Merge branch 'pm-cpufreq'

* pm-cpufreq: (53 commits)
  cpufreq: speedstep-lib: Use monotonic clock
  cpufreq: powernv: Increase the verbosity of OCC console messages
  cpufreq: sfi: use kmemdup rather than duplicating its implementation
  cpufreq: drop !cpufreq_driver check from cpufreq_parse_governor()
  cpufreq: rename cpufreq_real_policy as cpufreq_user_policy
  cpufreq: remove redundant 'policy' field from user_policy
  cpufreq: remove redundant 'governor' field from user_policy
  cpufreq: update user_policy.* on success
  cpufreq: use memcpy() to copy policy
  cpufreq: remove redundant CPUFREQ_INCOMPATIBLE notifier event
  cpufreq: mediatek: Add MT8173 cpufreq driver
  dt-bindings: mediatek: Add MT8173 CPU DVFS clock bindings
  intel_pstate: append more Oracle OEM table id to vendor bypass list
  intel_pstate: Add SKY-S support
  intel_pstate: Fix possible overflow complained by Coverity
  cpufreq: Correct a freq check in cpufreq_set_policy()
  cpufreq: Lock CPU online/offline in cpufreq_register_driver()
  cpufreq: Replace recover_policy with new_policy in cpufreq_online()
  cpufreq: Separate CPU device registration from CPU online
  cpufreq: powernv: Restore cpu frequency to policy->cur on unthrottling
  ...

Documentation/cpu-freq/core.txt

@@ -55,16 +55,13 @@ transition notifiers.
 ----------------------------
 
 These are notified when a new policy is intended to be set. Each
-CPUFreq policy notifier is called three times for a policy transition:
+CPUFreq policy notifier is called twice for a policy transition:
 
 1.) During CPUFREQ_ADJUST all CPUFreq notifiers may change the limit if
     they see a need for this - may it be thermal considerations or
     hardware limitations.
 
-2.) During CPUFREQ_INCOMPATIBLE only changes may be done in order to avoid
+2.) And during CPUFREQ_NOTIFY all notifiers are informed of the new policy
-    hardware failure.
-
-3.) And during CPUFREQ_NOTIFY all notifiers are informed of the new policy
    - if two hardware drivers failed to agree on a new policy before this
    stage, the incompatible hardware shall be shut down, and the user
    informed of this.

Documentation/devicetree/bindings/clock/mt8173-cpu-dvfs.txt

@@ -0,0 +1,83 @@
+Device Tree Clock bindins for CPU DVFS of Mediatek MT8173 SoC
+
+Required properties:
+- clocks: A list of phandle + clock-specifier pairs for the clocks listed in clock names.
+- clock-names: Should contain the following:
+	"cpu"		- The multiplexer for clock input of CPU cluster.
+	"intermediate"	- A parent of "cpu" clock which is used as "intermediate" clock
+			  source (usually MAINPLL) when the original CPU PLL is under
+			  transition and not stable yet.
+	Please refer to Documentation/devicetree/bindings/clk/clock-bindings.txt for
+	generic clock consumer properties.
+- proc-supply: Regulator for Vproc of CPU cluster.
+
+Optional properties:
+- sram-supply: Regulator for Vsram of CPU cluster. When present, the cpufreq driver
+	       needs to do "voltage tracking" to step by step scale up/down Vproc and
+	       Vsram to fit SoC specific needs. When absent, the voltage scaling
+	       flow is handled by hardware, hence no software "voltage tracking" is
+	       needed.
+
+Example:
+--------
+	cpu0: cpu@0 {
+		device_type = "cpu";
+		compatible = "arm,cortex-a53";
+		reg = <0x000>;
+		enable-method = "psci";
+		cpu-idle-states = <&CPU_SLEEP_0>;
+		clocks = <&infracfg CLK_INFRA_CA53SEL>,
+			 <&apmixedsys CLK_APMIXED_MAINPLL>;
+		clock-names = "cpu", "intermediate";
+	};
+
+	cpu1: cpu@1 {
+		device_type = "cpu";
+		compatible = "arm,cortex-a53";
+		reg = <0x001>;
+		enable-method = "psci";
+		cpu-idle-states = <&CPU_SLEEP_0>;
+		clocks = <&infracfg CLK_INFRA_CA53SEL>,
+			 <&apmixedsys CLK_APMIXED_MAINPLL>;
+		clock-names = "cpu", "intermediate";
+	};
+
+	cpu2: cpu@100 {
+		device_type = "cpu";
+		compatible = "arm,cortex-a57";
+		reg = <0x100>;
+		enable-method = "psci";
+		cpu-idle-states = <&CPU_SLEEP_0>;
+		clocks = <&infracfg CLK_INFRA_CA57SEL>,
+			 <&apmixedsys CLK_APMIXED_MAINPLL>;
+		clock-names = "cpu", "intermediate";
+	};
+
+	cpu3: cpu@101 {
+		device_type = "cpu";
+		compatible = "arm,cortex-a57";
+		reg = <0x101>;
+		enable-method = "psci";
+		cpu-idle-states = <&CPU_SLEEP_0>;
+		clocks = <&infracfg CLK_INFRA_CA57SEL>,
+			 <&apmixedsys CLK_APMIXED_MAINPLL>;
+		clock-names = "cpu", "intermediate";
+	};
+
+	&cpu0 {
+		proc-supply = <&mt6397_vpca15_reg>;
+	};
+
+	&cpu1 {
+		proc-supply = <&mt6397_vpca15_reg>;
+	};
+
+	&cpu2 {
+		proc-supply = <&da9211_vcpu_reg>;
+		sram-supply = <&mt6397_vsramca7_reg>;
+	};
+
+	&cpu3 {
+		proc-supply = <&da9211_vcpu_reg>;
+		sram-supply = <&mt6397_vsramca7_reg>;
+	};

arch/powerpc/include/asm/opal-api.h

@@ -361,6 +361,7 @@ enum opal_msg_type {
 	OPAL_MSG_HMI_EVT,
 	OPAL_MSG_DPO,
 	OPAL_MSG_PRD,
+	OPAL_MSG_OCC,
 	OPAL_MSG_TYPE_MAX,
 };
 
@@ -700,6 +701,17 @@ struct opal_prd_msg_header {
 
 struct opal_prd_msg;
 
+#define OCC_RESET                       0
+#define OCC_LOAD                        1
+#define OCC_THROTTLE                    2
+#define OCC_MAX_THROTTLE_STATUS         5
+
+struct opal_occ_msg {
+	__be64 type;
+	__be64 chip;
+	__be64 throttle_status;
+};
+
 /*
  * SG entries
  *

drivers/acpi/processor_perflib.c

@@ -83,7 +83,7 @@ static int acpi_processor_ppc_notifier(struct notifier_block *nb,
 	if (ignore_ppc)
 		return 0;
 
-	if (event != CPUFREQ_INCOMPATIBLE)
+	if (event != CPUFREQ_ADJUST)
 		return 0;
 
 	mutex_lock(&performance_mutex);
@@ -780,9 +780,7 @@ acpi_processor_register_performance(struct acpi_processor_performance
 
 EXPORT_SYMBOL(acpi_processor_register_performance);
 
-void
+void acpi_processor_unregister_performance(unsigned int cpu)
-acpi_processor_unregister_performance(struct acpi_processor_performance
-				      *performance, unsigned int cpu)
 {
 	struct acpi_processor *pr;
 

drivers/cpufreq/Kconfig.arm

@@ -130,6 +130,13 @@ config ARM_KIRKWOOD_CPUFREQ
 	  This adds the CPUFreq driver for Marvell Kirkwood
 	  SoCs.
 
+config ARM_MT8173_CPUFREQ
+	bool "Mediatek MT8173 CPUFreq support"
+	depends on ARCH_MEDIATEK && REGULATOR
+	select PM_OPP
+	help
+	  This adds the CPUFreq driver support for Mediatek MT8173 SoC.
+
 config ARM_OMAP2PLUS_CPUFREQ
 	bool "TI OMAP2+"
 	depends on ARCH_OMAP2PLUS

drivers/cpufreq/Makefile

@@ -62,6 +62,7 @@ obj-$(CONFIG_ARM_HISI_ACPU_CPUFREQ)	+= hisi-acpu-cpufreq.o
 obj-$(CONFIG_ARM_IMX6Q_CPUFREQ)		+= imx6q-cpufreq.o
 obj-$(CONFIG_ARM_INTEGRATOR)		+= integrator-cpufreq.o
 obj-$(CONFIG_ARM_KIRKWOOD_CPUFREQ)	+= kirkwood-cpufreq.o
+obj-$(CONFIG_ARM_MT8173_CPUFREQ)	+= mt8173-cpufreq.o
 obj-$(CONFIG_ARM_OMAP2PLUS_CPUFREQ)	+= omap-cpufreq.o
 obj-$(CONFIG_ARM_PXA2xx_CPUFREQ)	+= pxa2xx-cpufreq.o
 obj-$(CONFIG_PXA3xx)			+= pxa3xx-cpufreq.o

drivers/cpufreq/acpi-cpufreq.c

@@ -65,18 +65,21 @@ enum {
 #define MSR_K7_HWCR_CPB_DIS	(1ULL << 25)
 
 struct acpi_cpufreq_data {
-	struct acpi_processor_performance *acpi_data;
 	struct cpufreq_frequency_table *freq_table;
 	unsigned int resume;
 	unsigned int cpu_feature;
+	unsigned int acpi_perf_cpu;
 	cpumask_var_t freqdomain_cpus;
 };
 
-static DEFINE_PER_CPU(struct acpi_cpufreq_data *, acfreq_data);
-
 /* acpi_perf_data is a pointer to percpu data. */
 static struct acpi_processor_performance __percpu *acpi_perf_data;
 
+static inline struct acpi_processor_performance *to_perf_data(struct acpi_cpufreq_data *data)
+{
+	return per_cpu_ptr(acpi_perf_data, data->acpi_perf_cpu);
+}
+
 static struct cpufreq_driver acpi_cpufreq_driver;
 
 static unsigned int acpi_pstate_strict;
@@ -144,7 +147,7 @@ static int _store_boost(int val)
 
 static ssize_t show_freqdomain_cpus(struct cpufreq_policy *policy, char *buf)
 {
-	struct acpi_cpufreq_data *data = per_cpu(acfreq_data, policy->cpu);
+	struct acpi_cpufreq_data *data = policy->driver_data;
 
 	return cpufreq_show_cpus(data->freqdomain_cpus, buf);
 }
@@ -202,7 +205,7 @@ static unsigned extract_io(u32 value, struct acpi_cpufreq_data *data)
 	struct acpi_processor_performance *perf;
 	int i;
 
-	perf = data->acpi_data;
+	perf = to_perf_data(data);
 
 	for (i = 0; i < perf->state_count; i++) {
 		if (value == perf->states[i].status)
@@ -221,7 +224,7 @@ static unsigned extract_msr(u32 msr, struct acpi_cpufreq_data *data)
 	else
 		msr &= INTEL_MSR_RANGE;
 
-	perf = data->acpi_data;
+	perf = to_perf_data(data);
 
 	cpufreq_for_each_entry(pos, data->freq_table)
 		if (msr == perf->states[pos->driver_data].status)
@@ -327,7 +330,8 @@ static void drv_write(struct drv_cmd *cmd)
 	put_cpu();
 }
 
-static u32 get_cur_val(const struct cpumask *mask)
+static u32
+get_cur_val(const struct cpumask *mask, struct acpi_cpufreq_data *data)
 {
 	struct acpi_processor_performance *perf;
 	struct drv_cmd cmd;
@@ -335,7 +339,7 @@ static u32 get_cur_val(const struct cpumask *mask)
 	if (unlikely(cpumask_empty(mask)))
 		return 0;
 
