
+
+		memcpy(per_cpu(cpu_msrs, cpu).controls,
+		       per_cpu(cpu_msrs, 0).controls,
+		       sizeof(struct op_msr) * model->num_controls);
+
+		mux_clone(cpu);
+	}
+
+	nmi_enabled = 0;
+	ctr_running = 0;
+	barrier();
+	err = register_die_notifier(&profile_exceptions_nb);
+	if (err)
+		goto fail;
+
+	get_online_cpus();
+	register_cpu_notifier(&oprofile_cpu_nb);
+	on_each_cpu(nmi_cpu_setup, NULL, 1);
+	nmi_enabled = 1;
+	put_online_cpus();
+
+	return 0;
+fail:
+	free_msrs();
+	return err;
+}
+
+static void nmi_shutdown(void)
+{
+	struct op_msrs *msrs;
+
+	get_online_cpus();
+	unregister_cpu_notifier(&oprofile_cpu_nb);
+	on_each_cpu(nmi_cpu_shutdown, NULL, 1);
+	nmi_enabled = 0;
+	ctr_running = 0;
+	put_online_cpus();
+	barrier();
+	unregister_die_notifier(&profile_exceptions_nb);
+	msrs = &get_cpu_var(cpu_msrs);
+	model->shutdown(msrs);
+	free_msrs();
+	put_cpu_var(cpu_msrs);
+}
 
 #ifdef CONFIG_PM
 
@@ -598,6 +655,7 @@ static int __init ppro_init(char **cpu_type)
 	case 15: case 23:
 		*cpu_type = "i386/core_2";
 		break;
+	case 0x2e:
 	case 26:
 		spec = &op_arch_perfmon_spec;
 		*cpu_type = "i386/core_i7";
@@ -685,9 +743,6 @@ int __init op_nmi_init(struct oprofile_operations *ops)
 		return -ENODEV;
 	}
 
-#ifdef CONFIG_SMP
-	register_cpu_notifier(&oprofile_cpu_nb);
-#endif
 	/* default values, can be overwritten by model */
 	ops->create_files	= nmi_create_files;
 	ops->setup		= nmi_setup;
@@ -714,12 +769,6 @@ int __init op_nmi_init(struct oprofile_operations *ops)
 
 void op_nmi_exit(void)
 {
-	if (using_nmi) {
+	if (using_nmi)
 		exit_sysfs();
-#ifdef CONFIG_SMP
-		unregister_cpu_notifier(&oprofile_cpu_nb);
-#endif
-	}
-	if (model->exit)
-		model->exit();
 }