-	switch (per_cpu(acfreq_data, cpumask_first(mask))->cpu_feature) {
+	switch (data->cpu_feature) {
 	case SYSTEM_INTEL_MSR_CAPABLE:
 		cmd.type = SYSTEM_INTEL_MSR_CAPABLE;
 		cmd.addr.msr.reg = MSR_IA32_PERF_CTL;
@@ -346,7 +350,7 @@ static u32 get_cur_val(const struct cpumask *mask)
 		break;
 	case SYSTEM_IO_CAPABLE:
 		cmd.type = SYSTEM_IO_CAPABLE;
-		perf = per_cpu(acfreq_data, cpumask_first(mask))->acpi_data;
+		perf = to_perf_data(data);
 		cmd.addr.io.port = perf->control_register.address;
 		cmd.addr.io.bit_width = perf->control_register.bit_width;
 		break;
@@ -364,19 +368,24 @@ static u32 get_cur_val(const struct cpumask *mask)
 
 static unsigned int get_cur_freq_on_cpu(unsigned int cpu)
 {
-	struct acpi_cpufreq_data *data = per_cpu(acfreq_data, cpu);
+	struct acpi_cpufreq_data *data;
+	struct cpufreq_policy *policy;
 	unsigned int freq;
 	unsigned int cached_freq;
 
 	pr_debug("get_cur_freq_on_cpu (%d)\n", cpu);
 
-	if (unlikely(data == NULL ||
+	policy = cpufreq_cpu_get(cpu);
-		     data->acpi_data == NULL || data->freq_table == NULL)) {
+	if (unlikely(!policy))
 		return 0;
-	}
 
-	cached_freq = data->freq_table[data->acpi_data->state].frequency;
+	data = policy->driver_data;
-	freq = extract_freq(get_cur_val(cpumask_of(cpu)), data);
+	cpufreq_cpu_put(policy);
+	if (unlikely(!data || !data->freq_table))
+		return 0;
+
+	cached_freq = data->freq_table[to_perf_data(data)->state].frequency;
+	freq = extract_freq(get_cur_val(cpumask_of(cpu), data), data);
 	if (freq != cached_freq) {
 		/*
 		 * The dreaded BIOS frequency change behind our back.
@@ -397,7 +406,7 @@ static unsigned int check_freqs(const struct cpumask *mask, unsigned int freq,
 	unsigned int i;
 
 	for (i = 0; i < 100; i++) {
-		cur_freq = extract_freq(get_cur_val(mask), data);
+		cur_freq = extract_freq(get_cur_val(mask, data), data);
 		if (cur_freq == freq)
 			return 1;
 		udelay(10);
@@ -408,18 +417,17 @@ static unsigned int check_freqs(const struct cpumask *mask, unsigned int freq,
 static int acpi_cpufreq_target(struct cpufreq_policy *policy,
 			       unsigned int index)
 {
-	struct acpi_cpufreq_data *data = per_cpu(acfreq_data, policy->cpu);
+	struct acpi_cpufreq_data *data = policy->driver_data;
 	struct acpi_processor_performance *perf;
 	struct drv_cmd cmd;
 	unsigned int next_perf_state = 0; /* Index into perf table */
 	int result = 0;
 
-	if (unlikely(data == NULL ||
+	if (unlikely(data == NULL || data->freq_table == NULL)) {
-	     data->acpi_data == NULL || data->freq_table == NULL)) {
 		return -ENODEV;
 	}
 
-	perf = data->acpi_data;
+	perf = to_perf_data(data);
 	next_perf_state = data->freq_table[index].driver_data;
 	if (perf->state == next_perf_state) {
 		if (unlikely(data->resume)) {
@@ -482,8 +490,9 @@ out:
 static unsigned long
 acpi_cpufreq_guess_freq(struct acpi_cpufreq_data *data, unsigned int cpu)
 {
-	struct acpi_processor_performance *perf = data->acpi_data;
+	struct acpi_processor_performance *perf;
 
+	perf = to_perf_data(data);
 	if (cpu_khz) {
 		/* search the closest match to cpu_khz */
 		unsigned int i;
@@ -672,17 +681,17 @@ static int acpi_cpufreq_cpu_init(struct cpufreq_policy *policy)
 		goto err_free;
 	}
 
-	data->acpi_data = per_cpu_ptr(acpi_perf_data, cpu);
+	perf = per_cpu_ptr(acpi_perf_data, cpu);
-	per_cpu(acfreq_data, cpu) = data;
+	data->acpi_perf_cpu = cpu;
+	policy->driver_data = data;
 
 	if (cpu_has(c, X86_FEATURE_CONSTANT_TSC))
 		acpi_cpufreq_driver.flags |= CPUFREQ_CONST_LOOPS;
 
-	result = acpi_processor_register_performance(data->acpi_data, cpu);
+	result = acpi_processor_register_performance(perf, cpu);
 	if (result)
 		goto err_free_mask;
 
-	perf = data->acpi_data;
 	policy->shared_type = perf->shared_type;
 
 	/*
@@ -838,26 +847,25 @@ static int acpi_cpufreq_cpu_init(struct cpufreq_policy *policy)
 err_freqfree:
 	kfree(data->freq_table);
 err_unreg:
-	acpi_processor_unregister_performance(perf, cpu);
+	acpi_processor_unregister_performance(cpu);
 err_free_mask:
 	free_cpumask_var(data->freqdomain_cpus);
 err_free:
 	kfree(data);
-	per_cpu(acfreq_data, cpu) = NULL;
+	policy->driver_data = NULL;
 
 	return result;
 }
 
 static int acpi_cpufreq_cpu_exit(struct cpufreq_policy *policy)
 {
-	struct acpi_cpufreq_data *data = per_cpu(acfreq_data, policy->cpu);
+	struct acpi_cpufreq_data *data = policy->driver_data;
 
 	pr_debug("acpi_cpufreq_cpu_exit\n");
 
 	if (data) {
-		per_cpu(acfreq_data, policy->cpu) = NULL;
+		policy->driver_data = NULL;
-		acpi_processor_unregister_performance(data->acpi_data,
+		acpi_processor_unregister_performance(data->acpi_perf_cpu);
-						      policy->cpu);
 		free_cpumask_var(data->freqdomain_cpus);
 		kfree(data->freq_table);
 		kfree(data);
@@ -868,7 +876,7 @@ static int acpi_cpufreq_cpu_exit(struct cpufreq_policy *policy)
 
 static int acpi_cpufreq_resume(struct cpufreq_policy *policy)
 {
-	struct acpi_cpufreq_data *data = per_cpu(acfreq_data, policy->cpu);
+	struct acpi_cpufreq_data *data = policy->driver_data;
 
 	pr_debug("acpi_cpufreq_resume\n");
 
@@ -880,7 +888,9 @@ static int acpi_cpufreq_resume(struct cpufreq_policy *policy)
 static struct freq_attr *acpi_cpufreq_attr[] = {
 	&cpufreq_freq_attr_scaling_available_freqs,
 	&freqdomain_cpus,
-	NULL,	/* this is a placeholder for cpb, do not remove */
+#ifdef CONFIG_X86_ACPI_CPUFREQ_CPB
+	&cpb,
+#endif
 	NULL,
 };
 
@@ -953,17 +963,16 @@ static int __init acpi_cpufreq_init(void)
 	 * only if configured. This is considered legacy code, which
 	 * will probably be removed at some point in the future.
 	 */
-	if (check_amd_hwpstate_cpu(0)) {
+	if (!check_amd_hwpstate_cpu(0)) {
-		struct freq_attr **iter;
+		struct freq_attr **attr;
 
-		pr_debug("adding sysfs entry for cpb\n");
+		pr_debug("CPB unsupported, do not expose it\n");
 
-		for (iter = acpi_cpufreq_attr; *iter != NULL; iter++)
+		for (attr = acpi_cpufreq_attr; *attr; attr++)
-			;
+			if (*attr == &cpb) {
-
+				*attr = NULL;
-		/* make sure there is a terminator behind it */
+				break;
-		if (iter[1] == NULL)
+			}
-			*iter = &cpb;
 	}
 #endif
 	acpi_cpufreq_boost_init();

drivers/cpufreq/cpufreq.c

@@ -112,12 +112,6 @@ static inline bool has_target(void)
 	return cpufreq_driver->target_index || cpufreq_driver->target;
 }
 
-/*
- * rwsem to guarantee that cpufreq driver module doesn't unload during critical
- * sections
- */
-static DECLARE_RWSEM(cpufreq_rwsem);
-
 /* internal prototypes */
 static int __cpufreq_governor(struct cpufreq_policy *policy,
 		unsigned int event);
@@ -277,10 +271,6 @@ EXPORT_SYMBOL_GPL(cpufreq_generic_get);
  * If corresponding call cpufreq_cpu_put() isn't made, the policy wouldn't be
  * freed as that depends on the kobj count.
  *
- * It also takes a read-lock of 'cpufreq_rwsem' and doesn't put it back if a
- * valid policy is found. This is done to make sure the driver doesn't get
- * unregistered while the policy is being used.
- *
  * Return: A valid policy on success, otherwise NULL on failure.
  */
 struct cpufreq_policy *cpufreq_cpu_get(unsigned int cpu)
@@ -291,9 +281,6 @@ struct cpufreq_policy *cpufreq_cpu_get(unsigned int cpu)
 	if (WARN_ON(cpu >= nr_cpu_ids))
 		return NULL;
 
-	if (!down_read_trylock(&cpufreq_rwsem))
-		return NULL;
-
 	/* get the cpufreq driver */
 	read_lock_irqsave(&cpufreq_driver_lock, flags);
 
@@ -306,9 +293,6 @@ struct cpufreq_policy *cpufreq_cpu_get(unsigned int cpu)
 
 	read_unlock_irqrestore(&cpufreq_driver_lock, flags);
 
-	if (!policy)
-		up_read(&cpufreq_rwsem);
-
 	return policy;
 }
 EXPORT_SYMBOL_GPL(cpufreq_cpu_get);
@@ -320,13 +304,10 @@ EXPORT_SYMBOL_GPL(cpufreq_cpu_get);
  *
  * This decrements the kobject reference count incremented earlier by calling
  * cpufreq_cpu_get().
- *
- * It also drops the read-lock of 'cpufreq_rwsem' taken at cpufreq_cpu_get().
  */
 void cpufreq_cpu_put(struct cpufreq_policy *policy)
 {
 	kobject_put(&policy->kobj);
-	up_read(&cpufreq_rwsem);
 }
 EXPORT_SYMBOL_GPL(cpufreq_cpu_put);
 
@@ -539,9 +520,6 @@ static int cpufreq_parse_governor(char *str_governor, unsigned int *policy,
 {
 	int err = -EINVAL;
 
-	if (!cpufreq_driver)
-		goto out;
-
 	if (cpufreq_driver->setpolicy) {
 		if (!strncasecmp(str_governor, "performance", CPUFREQ_NAME_LEN)) {
 			*policy = CPUFREQ_POLICY_PERFORMANCE;
@@ -576,7 +554,6 @@ static int cpufreq_parse_governor(char *str_governor, unsigned int *policy,
 
 		mutex_unlock(&cpufreq_governor_mutex);
 	}
-out:
 	return err;
 }
 
@@ -625,9 +602,7 @@ static ssize_t store_##file_name					\
 	int ret, temp;							\
 	struct cpufreq_policy new_policy;				\
 									\
-	ret = cpufreq_get_policy(&new_policy, policy->cpu);		\
+	memcpy(&new_policy, policy, sizeof(*policy));			\
-	if (ret)							\
-		return -EINVAL;						\
 									\
 	ret = sscanf(buf, "%u", &new_policy.object);			\
 	if (ret != 1)							\
@@ -681,9 +656,7 @@ static ssize_t store_scaling_governor(struct cpufreq_policy *policy,
 	char	str_governor[16];
 	struct cpufreq_policy new_policy;
 
-	ret = cpufreq_get_policy(&new_policy, policy->cpu);
+	memcpy(&new_policy, policy, sizeof(*policy));
-	if (ret)
-		return ret;
 
 	ret = sscanf(buf, "%15s", str_governor);
 	if (ret != 1)
@@ -694,14 +667,7 @@ static ssize_t store_scaling_governor(struct cpufreq_policy *policy,
 		return -EINVAL;
 
 	ret = cpufreq_set_policy(policy, &new_policy);
-
+	return ret ? ret : count;
-	policy->user_policy.policy = policy->policy;
-	policy->user_policy.governor = policy->governor;
-
-	if (ret)
-		return ret;
-	else
-		return count;
 }
 
 /**
@@ -851,9 +817,6 @@ static ssize_t show(struct kobject *kobj, struct attribute *attr, char *buf)
 	struct freq_attr *fattr = to_attr(attr);
 	ssize_t ret;
 
-	if (!down_read_trylock(&cpufreq_rwsem))
-		return -EINVAL;
-
 	down_read(&policy->rwsem);
 
 	if (fattr->show)
@@ -862,7 +825,6 @@ static ssize_t show(struct kobject *kobj, struct attribute *attr, char *buf)
 		ret = -EIO;
 
 	up_read(&policy->rwsem);
-	up_read(&cpufreq_rwsem);
 
 	return ret;
 }
@@ -879,9 +841,6 @@ static ssize_t store(struct kobject *kobj, struct attribute *attr,
 	if (!cpu_online(policy->cpu))
 		goto unlock;
 
-	if (!down_read_trylock(&cpufreq_rwsem))
-		goto unlock;
-
 	down_write(&policy->rwsem);
 
 	/* Updating inactive policies is invalid, so avoid doing that. */
@@ -897,8 +856,6 @@ static ssize_t store(struct kobject *kobj, struct attribute *attr,
 
 unlock_policy_rwsem:
 	up_write(&policy->rwsem);
-
-	up_read(&cpufreq_rwsem);
 unlock:
 	put_online_cpus();
 
@@ -1027,8 +984,7 @@ static void cpufreq_remove_dev_symlink(struct cpufreq_policy *policy)
 	}
 }
 
-static int cpufreq_add_dev_interface(struct cpufreq_policy *policy,
+static int cpufreq_add_dev_interface(struct cpufreq_policy *policy)
-				     struct device *dev)
 {
 	struct freq_attr **drv_attr;
 	int ret = 0;
@@ -1060,11 +1016,10 @@ static int cpufreq_add_dev_interface(struct cpufreq_policy *policy,
 	return cpufreq_add_dev_symlink(policy);
 }
 
-static void cpufreq_init_policy(struct cpufreq_policy *policy)
+static int cpufreq_init_policy(struct cpufreq_policy *policy)
 {
 	struct cpufreq_governor *gov = NULL;
 	struct cpufreq_policy new_policy;
-	int ret = 0;
 
 	memcpy(&new_policy, policy, sizeof(*policy));
 
@@ -1083,16 +1038,10 @@ static void cpufreq_init_policy(struct cpufreq_policy *policy)
 		cpufreq_parse_governor(gov->name, &new_policy.policy, NULL);
 
 	/* set default policy */
-	ret = cpufreq_set_policy(policy, &new_policy);
+	return cpufreq_set_policy(policy, &new_policy);
-	if (ret) {
-		pr_debug("setting policy failed\n");
-		if (cpufreq_driver->exit)
-			cpufreq_driver->exit(policy);
-	}
 }
 
-static int cpufreq_add_policy_cpu(struct cpufreq_policy *policy,
+static int cpufreq_add_policy_cpu(struct cpufreq_policy *policy, unsigned int cpu)
-				  unsigned int cpu, struct device *dev)
 {
 	int ret = 0;
 
@@ -1126,33 +1075,15 @@ static int cpufreq_add_policy_cpu(struct cpufreq_policy *policy,
 	return 0;
 }
 
-static struct cpufreq_policy *cpufreq_policy_restore(unsigned int cpu)
+static struct cpufreq_policy *cpufreq_policy_alloc(unsigned int cpu)
-{
-	struct cpufreq_policy *policy;
-	unsigned long flags;
-
-	read_lock_irqsave(&cpufreq_driver_lock, flags);
-	policy = per_cpu(cpufreq_cpu_data, cpu);
-	read_unlock_irqrestore(&cpufreq_driver_lock, flags);
-
-	if (likely(policy)) {
-		/* Policy should be inactive here */
-		WARN_ON(!policy_is_inactive(policy));
-
-		down_write(&policy->rwsem);
-		policy->cpu = cpu;
-		policy->governor = NULL;
-		up_write(&policy->rwsem);
-	}
-
-	return policy;
-}
-
-static struct cpufreq_policy *cpufreq_policy_alloc(struct device *dev)
 {
+	struct device *dev = get_cpu_device(cpu);
 	struct cpufreq_policy *policy;
 	int ret;
 
+	if (WARN_ON(!dev))
+		return NULL;
+
 	policy = kzalloc(sizeof(*policy), GFP_KERNEL);
 	if (!policy)
 		return NULL;
@@ -1180,10 +1111,10 @@ static struct cpufreq_policy *cpufreq_policy_alloc(struct device *dev)
 	init_completion(&policy->kobj_unregister);
 	INIT_WORK(&policy->update, handle_update);
 
-	policy->cpu = dev->id;
+	policy->cpu = cpu;
 
 	/* Set this once on allocation */
-	policy->kobj_cpu = dev->id;
+	policy->kobj_cpu = cpu;
 
 	return policy;
 
@@ -1245,59 +1176,34 @@ static void cpufreq_policy_free(struct cpufreq_policy *policy, bool notify)
 	kfree(policy);
 }
 
-/**
+static int cpufreq_online(unsigned int cpu)
- * cpufreq_add_dev - add a CPU device
- *
- * Adds the cpufreq interface for a CPU device.
- *
- * The Oracle says: try running cpufreq registration/unregistration concurrently
- * with with cpu hotplugging and all hell will break loose. Tried to clean this
- * mess up, but more thorough testing is needed. - Mathieu
- */
-static int cpufreq_add_dev(struct device *dev, struct subsys_interface *sif)
 {
-	unsigned int j, cpu = dev->id;
-	int ret = -ENOMEM;
 	struct cpufreq_policy *policy;
+	bool new_policy;
 	unsigned long flags;
-	bool recover_policy = !sif;
+	unsigned int j;
+	int ret;
 
-	pr_debug("adding CPU %u\n", cpu);
+	pr_debug("%s: bringing CPU%u online\n", __func__, cpu);
-
-	if (cpu_is_offline(cpu)) {
-		/*
-		 * Only possible if we are here from the subsys_interface add
-		 * callback.  A hotplug notifier will follow and we will handle
-		 * it as CPU online then.  For now, just create the sysfs link,
-		 * unless there is no policy or the link is already present.
-		 */
-		policy = per_cpu(cpufreq_cpu_data, cpu);
-		return policy && !cpumask_test_and_set_cpu(cpu, policy->real_cpus)
-			? add_cpu_dev_symlink(policy, cpu) : 0;
-	}
-
-	if (!down_read_trylock(&cpufreq_rwsem))
-		return 0;
 
 	/* Check if this CPU already has a policy to manage it */
 	policy = per_cpu(cpufreq_cpu_data, cpu);
-	if (policy && !policy_is_inactive(policy)) {
+	if (policy) {
 		WARN_ON(!cpumask_test_cpu(cpu, policy->related_cpus));
-		ret = cpufreq_add_policy_cpu(policy, cpu, dev);
+		if (!policy_is_inactive(policy))
-		up_read(&cpufreq_rwsem);
+			return cpufreq_add_policy_cpu(policy, cpu);
-		return ret;
-	}
 
-	/*
+		/* This is the only online CPU for the policy.  Start over. */
-	 * Restore the saved policy when doing light-weight init and fall back
+		new_policy = false;
-	 * to the full init if that fails.
+		down_write(&policy->rwsem);
-	 */
+		policy->cpu = cpu;
-	policy = recover_policy ? cpufreq_policy_restore(cpu) : NULL;
+		policy->governor = NULL;
-	if (!policy) {
+		up_write(&policy->rwsem);
-		recover_policy = false;
+	} else {
-		policy = cpufreq_policy_alloc(dev);
+		new_policy = true;
+		policy = cpufreq_policy_alloc(cpu);
 		if (!policy)
-			goto nomem_out;
+			return -ENOMEM;
 	}
 
 	cpumask_copy(policy->cpus, cpumask_of(cpu));
@@ -1308,17 +1214,17 @@ static int cpufreq_add_dev(struct device *dev, struct subsys_interface *sif)
 	ret = cpufreq_driver->init(policy);
 	if (ret) {
 		pr_debug("initialization failed\n");
-		goto err_set_policy_cpu;
+		goto out_free_policy;
 	}
 
 	down_write(&policy->rwsem);
 
-	/* related cpus should atleast have policy->cpus */
+	if (new_policy) {
-	cpumask_or(policy->related_cpus, policy->related_cpus, policy->cpus);
+		/* related_cpus should at least include policy->cpus. */
-
+		cpumask_or(policy->related_cpus, policy->related_cpus, policy->cpus);
-	/* Remember which CPUs have been present at the policy creation time. */
+		/* Remember CPUs present at the policy creation time. */
-	if (!recover_policy)
 		cpumask_and(policy->real_cpus, policy->cpus, cpu_present_mask);
+	}
 
 	/*
 	 * affected cpus must always be the one, which are online. We aren't
@@ -1326,7 +1232,7 @@ static int cpufreq_add_dev(struct device *dev, struct subsys_interface *sif)
 	 */
 	cpumask_and(policy->cpus, policy->cpus, cpu_online_mask);
 
-	if (!recover_policy) {
+	if (new_policy) {
 		policy->user_policy.min = policy->min;
 		policy->user_policy.max = policy->max;
 
@@ -1340,7 +1246,7 @@ static int cpufreq_add_dev(struct device *dev, struct subsys_interface *sif)
 		policy->cur = cpufreq_driver->get(policy->cpu);
 		if (!policy->cur) {
 			pr_err("%s: ->get() failed\n", __func__);
-			goto err_get_freq;
+			goto out_exit_policy;
 		}
 	}
 
@@ -1387,10 +1293,10 @@ static int cpufreq_add_dev(struct device *dev, struct subsys_interface *sif)
 	blocking_notifier_call_chain(&cpufreq_policy_notifier_list,
 				     CPUFREQ_START, policy);
 
-	if (!recover_policy) {
+	if (new_policy) {
-		ret = cpufreq_add_dev_interface(policy, dev);
+		ret = cpufreq_add_dev_interface(policy);
 		if (ret)
-			goto err_out_unregister;
+			goto out_exit_policy;
 		blocking_notifier_call_chain(&cpufreq_policy_notifier_list,
 				CPUFREQ_CREATE_POLICY, policy);
 
@@ -1399,18 +1305,19 @@ static int cpufreq_add_dev(struct device *dev, struct subsys_interface *sif)
 		write_unlock_irqrestore(&cpufreq_driver_lock, flags);
 	}
 
-	cpufreq_init_policy(policy);
+	ret = cpufreq_init_policy(policy);
-
+	if (ret) {
-	if (!recover_policy) {
+		pr_err("%s: Failed to initialize policy for cpu: %d (%d)\n",
-		policy->user_policy.policy = policy->policy;
+		       __func__, cpu, ret);
-		policy->user_policy.governor = policy->governor;
+		/* cpufreq_policy_free() will notify based on this */
+		new_policy = false;
+		goto out_exit_policy;
 	}
+
 	up_write(&policy->rwsem);
 
 	kobject_uevent(&policy->kobj, KOBJ_ADD);
 
-	up_read(&cpufreq_rwsem);
-
 	/* Callback for handling stuff after policy is ready */
 	if (cpufreq_driver->ready)
 		cpufreq_driver->ready(policy);
@@ -1419,24 +1326,47 @@ static int cpufreq_add_dev(struct device *dev, struct subsys_interface *sif)
 
 	return 0;
 
-err_out_unregister:
+out_exit_policy:
-err_get_freq:
 	up_write(&policy->rwsem);
 
 	if (cpufreq_driver->exit)
 		cpufreq_driver->exit(policy);
-err_set_policy_cpu:
+out_free_policy:
-	cpufreq_policy_free(policy, recover_policy);
+	cpufreq_policy_free(policy, !new_policy);
-nomem_out:
+	return ret;
-	up_read(&cpufreq_rwsem);
+}
+
+/**
+ * cpufreq_add_dev - the cpufreq interface for a CPU device.
+ * @dev: CPU device.
+ * @sif: Subsystem interface structure pointer (not used)
+ */
+static int cpufreq_add_dev(struct device *dev, struct subsys_interface *sif)
+{
+	unsigned cpu = dev->id;
+	int ret;
+
+	dev_dbg(dev, "%s: adding CPU%u\n", __func__, cpu);
+
+	if (cpu_online(cpu)) {
+		ret = cpufreq_online(cpu);
+	} else {
+		/*
+		 * A hotplug notifier will follow and we will handle it as CPU
+		 * online then.  For now, just create the sysfs link, unless
+		 * there is no policy or the link is already present.
+		 */
+		struct cpufreq_policy *policy = per_cpu(cpufreq_cpu_data, cpu);
+
+		ret = policy && !cpumask_test_and_set_cpu(cpu, policy->real_cpus)
+			? add_cpu_dev_symlink(policy, cpu) : 0;
+	}
 
 	return ret;
 }
 
-static int __cpufreq_remove_dev_prepare(struct device *dev)
+static void cpufreq_offline_prepare(unsigned int cpu)
 {
-	unsigned int cpu = dev->id;
-	int ret = 0;
 	struct cpufreq_policy *policy;
 
 	pr_debug("%s: unregistering CPU %u\n", __func__, cpu);
@@ -1444,11 +1374,11 @@ static int __cpufreq_remove_dev_prepare(struct device *dev)
 	policy = cpufreq_cpu_get_raw(cpu);
 	if (!policy) {
 		pr_debug("%s: No cpu_data found\n", __func__);
-		return -EINVAL;
+		return;
 	}
 
 	if (has_target()) {
-		ret = __cpufreq_governor(policy, CPUFREQ_GOV_STOP);
+		int ret = __cpufreq_governor(policy, CPUFREQ_GOV_STOP);
 		if (ret)
 			pr_err("%s: Failed to stop governor\n", __func__);
 	}
@@ -1469,7 +1399,7 @@ static int __cpufreq_remove_dev_prepare(struct device *dev)
 	/* Start governor again for active policy */
 	if (!policy_is_inactive(policy)) {
 		if (has_target()) {
-			ret = __cpufreq_governor(policy, CPUFREQ_GOV_START);
+			int ret = __cpufreq_governor(policy, CPUFREQ_GOV_START);
 			if (!ret)
 				ret = __cpufreq_governor(policy, CPUFREQ_GOV_LIMITS);
 
@@ -1479,28 +1409,24 @@ static int __cpufreq_remove_dev_prepare(struct device *dev)
 	} else if (cpufreq_driver->stop_cpu) {
 		cpufreq_driver->stop_cpu(policy);
 	}
-
-	return ret;
 }
 
-static int __cpufreq_remove_dev_finish(struct device *dev)
+static void cpufreq_offline_finish(unsigned int cpu)
 {
-	unsigned int cpu = dev->id;
-	int ret;
 	struct cpufreq_policy *policy = per_cpu(cpufreq_cpu_data, cpu);
 
 	if (!policy) {
 		pr_debug("%s: No cpu_data found\n", __func__);
-		return -EINVAL;
+		return;
 	}
 
 	/* Only proceed for inactive policies */
 	if (!policy_is_inactive(policy))
-		return 0;
+		return;
 
 	/* If cpu is last user of policy, free policy */
 	if (has_target()) {
-		ret = __cpufreq_governor(policy, CPUFREQ_GOV_POLICY_EXIT);
+		int ret = __cpufreq_governor(policy, CPUFREQ_GOV_POLICY_EXIT);
 		if (ret)
 			pr_err("%s: Failed to exit governor\n", __func__);
 	}
@@ -1512,8 +1438,6 @@ static int __cpufreq_remove_dev_finish(struct device *dev)
 	 */
 	if (cpufreq_driver->exit)
 		cpufreq_driver->exit(policy);
-
-	return 0;
 }
 
 /**
@@ -1530,8 +1454,8 @@ static int cpufreq_remove_dev(struct device *dev, struct subsys_interface *sif)
 		return 0;
 
 	if (cpu_online(cpu)) {
-		__cpufreq_remove_dev_prepare(dev);
+		cpufreq_offline_prepare(cpu);
-		__cpufreq_remove_dev_finish(dev);
+		cpufreq_offline_finish(cpu);
 	}
 
 	cpumask_clear_cpu(cpu, policy->real_cpus);
@@ -2247,7 +2171,11 @@ static int cpufreq_set_policy(struct cpufreq_policy *policy,
 
 	memcpy(&new_policy->cpuinfo, &policy->cpuinfo, sizeof(policy->cpuinfo));
 
-	if (new_policy->min > policy->max || new_policy->max < policy->min)
+	/*
+	* This check works well when we store new min/max freq attributes,
+	* because new_policy is a copy of policy with one field updated.
+	*/
+	if (new_policy->min > new_policy->max)
 		return -EINVAL;
 
 	/* verify the cpu speed can be set within this limit */
@@ -2259,10 +2187,6 @@ static int cpufreq_set_policy(struct cpufreq_policy *policy,
 	blocking_notifier_call_chain(&cpufreq_policy_notifier_list,
 			CPUFREQ_ADJUST, new_policy);
 
-	/* adjust if necessary - hardware incompatibility*/
-	blocking_notifier_call_chain(&cpufreq_policy_notifier_list,
-			CPUFREQ_INCOMPATIBLE, new_policy);
-
 	/*
 	 * verify the cpu speed can be set within this limit, which might be
 	 * different to the first one
@@ -2296,16 +2220,31 @@ static int cpufreq_set_policy(struct cpufreq_policy *policy,
 	old_gov = policy->governor;
 	/* end old governor */
 	if (old_gov) {
-		__cpufreq_governor(policy, CPUFREQ_GOV_STOP);
+		ret = __cpufreq_governor(policy, CPUFREQ_GOV_STOP);
+		if (ret) {
+			/* This can happen due to race with other operations */
+			pr_debug("%s: Failed to Stop Governor: %s (%d)\n",
+				 __func__, old_gov->name, ret);
+			return ret;
+		}
+
 		up_write(&policy->rwsem);
-		__cpufreq_governor(policy, CPUFREQ_GOV_POLICY_EXIT);
+		ret = __cpufreq_governor(policy, CPUFREQ_GOV_POLICY_EXIT);
 		down_write(&policy->rwsem);
+
+		if (ret) {
+			pr_err("%s: Failed to Exit Governor: %s (%d)\n",
+			       __func__, old_gov->name, ret);
+			return ret;
+		}
 	}
 
 	/* start new governor */
 	policy->governor = new_policy->governor;
-	if (!__cpufreq_governor(policy, CPUFREQ_GOV_POLICY_INIT)) {
+	ret = __cpufreq_governor(policy, CPUFREQ_GOV_POLICY_INIT);
-		if (!__cpufreq_governor(policy, CPUFREQ_GOV_START))
+	if (!ret) {
+		ret = __cpufreq_governor(policy, CPUFREQ_GOV_START);
+		if (!ret)
 			goto out;
 
 		up_write(&policy->rwsem);
@@ -2317,11 +2256,13 @@ static int cpufreq_set_policy(struct cpufreq_policy *policy,
 	pr_debug("starting governor %s failed\n", policy->governor->name);
 	if (old_gov) {
 		policy->governor = old_gov;
-		__cpufreq_governor(policy, CPUFREQ_GOV_POLICY_INIT);
+		if (__cpufreq_governor(policy, CPUFREQ_GOV_POLICY_INIT))
-		__cpufreq_governor(policy, CPUFREQ_GOV_START);
+			policy->governor = NULL;
+		else
+			__cpufreq_governor(policy, CPUFREQ_GOV_START);
 	}
 
-	return -EINVAL;
+	return ret;
 
  out:
 	pr_debug("governor: change or update limits\n");
@@ -2350,8 +2291,6 @@ int cpufreq_update_policy(unsigned int cpu)
 	memcpy(&new_policy, policy, sizeof(*policy));
 	new_policy.min = policy->user_policy.min;
 	new_policy.max = policy->user_policy.max;
-	new_policy.policy = policy->user_policy.policy;
-	new_policy.governor = policy->user_policy.governor;
 
 	/*
 	 * BIOS might change freq behind our back
@@ -2387,27 +2326,23 @@ static int cpufreq_cpu_callback(struct notifier_block *nfb,
 					unsigned long action, void *hcpu)
 {
 	unsigned int cpu = (unsigned long)hcpu;
-	struct device *dev;
 
-	dev = get_cpu_device(cpu);
+	switch (action & ~CPU_TASKS_FROZEN) {
-	if (dev) {
+	case CPU_ONLINE:
-		switch (action & ~CPU_TASKS_FROZEN) {
+		cpufreq_online(cpu);
-		case CPU_ONLINE:
+		break;
-			cpufreq_add_dev(dev, NULL);
-			break;
 
-		case CPU_DOWN_PREPARE:
+	case CPU_DOWN_PREPARE:
-			__cpufreq_remove_dev_prepare(dev);
+		cpufreq_offline_prepare(cpu);
-			break;
+		break;
 
-		case CPU_POST_DEAD:
+	case CPU_POST_DEAD:
-			__cpufreq_remove_dev_finish(dev);
+		cpufreq_offline_finish(cpu);
-			break;
+		break;
 
-		case CPU_DOWN_FAILED:
+	case CPU_DOWN_FAILED:
-			cpufreq_add_dev(dev, NULL);
+		cpufreq_online(cpu);
-			break;
+		break;
-		}
 	}
 	return NOTIFY_OK;
 }
@@ -2515,10 +2450,14 @@ int cpufreq_register_driver(struct cpufreq_driver *driver_data)
 
 	pr_debug("trying to register driver %s\n", driver_data->name);
 
+	/* Protect against concurrent CPU online/offline. */
+	get_online_cpus();
+
 	write_lock_irqsave(&cpufreq_driver_lock, flags);
 	if (cpufreq_driver) {
 		write_unlock_irqrestore(&cpufreq_driver_lock, flags);
-		return -EEXIST;
+		ret = -EEXIST;
+		goto out;
 	}
 	cpufreq_driver = driver_data;
 	write_unlock_irqrestore(&cpufreq_driver_lock, flags);
@@ -2557,7 +2496,10 @@ int cpufreq_register_driver(struct cpufreq_driver *driver_data)
 	register_hotcpu_notifier(&cpufreq_cpu_notifier);
 	pr_debug("driver %s up and running\n", driver_data->name);
 
-	return 0;
+out:
+	put_online_cpus();
+	return ret;
+
 err_if_unreg:
 	subsys_interface_unregister(&cpufreq_interface);
 err_boost_unreg:
@@ -2567,7 +2509,7 @@ err_null_driver:
 	write_lock_irqsave(&cpufreq_driver_lock, flags);
 	cpufreq_driver = NULL;
 	write_unlock_irqrestore(&cpufreq_driver_lock, flags);
-	return ret;
+	goto out;
 }
 EXPORT_SYMBOL_GPL(cpufreq_register_driver);
 
@@ -2588,19 +2530,20 @@ int cpufreq_unregister_driver(struct cpufreq_driver *driver)
 
 	pr_debug("unregistering driver %s\n", driver->name);
 
+	/* Protect against concurrent cpu hotplug */
+	get_online_cpus();
 	subsys_interface_unregister(&cpufreq_interface);
 	if (cpufreq_boost_supported())
 		cpufreq_sysfs_remove_file(&boost.attr);
 
 	unregister_hotcpu_notifier(&cpufreq_cpu_notifier);
 
-	down_write(&cpufreq_rwsem);
 	write_lock_irqsave(&cpufreq_driver_lock, flags);
 
 	cpufreq_driver = NULL;
 
 	write_unlock_irqrestore(&cpufreq_driver_lock, flags);
-	up_write(&cpufreq_rwsem);
+	put_online_cpus();
 
 	return 0;
 }

drivers/cpufreq/cpufreq_conservative.c

@@ -47,7 +47,7 @@ static inline unsigned int get_freq_target(struct cs_dbs_tuners *cs_tuners,
 static void cs_check_cpu(int cpu, unsigned int load)
 {
 	struct cs_cpu_dbs_info_s *dbs_info = &per_cpu(cs_cpu_dbs_info, cpu);
-	struct cpufreq_policy *policy = dbs_info->cdbs.cur_policy;
+	struct cpufreq_policy *policy = dbs_info->cdbs.shared->policy;
 	struct dbs_data *dbs_data = policy->governor_data;
 	struct cs_dbs_tuners *cs_tuners = dbs_data->tuners;
 
@@ -102,26 +102,15 @@ static void cs_check_cpu(int cpu, unsigned int load)
 	}
 }
 
-static void cs_dbs_timer(struct work_struct *work)
+static unsigned int cs_dbs_timer(struct cpu_dbs_info *cdbs,
+				 struct dbs_data *dbs_data, bool modify_all)
 {
-	struct cs_cpu_dbs_info_s *dbs_info = container_of(work,
-			struct cs_cpu_dbs_info_s, cdbs.work.work);
-	unsigned int cpu = dbs_info->cdbs.cur_policy->cpu;
-	struct cs_cpu_dbs_info_s *core_dbs_info = &per_cpu(cs_cpu_dbs_info,
-			cpu);
-	struct dbs_data *dbs_data = dbs_info->cdbs.cur_policy->governor_data;
 	struct cs_dbs_tuners *cs_tuners = dbs_data->tuners;
-	int delay = delay_for_sampling_rate(cs_tuners->sampling_rate);
-	bool modify_all = true;
 
-	mutex_lock(&core_dbs_info->cdbs.timer_mutex);
+	if (modify_all)
-	if (!need_load_eval(&core_dbs_info->cdbs, cs_tuners->sampling_rate))
+		dbs_check_cpu(dbs_data, cdbs->shared->policy->cpu);
-		modify_all = false;
-	else
-		dbs_check_cpu(dbs_data, cpu);
 
-	gov_queue_work(dbs_data, dbs_info->cdbs.cur_policy, delay, modify_all);
+	return delay_for_sampling_rate(cs_tuners->sampling_rate);
-	mutex_unlock(&core_dbs_info->cdbs.timer_mutex);
 }
 
 static int dbs_cpufreq_notifier(struct notifier_block *nb, unsigned long val,
@@ -135,7 +124,7 @@ static int dbs_cpufreq_notifier(struct notifier_block *nb, unsigned long val,
 	if (!dbs_info->enable)
 		return 0;
 
-	policy = dbs_info->cdbs.cur_policy;
+	policy = dbs_info->cdbs.shared->policy;
 
 	/*
 	 * we only care if our internally tracked freq moves outside the 'valid'

drivers/cpufreq/cpufreq_governor.c

@@ -32,10 +32,10 @@ static struct attribute_group *get_sysfs_attr(struct dbs_data *dbs_data)
 
 void dbs_check_cpu(struct dbs_data *dbs_data, int cpu)
 {
-	struct cpu_dbs_common_info *cdbs = dbs_data->cdata->get_cpu_cdbs(cpu);
+	struct cpu_dbs_info *cdbs = dbs_data->cdata->get_cpu_cdbs(cpu);
 	struct od_dbs_tuners *od_tuners = dbs_data->tuners;
 	struct cs_dbs_tuners *cs_tuners = dbs_data->tuners;
-	struct cpufreq_policy *policy;
+	struct cpufreq_policy *policy = cdbs->shared->policy;
 	unsigned int sampling_rate;
 	unsigned int max_load = 0;
 	unsigned int ignore_nice;
@@ -60,11 +60,9 @@ void dbs_check_cpu(struct dbs_data *dbs_data, int cpu)
 		ignore_nice = cs_tuners->ignore_nice_load;
 	}
 
-	policy = cdbs->cur_policy;
-
 	/* Get Absolute Load */
 	for_each_cpu(j, policy->cpus) {
-		struct cpu_dbs_common_info *j_cdbs;
+		struct cpu_dbs_info *j_cdbs;
 		u64 cur_wall_time, cur_idle_time;
 		unsigned int idle_time, wall_time;
 		unsigned int load;
@@ -163,9 +161,9 @@ EXPORT_SYMBOL_GPL(dbs_check_cpu);
 static inline void __gov_queue_work(int cpu, struct dbs_data *dbs_data,
 		unsigned int delay)
 {
-	struct cpu_dbs_common_info *cdbs = dbs_data->cdata->get_cpu_cdbs(cpu);
+	struct cpu_dbs_info *cdbs = dbs_data->cdata->get_cpu_cdbs(cpu);
 
-	mod_delayed_work_on(cpu, system_wq, &cdbs->work, delay);
+	mod_delayed_work_on(cpu, system_wq, &cdbs->dwork, delay);
 }
 
 void gov_queue_work(struct dbs_data *dbs_data, struct cpufreq_policy *policy,
@@ -199,33 +197,63 @@ EXPORT_SYMBOL_GPL(gov_queue_work);
 static inline void gov_cancel_work(struct dbs_data *dbs_data,
 		struct cpufreq_policy *policy)
 {
-	struct cpu_dbs_common_info *cdbs;
+	struct cpu_dbs_info *cdbs;
 	int i;
 
 	for_each_cpu(i, policy->cpus) {
 		cdbs = dbs_data->cdata->get_cpu_cdbs(i);
-		cancel_delayed_work_sync(&cdbs->work);
+		cancel_delayed_work_sync(&cdbs->dwork);
 	}
 }
 
 /* Will return if we need to evaluate cpu load again or not */
-bool need_load_eval(struct cpu_dbs_common_info *cdbs,
+static bool need_load_eval(struct cpu_common_dbs_info *shared,
-		unsigned int sampling_rate)
+			   unsigned int sampling_rate)
 {
-	if (policy_is_shared(cdbs->cur_policy)) {
+	if (policy_is_shared(shared->policy)) {
 		ktime_t time_now = ktime_get();
-		s64 delta_us = ktime_us_delta(time_now, cdbs->time_stamp);
+		s64 delta_us = ktime_us_delta(time_now, shared->time_stamp);
 
 		/* Do nothing if we recently have sampled */
 		if (delta_us < (s64)(sampling_rate / 2))
 			return false;
 		else
-			cdbs->time_stamp = time_now;
+			shared->time_stamp = time_now;
 	}
 
 	return true;
 }
-EXPORT_SYMBOL_GPL(need_load_eval);
+
+static void dbs_timer(struct work_struct *work)
+{
+	struct cpu_dbs_info *cdbs = container_of(work, struct cpu_dbs_info,
+						 dwork.work);
+	struct cpu_common_dbs_info *shared = cdbs->shared;
+	struct cpufreq_policy *policy = shared->policy;
+	struct dbs_data *dbs_data = policy->governor_data;
+	unsigned int sampling_rate, delay;
+	bool modify_all = true;
+
+	mutex_lock(&shared->timer_mutex);
+
+	if (dbs_data->cdata->governor == GOV_CONSERVATIVE) {
+		struct cs_dbs_tuners *cs_tuners = dbs_data->tuners;
+
+		sampling_rate = cs_tuners->sampling_rate;
+	} else {
+		struct od_dbs_tuners *od_tuners = dbs_data->tuners;
+
+		sampling_rate = od_tuners->sampling_rate;
+	}
+
+	if (!need_load_eval(cdbs->shared, sampling_rate))
+		modify_all = false;
+
+	delay = dbs_data->cdata->gov_dbs_timer(cdbs, dbs_data, modify_all);
+	gov_queue_work(dbs_data, policy, delay, modify_all);
+
+	mutex_unlock(&shared->timer_mutex);
+}
 
 static void set_sampling_rate(struct dbs_data *dbs_data,
 		unsigned int sampling_rate)
@@ -239,6 +267,37 @@ static void set_sampling_rate(struct dbs_data *dbs_data,
 	}
 }
 
+static int alloc_common_dbs_info(struct cpufreq_policy *policy,
+				 struct common_dbs_data *cdata)
+{
+	struct cpu_common_dbs_info *shared;
+	int j;
+
+	/* Allocate memory for the common information for policy->cpus */
+	shared = kzalloc(sizeof(*shared), GFP_KERNEL);
+	if (!shared)
+		return -ENOMEM;
+
+	/* Set shared for all CPUs, online+offline */
+	for_each_cpu(j, policy->related_cpus)
+		cdata->get_cpu_cdbs(j)->shared = shared;
+
+	return 0;
+}
+
+static void free_common_dbs_info(struct cpufreq_policy *policy,
+				 struct common_dbs_data *cdata)
+{
+	struct cpu_dbs_info *cdbs = cdata->get_cpu_cdbs(policy->cpu);
+	struct cpu_common_dbs_info *shared = cdbs->shared;
+	int j;
+
+	for_each_cpu(j, policy->cpus)
+		cdata->get_cpu_cdbs(j)->shared = NULL;
+
+	kfree(shared);
+}
+
 static int cpufreq_governor_init(struct cpufreq_policy *policy,
 				 struct dbs_data *dbs_data,
 				 struct common_dbs_data *cdata)
@@ -246,9 +305,18 @@ static int cpufreq_governor_init(struct cpufreq_policy *policy,
 	unsigned int latency;
 	int ret;
 
+	/* State should be equivalent to EXIT */
+	if (policy->governor_data)
+		return -EBUSY;
+
 	if (dbs_data) {
 		if (WARN_ON(have_governor_per_policy()))
 			return -EINVAL;
+
+		ret = alloc_common_dbs_info(policy, cdata);
+		if (ret)
+			return ret;
+
 		dbs_data->usage_count++;
 		policy->governor_data = dbs_data;
 		return 0;
@@ -258,12 +326,16 @@ static int cpufreq_governor_init(struct cpufreq_policy *policy,
 	if (!dbs_data)
 		return -ENOMEM;
 
+	ret = alloc_common_dbs_info(policy, cdata);
+	if (ret)
+		goto free_dbs_data;
+
 	dbs_data->cdata = cdata;
 	dbs_data->usage_count = 1;
 
 	ret = cdata->init(dbs_data, !policy->governor->initialized);
 	if (ret)
-		goto free_dbs_data;
+		goto free_common_dbs_info;
 
 	/* policy latency is in ns. Convert it to us first */
 	latency = policy->cpuinfo.transition_latency / 1000;
@@ -300,15 +372,22 @@ put_kobj:
 	}
 cdata_exit:
 	cdata->exit(dbs_data, !policy->governor->initialized);
+free_common_dbs_info:
+	free_common_dbs_info(policy, cdata);
 free_dbs_data:
 	kfree(dbs_data);
 	return ret;
 }
 
-static void cpufreq_governor_exit(struct cpufreq_policy *policy,
+static int cpufreq_governor_exit(struct cpufreq_policy *policy,
-				  struct dbs_data *dbs_data)
+				 struct dbs_data *dbs_data)
 {
 	struct common_dbs_data *cdata = dbs_data->cdata;
+	struct cpu_dbs_info *cdbs = cdata->get_cpu_cdbs(policy->cpu);
+
+	/* State should be equivalent to INIT */
+	if (!cdbs->shared || cdbs->shared->policy)
+		return -EBUSY;
 
 	policy->governor_data = NULL;
 	if (!--dbs_data->usage_count) {
@@ -323,6 +402,9 @@ static void cpufreq_governor_exit(struct cpufreq_policy *policy,
 		cdata->exit(dbs_data, policy->governor->initialized == 1);
 		kfree(dbs_data);
 	}
+
+	free_common_dbs_info(policy, cdata);
+	return 0;
 }
 
 static int cpufreq_governor_start(struct cpufreq_policy *policy,
@@ -330,12 +412,17 @@ static int cpufreq_governor_start(struct cpufreq_policy *policy,
 {
 	struct common_dbs_data *cdata = dbs_data->cdata;
 	unsigned int sampling_rate, ignore_nice, j, cpu = policy->cpu;
-	struct cpu_dbs_common_info *cpu_cdbs = cdata->get_cpu_cdbs(cpu);
+	struct cpu_dbs_info *cdbs = cdata->get_cpu_cdbs(cpu);
+	struct cpu_common_dbs_info *shared = cdbs->shared;
 	int io_busy = 0;
 
 	if (!policy->cur)
 		return -EINVAL;
 
+	/* State should be equivalent to INIT */
+	if (!shared || shared->policy)
+		return -EBUSY;
+
 	if (cdata->governor == GOV_CONSERVATIVE) {
 		struct cs_dbs_tuners *cs_tuners = dbs_data->tuners;
 
@@ -349,12 +436,14 @@ static int cpufreq_governor_start(struct cpufreq_policy *policy,
 		io_busy = od_tuners->io_is_busy;
 	}
 
+	shared->policy = policy;
+	shared->time_stamp = ktime_get();
+	mutex_init(&shared->timer_mutex);
+
 	for_each_cpu(j, policy->cpus) {
-		struct cpu_dbs_common_info *j_cdbs = cdata->get_cpu_cdbs(j);
+		struct cpu_dbs_info *j_cdbs = cdata->get_cpu_cdbs(j);
 		unsigned int prev_load;
 
-		j_cdbs->cpu = j;
-		j_cdbs->cur_policy = policy;
 		j_cdbs->prev_cpu_idle =
 			get_cpu_idle_time(j, &j_cdbs->prev_cpu_wall, io_busy);
 
@@ -366,8 +455,7 @@ static int cpufreq_governor_start(struct cpufreq_policy *policy,
 		if (ignore_nice)
 			j_cdbs->prev_cpu_nice = kcpustat_cpu(j).cpustat[CPUTIME_NICE];
 
-		mutex_init(&j_cdbs->timer_mutex);
+		INIT_DEFERRABLE_WORK(&j_cdbs->dwork, dbs_timer);
-		INIT_DEFERRABLE_WORK(&j_cdbs->work, cdata->gov_dbs_timer);
 	}
 
 	if (cdata->governor == GOV_CONSERVATIVE) {
@@ -386,20 +474,24 @@ static int cpufreq_governor_start(struct cpufreq_policy *policy,
 		od_ops->powersave_bias_init_cpu(cpu);
 	}
 
-	/* Initiate timer time stamp */
-	cpu_cdbs->time_stamp = ktime_get();
-
 	gov_queue_work(dbs_data, policy, delay_for_sampling_rate(sampling_rate),
 		       true);
 	return 0;
 }
 
-static void cpufreq_governor_stop(struct cpufreq_policy *policy,
+static int cpufreq_governor_stop(struct cpufreq_policy *policy,
-				  struct dbs_data *dbs_data)
+				 struct dbs_data *dbs_data)
 {
 	struct common_dbs_data *cdata = dbs_data->cdata;
 	unsigned int cpu = policy->cpu;
-	struct cpu_dbs_common_info *cpu_cdbs = cdata->get_cpu_cdbs(cpu);
+	struct cpu_dbs_info *cdbs = cdata->get_cpu_cdbs(cpu);
+	struct cpu_common_dbs_info *shared = cdbs->shared;
+
+	/* State should be equivalent to START */
+	if (!shared || !shared->policy)
+		return -EBUSY;
+
+	gov_cancel_work(dbs_data, policy);
 
 	if (cdata->governor == GOV_CONSERVATIVE) {
 		struct cs_cpu_dbs_info_s *cs_dbs_info =
@@ -408,38 +500,40 @@ static void cpufreq_governor_stop(struct cpufreq_policy *policy,
 		cs_dbs_info->enable = 0;
 	}
 
-	gov_cancel_work(dbs_data, policy);
+	shared->policy = NULL;
-
+	mutex_destroy(&shared->timer_mutex);
-	mutex_destroy(&cpu_cdbs->timer_mutex);
+	return 0;
-	cpu_cdbs->cur_policy = NULL;
 }
 
-static void cpufreq_governor_limits(struct cpufreq_policy *policy,
+static int cpufreq_governor_limits(struct cpufreq_policy *policy,
-				    struct dbs_data *dbs_data)
+				   struct dbs_data *dbs_data)
 {
 	struct common_dbs_data *cdata = dbs_data->cdata;
 	unsigned int cpu = policy->cpu;
-	struct cpu_dbs_common_info *cpu_cdbs = cdata->get_cpu_cdbs(cpu);
+	struct cpu_dbs_info *cdbs = cdata->get_cpu_cdbs(cpu);
 
-	if (!cpu_cdbs->cur_policy)
+	/* State should be equivalent to START */
-		return;
+	if (!cdbs->shared || !cdbs->shared->policy)
+		return -EBUSY;
 
-	mutex_lock(&cpu_cdbs->timer_mutex);
+	mutex_lock(&cdbs->shared->timer_mutex);
-	if (policy->max < cpu_cdbs->cur_policy->cur)
+	if (policy->max < cdbs->shared->policy->cur)
-		__cpufreq_driver_target(cpu_cdbs->cur_policy, policy->max,
+		__cpufreq_driver_target(cdbs->shared->policy, policy->max,
 					CPUFREQ_RELATION_H);
-	else if (policy->min > cpu_cdbs->cur_policy->cur)
+	else if (policy->min > cdbs->shared->policy->cur)
-		__cpufreq_driver_target(cpu_cdbs->cur_policy, policy->min,
+		__cpufreq_driver_target(cdbs->shared->policy, policy->min,
 					CPUFREQ_RELATION_L);
 	dbs_check_cpu(dbs_data, cpu);
-	mutex_unlock(&cpu_cdbs->timer_mutex);
+	mutex_unlock(&cdbs->shared->timer_mutex);
+
+	return 0;
 }
 
 int cpufreq_governor_dbs(struct cpufreq_policy *policy,
 			 struct common_dbs_data *cdata, unsigned int event)
 {
 	struct dbs_data *dbs_data;
-	int ret = 0;
+	int ret;
 
 	/* Lock governor to block concurrent initialization of governor */
 	mutex_lock(&cdata->mutex);
@@ -449,7 +543,7 @@ int cpufreq_governor_dbs(struct cpufreq_policy *policy,
 	else
 		dbs_data = cdata->gdbs_data;
 
-	if (WARN_ON(!dbs_data && (event != CPUFREQ_GOV_POLICY_INIT))) {
+	if (!dbs_data && (event != CPUFREQ_GOV_POLICY_INIT)) {
 		ret = -EINVAL;
 		goto unlock;
 	}
@@ -459,17 +553,19 @@ int cpufreq_governor_dbs(struct cpufreq_policy *policy,
 		ret = cpufreq_governor_init(policy, dbs_data, cdata);
 		break;
 	case CPUFREQ_GOV_POLICY_EXIT:
-		cpufreq_governor_exit(policy, dbs_data);
+		ret = cpufreq_governor_exit(policy, dbs_data);
 		break;
 	case CPUFREQ_GOV_START:
 		ret = cpufreq_governor_start(policy, dbs_data);
 		break;
 	case CPUFREQ_GOV_STOP:
-		cpufreq_governor_stop(policy, dbs_data);
+		ret = cpufreq_governor_stop(policy, dbs_data);
 		break;
 	case CPUFREQ_GOV_LIMITS:
-		cpufreq_governor_limits(policy, dbs_data);
+		ret = cpufreq_governor_limits(policy, dbs_data);
 		break;
+	default:
+		ret = -EINVAL;
 	}
 
 unlock:

drivers/cpufreq/cpufreq_governor.h

@@ -109,7 +109,7 @@ store_one(_gov, file_name)
 
 /* create helper routines */
 #define define_get_cpu_dbs_routines(_dbs_info)				\
-static struct cpu_dbs_common_info *get_cpu_cdbs(int cpu)		\
+static struct cpu_dbs_info *get_cpu_cdbs(int cpu)			\
 {									\
 	return &per_cpu(_dbs_info, cpu).cdbs;				\
 }									\
@@ -128,9 +128,20 @@ static void *get_cpu_dbs_info_s(int cpu)				\
  * cs_*: Conservative governor
  */
 
+/* Common to all CPUs of a policy */
+struct cpu_common_dbs_info {
+	struct cpufreq_policy *policy;
+	/*
+	 * percpu mutex that serializes governor limit change with dbs_timer
+	 * invocation. We do not want dbs_timer to run when user is changing
+	 * the governor or limits.
+	 */
+	struct mutex timer_mutex;
+	ktime_t time_stamp;
+};
+
 /* Per cpu structures */
-struct cpu_dbs_common_info {
+struct cpu_dbs_info {
-	int cpu;
 	u64 prev_cpu_idle;
 	u64 prev_cpu_wall;
 	u64 prev_cpu_nice;
@@ -141,19 +152,12 @@ struct cpu_dbs_common_info {
 	 * wake-up from idle.
 	 */
 	unsigned int prev_load;
-	struct cpufreq_policy *cur_policy;
+	struct delayed_work dwork;
-	struct delayed_work work;
+	struct cpu_common_dbs_info *shared;
-	/*
-	 * percpu mutex that serializes governor limit change with gov_dbs_timer
-	 * invocation. We do not want gov_dbs_timer to run when user is changing
-	 * the governor or limits.
-	 */
-	struct mutex timer_mutex;
-	ktime_t time_stamp;
 };
 
 struct od_cpu_dbs_info_s {
-	struct cpu_dbs_common_info cdbs;
+	struct cpu_dbs_info cdbs;
 	struct cpufreq_frequency_table *freq_table;
 	unsigned int freq_lo;
 	unsigned int freq_lo_jiffies;
@@ -163,7 +167,7 @@ struct od_cpu_dbs_info_s {
 };
 
 struct cs_cpu_dbs_info_s {
-	struct cpu_dbs_common_info cdbs;
+	struct cpu_dbs_info cdbs;
 	unsigned int down_skip;
 	unsigned int requested_freq;
 	unsigned int enable:1;
@@ -204,9 +208,11 @@ struct common_dbs_data {
 	 */
 	struct dbs_data *gdbs_data;
 
-	struct cpu_dbs_common_info *(*get_cpu_cdbs)(int cpu);
+	struct cpu_dbs_info *(*get_cpu_cdbs)(int cpu);
 	void *(*get_cpu_dbs_info_s)(int cpu);
-	void (*gov_dbs_timer)(struct work_struct *work);
+	unsigned int (*gov_dbs_timer)(struct cpu_dbs_info *cdbs,
+				      struct dbs_data *dbs_data,
+				      bool modify_all);
 	void (*gov_check_cpu)(int cpu, unsigned int load);
 	int (*init)(struct dbs_data *dbs_data, bool notify);
 	void (*exit)(struct dbs_data *dbs_data, bool notify);
@@ -265,8 +271,6 @@ static ssize_t show_sampling_rate_min_gov_pol				\
 extern struct mutex cpufreq_governor_lock;
 
 void dbs_check_cpu(struct dbs_data *dbs_data, int cpu);
-bool need_load_eval(struct cpu_dbs_common_info *cdbs,
-		unsigned int sampling_rate);
 int cpufreq_governor_dbs(struct cpufreq_policy *policy,
 		struct common_dbs_data *cdata, unsigned int event);
 void gov_queue_work(struct dbs_data *dbs_data, struct cpufreq_policy *policy,

drivers/cpufreq/cpufreq_ondemand.c

@@ -155,7 +155,7 @@ static void dbs_freq_increase(struct cpufreq_policy *policy, unsigned int freq)
 static void od_check_cpu(int cpu, unsigned int load)
 {
 	struct od_cpu_dbs_info_s *dbs_info = &per_cpu(od_cpu_dbs_info, cpu);
-	struct cpufreq_policy *policy = dbs_info->cdbs.cur_policy;
+	struct cpufreq_policy *policy = dbs_info->cdbs.shared->policy;
 	struct dbs_data *dbs_data = policy->governor_data;
 	struct od_dbs_tuners *od_tuners = dbs_data->tuners;
 
@@ -191,46 +191,40 @@ static void od_check_cpu(int cpu, unsigned int load)
 	}
 }
 
-static void od_dbs_timer(struct work_struct *work)
+static unsigned int od_dbs_timer(struct cpu_dbs_info *cdbs,
+				 struct dbs_data *dbs_data, bool modify_all)
 {
-	struct od_cpu_dbs_info_s *dbs_info =
+	struct cpufreq_policy *policy = cdbs->shared->policy;
-		container_of(work, struct od_cpu_dbs_info_s, cdbs.work.work);
+	unsigned int cpu = policy->cpu;
-	unsigned int cpu = dbs_info->cdbs.cur_policy->cpu;
+	struct od_cpu_dbs_info_s *dbs_info = &per_cpu(od_cpu_dbs_info,
-	struct od_cpu_dbs_info_s *core_dbs_info = &per_cpu(od_cpu_dbs_info,
 			cpu);
-	struct dbs_data *dbs_data = dbs_info->cdbs.cur_policy->governor_data;
 	struct od_dbs_tuners *od_tuners = dbs_data->tuners;
-	int delay = 0, sample_type = core_dbs_info->sample_type;
+	int delay = 0, sample_type = dbs_info->sample_type;
-	bool modify_all = true;
 
-	mutex_lock(&core_dbs_info->cdbs.timer_mutex);
+	if (!modify_all)
-	if (!need_load_eval(&core_dbs_info->cdbs, od_tuners->sampling_rate)) {
-		modify_all = false;
 		goto max_delay;
-	}
 
 	/* Common NORMAL_SAMPLE setup */
-	core_dbs_info->sample_type = OD_NORMAL_SAMPLE;
+	dbs_info->sample_type = OD_NORMAL_SAMPLE;
 	if (sample_type == OD_SUB_SAMPLE) {
-		delay = core_dbs_info->freq_lo_jiffies;
+		delay = dbs_info->freq_lo_jiffies;
-		__cpufreq_driver_target(core_dbs_info->cdbs.cur_policy,
+		__cpufreq_driver_target(policy, dbs_info->freq_lo,
-				core_dbs_info->freq_lo, CPUFREQ_RELATION_H);
+					CPUFREQ_RELATION_H);
 	} else {
 		dbs_check_cpu(dbs_data, cpu);
-		if (core_dbs_info->freq_lo) {
+		if (dbs_info->freq_lo) {
 			/* Setup timer for SUB_SAMPLE */
-			core_dbs_info->sample_type = OD_SUB_SAMPLE;
+			dbs_info->sample_type = OD_SUB_SAMPLE;
-			delay = core_dbs_info->freq_hi_jiffies;
+			delay = dbs_info->freq_hi_jiffies;
 		}
 	}
 
 max_delay:
 	if (!delay)
 		delay = delay_for_sampling_rate(od_tuners->sampling_rate
-				* core_dbs_info->rate_mult);
+				* dbs_info->rate_mult);
 
-	gov_queue_work(dbs_data, dbs_info->cdbs.cur_policy, delay, modify_all);
+	return delay;
-	mutex_unlock(&core_dbs_info->cdbs.timer_mutex);
 }
 
 /************************** sysfs interface ************************/
@@ -273,27 +267,27 @@ static void update_sampling_rate(struct dbs_data *dbs_data,
 		dbs_info = &per_cpu(od_cpu_dbs_info, cpu);
 		cpufreq_cpu_put(policy);
 
-		mutex_lock(&dbs_info->cdbs.timer_mutex);
+		mutex_lock(&dbs_info->cdbs.shared->timer_mutex);
 
-		if (!delayed_work_pending(&dbs_info->cdbs.work)) {
+		if (!delayed_work_pending(&dbs_info->cdbs.dwork)) {
-			mutex_unlock(&dbs_info->cdbs.timer_mutex);
+			mutex_unlock(&dbs_info->cdbs.shared->timer_mutex);
 			continue;
 		}
 
 		next_sampling = jiffies + usecs_to_jiffies(new_rate);
-		appointed_at = dbs_info->cdbs.work.timer.expires;
+		appointed_at = dbs_info->cdbs.dwork.timer.expires;
 
 		if (time_before(next_sampling, appointed_at)) {
 
-			mutex_unlock(&dbs_info->cdbs.timer_mutex);
+			mutex_unlock(&dbs_info->cdbs.shared->timer_mutex);
-			cancel_delayed_work_sync(&dbs_info->cdbs.work);
+			cancel_delayed_work_sync(&dbs_info->cdbs.dwork);
-			mutex_lock(&dbs_info->cdbs.timer_mutex);
+			mutex_lock(&dbs_info->cdbs.shared->timer_mutex);
 
-			gov_queue_work(dbs_data, dbs_info->cdbs.cur_policy,
+			gov_queue_work(dbs_data, policy,
-					usecs_to_jiffies(new_rate), true);
+				       usecs_to_jiffies(new_rate), true);
 
 		}
-		mutex_unlock(&dbs_info->cdbs.timer_mutex);
+		mutex_unlock(&dbs_info->cdbs.shared->timer_mutex);
 	}
 }
 
@@ -556,13 +550,16 @@ static void od_set_powersave_bias(unsigned int powersave_bias)
 
 	get_online_cpus();
 	for_each_online_cpu(cpu) {
+		struct cpu_common_dbs_info *shared;
+
 		if (cpumask_test_cpu(cpu, &done))
 			continue;
 
-		policy = per_cpu(od_cpu_dbs_info, cpu).cdbs.cur_policy;
+		shared = per_cpu(od_cpu_dbs_info, cpu).cdbs.shared;
-		if (!policy)
+		if (!shared)
 			continue;
 
+		policy = shared->policy;
 		cpumask_or(&done, &done, policy->cpus);
 
 		if (policy->governor != &cpufreq_gov_ondemand)

drivers/cpufreq/e_powersaver.c

@@ -78,7 +78,7 @@ static int eps_acpi_init(void)
 static int eps_acpi_exit(struct cpufreq_policy *policy)
 {
 	if (eps_acpi_cpu_perf) {
-		acpi_processor_unregister_performance(eps_acpi_cpu_perf, 0);
+		acpi_processor_unregister_performance(0);
 		free_cpumask_var(eps_acpi_cpu_perf->shared_cpu_map);
 		kfree(eps_acpi_cpu_perf);
 		eps_acpi_cpu_perf = NULL;

drivers/cpufreq/ia64-acpi-cpufreq.c

@@ -29,7 +29,6 @@ MODULE_LICENSE("GPL");
 
 struct cpufreq_acpi_io {
 	struct acpi_processor_performance	acpi_data;
-	struct cpufreq_frequency_table		*freq_table;
 	unsigned int				resume;
 };
 
@@ -221,6 +220,7 @@ acpi_cpufreq_cpu_init (
 	unsigned int		cpu = policy->cpu;
 	struct cpufreq_acpi_io	*data;
 	unsigned int		result = 0;
+	struct cpufreq_frequency_table *freq_table;
 
 	pr_debug("acpi_cpufreq_cpu_init\n");
 
@@ -254,10 +254,10 @@ acpi_cpufreq_cpu_init (
 	}
 
 	/* alloc freq_table */
-	data->freq_table = kzalloc(sizeof(*data->freq_table) *
+	freq_table = kzalloc(sizeof(*freq_table) *
 	                           (data->acpi_data.state_count + 1),
 	                           GFP_KERNEL);
-	if (!data->freq_table) {
+	if (!freq_table) {
 		result = -ENOMEM;
 		goto err_unreg;
 	}
@@ -276,14 +276,14 @@ acpi_cpufreq_cpu_init (
 	for (i = 0; i <= data->acpi_data.state_count; i++)
 	{
 		if (i < data->acpi_data.state_count) {
-			data->freq_table[i].frequency =
+			freq_table[i].frequency =
 			      data->acpi_data.states[i].core_frequency * 1000;
 		} else {
-			data->freq_table[i].frequency = CPUFREQ_TABLE_END;
+			freq_table[i].frequency = CPUFREQ_TABLE_END;
 		}
 	}
 
-	result = cpufreq_table_validate_and_show(policy, data->freq_table);
+	result = cpufreq_table_validate_and_show(policy, freq_table);
 	if (result) {
 		goto err_freqfree;
 	}
@@ -311,9 +311,9 @@ acpi_cpufreq_cpu_init (
 	return (result);
 
  err_freqfree:
-	kfree(data->freq_table);
+	kfree(freq_table);
  err_unreg:
-	acpi_processor_unregister_performance(&data->acpi_data, cpu);
+	acpi_processor_unregister_performance(cpu);
  err_free:
 	kfree(data);
 	acpi_io_data[cpu] = NULL;
@@ -332,8 +332,8 @@ acpi_cpufreq_cpu_exit (
 
 	if (data) {
 		acpi_io_data[policy->cpu] = NULL;
-		acpi_processor_unregister_performance(&data->acpi_data,
+		acpi_processor_unregister_performance(policy->cpu);
-		                                      policy->cpu);
+		kfree(policy->freq_table);
 		kfree(data);
 	}
 

drivers/cpufreq/integrator-cpufreq.c

@@ -98,11 +98,10 @@ static int integrator_set_target(struct cpufreq_policy *policy,
 	/* get current setting */
 	cm_osc = __raw_readl(cm_base + INTEGRATOR_HDR_OSC_OFFSET);
 
-	if (machine_is_integrator()) {
+	if (machine_is_integrator())
 		vco.s = (cm_osc >> 8) & 7;
-	} else if (machine_is_cintegrator()) {
+	else if (machine_is_cintegrator())
 		vco.s = 1;
-	}
 	vco.v = cm_osc & 255;
 	vco.r = 22;
 	freqs.old = icst_hz(&cclk_params, vco) / 1000;
@@ -163,11 +162,10 @@ static unsigned int integrator_get(unsigned int cpu)
 	/* detect memory etc. */
 	cm_osc = __raw_readl(cm_base + INTEGRATOR_HDR_OSC_OFFSET);
 
-	if (machine_is_integrator()) {
+	if (machine_is_integrator())
 		vco.s = (cm_osc >> 8) & 7;
-	} else {
+	else
 		vco.s = 1;
-	}
 	vco.v = cm_osc & 255;
 	vco.r = 22;
 
@@ -203,7 +201,7 @@ static int __init integrator_cpufreq_probe(struct platform_device *pdev)
 	struct resource *res;
 
 	res = platform_get_resource(pdev, IORESOURCE_MEM, 0);
-        if (!res)
+	if (!res)
 		return -ENODEV;
 
 	cm_base = devm_ioremap(&pdev->dev, res->start, resource_size(res));
@@ -234,6 +232,6 @@ static struct platform_driver integrator_cpufreq_driver = {
 module_platform_driver_probe(integrator_cpufreq_driver,
 			     integrator_cpufreq_probe);
 
-MODULE_AUTHOR ("Russell M. King");
+MODULE_AUTHOR("Russell M. King");
-MODULE_DESCRIPTION ("cpufreq driver for ARM Integrator CPUs");
+MODULE_DESCRIPTION("cpufreq driver for ARM Integrator CPUs");
-MODULE_LICENSE ("GPL");
+MODULE_LICENSE("GPL");

drivers/cpufreq/intel_pstate.c

@@ -484,12 +484,11 @@ static void __init intel_pstate_sysfs_expose_params(void)
 }
 /************************** sysfs end ************************/
 
-static void intel_pstate_hwp_enable(void)
+static void intel_pstate_hwp_enable(struct cpudata *cpudata)
 {
-	hwp_active++;
 	pr_info("intel_pstate: HWP enabled\n");
 
-	wrmsrl( MSR_PM_ENABLE, 0x1);
+	wrmsrl_on_cpu(cpudata->cpu, MSR_PM_ENABLE, 0x1);
 }
 
 static int byt_get_min_pstate(void)
@@ -522,7 +521,7 @@ static void byt_set_pstate(struct cpudata *cpudata, int pstate)
 	int32_t vid_fp;
 	u32 vid;
 
-	val = pstate << 8;
+	val = (u64)pstate << 8;
 	if (limits.no_turbo && !limits.turbo_disabled)
 		val |= (u64)1 << 32;
 
@@ -611,7 +610,7 @@ static void core_set_pstate(struct cpudata *cpudata, int pstate)
 {
 	u64 val;
 
-	val = pstate << 8;
+	val = (u64)pstate << 8;
 	if (limits.no_turbo && !limits.turbo_disabled)
 		val |= (u64)1 << 32;
 
@@ -909,6 +908,7 @@ static const struct x86_cpu_id intel_pstate_cpu_ids[] = {
 	ICPU(0x4c, byt_params),
 	ICPU(0x4e, core_params),
 	ICPU(0x4f, core_params),
+	ICPU(0x5e, core_params),
 	ICPU(0x56, core_params),
 	ICPU(0x57, knl_params),
 	{}
@@ -933,6 +933,10 @@ static int intel_pstate_init_cpu(unsigned int cpunum)
 	cpu = all_cpu_data[cpunum];
 
 	cpu->cpu = cpunum;
+
+	if (hwp_active)
+		intel_pstate_hwp_enable(cpu);
+
 	intel_pstate_get_cpu_pstates(cpu);
 
 	init_timer_deferrable(&cpu->timer);
@@ -1170,6 +1174,10 @@ static struct hw_vendor_info vendor_info[] = {
 	{1, "ORACLE", "X4270M3 ", PPC},
 	{1, "ORACLE", "X4270M2 ", PPC},
 	{1, "ORACLE", "X4170M2 ", PPC},
+	{1, "ORACLE", "X4170 M3", PPC},
+	{1, "ORACLE", "X4275 M3", PPC},
+	{1, "ORACLE", "X6-2    ", PPC},
+	{1, "ORACLE", "Sudbury ", PPC},
 	{0, "", ""},
 };
 
@@ -1246,7 +1254,7 @@ static int __init intel_pstate_init(void)
 		return -ENOMEM;
 
 	if (static_cpu_has_safe(X86_FEATURE_HWP) && !no_hwp)
-		intel_pstate_hwp_enable();
+		hwp_active++;
 
 	if (!hwp_active && hwp_only)
 		goto out;

drivers/cpufreq/mt8173-cpufreq.c

@@ -0,0 +1,527 @@
+/*
+ * Copyright (c) 2015 Linaro Ltd.
+ * Author: Pi-Cheng Chen <pi-cheng.chen@linaro.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it will be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ */
+
+#include <linux/clk.h>
+#include <linux/cpu.h>
+#include <linux/cpu_cooling.h>
+#include <linux/cpufreq.h>
+#include <linux/cpumask.h>
+#include <linux/of.h>
+#include <linux/platform_device.h>
+#include <linux/pm_opp.h>
+#include <linux/regulator/consumer.h>
+#include <linux/slab.h>
+#include <linux/thermal.h>
+
+#define MIN_VOLT_SHIFT		(100000)
+#define MAX_VOLT_SHIFT		(200000)
+#define MAX_VOLT_LIMIT		(1150000)
+#define VOLT_TOL		(10000)
+
+/*
+ * The struct mtk_cpu_dvfs_info holds necessary information for doing CPU DVFS
+ * on each CPU power/clock domain of Mediatek SoCs. Each CPU cluster in
+ * Mediatek SoCs has two voltage inputs, Vproc and Vsram. In some cases the two
+ * voltage inputs need to be controlled under a hardware limitation:
+ * 100mV < Vsram - Vproc < 200mV
+ *
+ * When scaling the clock frequency of a CPU clock domain, the clock source
+ * needs to be switched to another stable PLL clock temporarily until
+ * the original PLL becomes stable at target frequency.
+ */
+struct mtk_cpu_dvfs_info {
+	struct device *cpu_dev;
+	struct regulator *proc_reg;
+	struct regulator *sram_reg;
+	struct clk *cpu_clk;
+	struct clk *inter_clk;
+	struct thermal_cooling_device *cdev;
+	int intermediate_voltage;
+	bool need_voltage_tracking;
+};
+
+static int mtk_cpufreq_voltage_tracking(struct mtk_cpu_dvfs_info *info,
+					int new_vproc)
+{
+	struct regulator *proc_reg = info->proc_reg;
+	struct regulator *sram_reg = info->sram_reg;
+	int old_vproc, old_vsram, new_vsram, vsram, vproc, ret;
+
+	old_vproc = regulator_get_voltage(proc_reg);
+	old_vsram = regulator_get_voltage(sram_reg);
+	/* Vsram should not exceed the maximum allowed voltage of SoC. */
+	new_vsram = min(new_vproc + MIN_VOLT_SHIFT, MAX_VOLT_LIMIT);
+
+	if (old_vproc < new_vproc) {
+		/*
+		 * When scaling up voltages, Vsram and Vproc scale up step
+		 * by step. At each step, set Vsram to (Vproc + 200mV) first,
+		 * then set Vproc to (Vsram - 100mV).
+		 * Keep doing it until Vsram and Vproc hit target voltages.
+		 */
+		do {
+			old_vsram = regulator_get_voltage(sram_reg);
+			old_vproc = regulator_get_voltage(proc_reg);
+
+			vsram = min(new_vsram, old_vproc + MAX_VOLT_SHIFT);
+
+			if (vsram + VOLT_TOL >= MAX_VOLT_LIMIT) {
+				vsram = MAX_VOLT_LIMIT;
+
+				/*
+				 * If the target Vsram hits the maximum voltage,
+				 * try to set the exact voltage value first.
+				 */
+				ret = regulator_set_voltage(sram_reg, vsram,
+							    vsram);
+				if (ret)
+					ret = regulator_set_voltage(sram_reg,
+							vsram - VOLT_TOL,
+							vsram);
+
+				vproc = new_vproc;
+			} else {
+				ret = regulator_set_voltage(sram_reg, vsram,
+							    vsram + VOLT_TOL);
+
+				vproc = vsram - MIN_VOLT_SHIFT;
+			}
+			if (ret)
+				return ret;
+
+			ret = regulator_set_voltage(proc_reg, vproc,
+						    vproc + VOLT_TOL);
+			if (ret) {
+				regulator_set_voltage(sram_reg, old_vsram,
+						      old_vsram);
+				return ret;
+			}
+		} while (vproc < new_vproc || vsram < new_vsram);
+	} else if (old_vproc > new_vproc) {
+		/*
+		 * When scaling down voltages, Vsram and Vproc scale down step
+		 * by step. At each step, set Vproc to (Vsram - 200mV) first,
+		 * then set Vproc to (Vproc + 100mV).
+		 * Keep doing it until Vsram and Vproc hit target voltages.
+		 */
+		do {
+			old_vproc = regulator_get_voltage(proc_reg);
+			old_vsram = regulator_get_voltage(sram_reg);
+
+			vproc = max(new_vproc, old_vsram - MAX_VOLT_SHIFT);
+			ret = regulator_set_voltage(proc_reg, vproc,
+						    vproc + VOLT_TOL);
+			if (ret)
+				return ret;
+
+			if (vproc == new_vproc)
+				vsram = new_vsram;
+			else
+				vsram = max(new_vsram, vproc + MIN_VOLT_SHIFT);
+
+			if (vsram + VOLT_TOL >= MAX_VOLT_LIMIT) {
+				vsram = MAX_VOLT_LIMIT;
+
+				/*
+				 * If the target Vsram hits the maximum voltage,
+				 * try to set the exact voltage value first.
+				 */
+				ret = regulator_set_voltage(sram_reg, vsram,
+							    vsram);
+				if (ret)
+					ret = regulator_set_voltage(sram_reg,
+							vsram - VOLT_TOL,
+							vsram);
+			} else {
+				ret = regulator_set_voltage(sram_reg, vsram,
+							    vsram + VOLT_TOL);
+			}
+
+			if (ret) {
+				regulator_set_voltage(proc_reg, old_vproc,
+						      old_vproc);
+				return ret;
+			}
+		} while (vproc > new_vproc + VOLT_TOL ||
+			 vsram > new_vsram + VOLT_TOL);
+	}
+
+	return 0;
+}
+
+static int mtk_cpufreq_set_voltage(struct mtk_cpu_dvfs_info *info, int vproc)
+{
+	if (info->need_voltage_tracking)
+		return mtk_cpufreq_voltage_tracking(info, vproc);
+	else
+		return regulator_set_voltage(info->proc_reg, vproc,
+					     vproc + VOLT_TOL);
+}
+
+static int mtk_cpufreq_set_target(struct cpufreq_policy *policy,
+				  unsigned int index)
+{
+	struct cpufreq_frequency_table *freq_table = policy->freq_table;
+	struct clk *cpu_clk = policy->clk;
+	struct clk *armpll = clk_get_parent(cpu_clk);
+	struct mtk_cpu_dvfs_info *info = policy->driver_data;
+	struct device *cpu_dev = info->cpu_dev;
+	struct dev_pm_opp *opp;
+	long freq_hz, old_freq_hz;
+	int vproc, old_vproc, inter_vproc, target_vproc, ret;
+
+	inter_vproc = info->intermediate_voltage;
+
+	old_freq_hz = clk_get_rate(cpu_clk);
+	old_vproc = regulator_get_voltage(info->proc_reg);
+
+	freq_hz = freq_table[index].frequency * 1000;
+
+	rcu_read_lock();
+	opp = dev_pm_opp_find_freq_ceil(cpu_dev, &freq_hz);
+	if (IS_ERR(opp)) {
+		rcu_read_unlock();
+		pr_err("cpu%d: failed to find OPP for %ld\n",
+		       policy->cpu, freq_hz);
+		return PTR_ERR(opp);
+	}
+	vproc = dev_pm_opp_get_voltage(opp);
+	rcu_read_unlock();
+
+	/*
+	 * If the new voltage or the intermediate voltage is higher than the
+	 * current voltage, scale up voltage first.
+	 */
+	target_vproc = (inter_vproc > vproc) ? inter_vproc : vproc;
+	if (old_vproc < target_vproc) {
+		ret = mtk_cpufreq_set_voltage(info, target_vproc);
+		if (ret) {
+			pr_err("cpu%d: failed to scale up voltage!\n",
+			       policy->cpu);
+			mtk_cpufreq_set_voltage(info, old_vproc);
+			return ret;
+		}
+	}
+
+	/* Reparent the CPU clock to intermediate clock. */
+	ret = clk_set_parent(cpu_clk, info->inter_clk);
+	if (ret) {
+		pr_err("cpu%d: failed to re-parent cpu clock!\n",
+		       policy->cpu);
+		mtk_cpufreq_set_voltage(info, old_vproc);
+		WARN_ON(1);
+		return ret;
+	}
+
+	/* Set the original PLL to target rate. */
+	ret = clk_set_rate(armpll, freq_hz);
+	if (ret) {
+		pr_err("cpu%d: failed to scale cpu clock rate!\n",
+		       policy->cpu);
+		clk_set_parent(cpu_clk, armpll);
+		mtk_cpufreq_set_voltage(info, old_vproc);
+		return ret;
+	}
+
+	/* Set parent of CPU clock back to the original PLL. */
+	ret = clk_set_parent(cpu_clk, armpll);
+	if (ret) {
+		pr_err("cpu%d: failed to re-parent cpu clock!\n",
+		       policy->cpu);
+		mtk_cpufreq_set_voltage(info, inter_vproc);
+		WARN_ON(1);
+		return ret;
+	}
+
+	/*
+	 * If the new voltage is lower than the intermediate voltage or the
+	 * original voltage, scale down to the new voltage.
+	 */
+	if (vproc < inter_vproc || vproc < old_vproc) {
+		ret = mtk_cpufreq_set_voltage(info, vproc);
+		if (ret) {
+			pr_err("cpu%d: failed to scale down voltage!\n",
+			       policy->cpu);
+			clk_set_parent(cpu_clk, info->inter_clk);
+			clk_set_rate(armpll, old_freq_hz);
+			clk_set_parent(cpu_clk, armpll);
+			return ret;
+		}
+	}
+
+	return 0;
+}
+
+static void mtk_cpufreq_ready(struct cpufreq_policy *policy)
+{
+	struct mtk_cpu_dvfs_info *info = policy->driver_data;
+	struct device_node *np = of_node_get(info->cpu_dev->of_node);
+
+	if (WARN_ON(!np))
+		return;
+
+	if (of_find_property(np, "#cooling-cells", NULL)) {
+		info->cdev = of_cpufreq_cooling_register(np,
+							 policy->related_cpus);
+
+		if (IS_ERR(info->cdev)) {
+			dev_err(info->cpu_dev,
+				"running cpufreq without cooling device: %ld\n",
+				PTR_ERR(info->cdev));
+
+			info->cdev = NULL;
+		}
+	}
+
+	of_node_put(np);
+}
+
+static int mtk_cpu_dvfs_info_init(struct mtk_cpu_dvfs_info *info, int cpu)
+{
+	struct device *cpu_dev;
+	struct regulator *proc_reg = ERR_PTR(-ENODEV);
+	struct regulator *sram_reg = ERR_PTR(-ENODEV);
+	struct clk *cpu_clk = ERR_PTR(-ENODEV);
+	struct clk *inter_clk = ERR_PTR(-ENODEV);
+	struct dev_pm_opp *opp;
+	unsigned long rate;
+	int ret;
+
+	cpu_dev = get_cpu_device(cpu);
+	if (!cpu_dev) {
+		pr_err("failed to get cpu%d device\n", cpu);
+		return -ENODEV;
+	}
+
+	cpu_clk = clk_get(cpu_dev, "cpu");
+	if (IS_ERR(cpu_clk)) {
+		if (PTR_ERR(cpu_clk) == -EPROBE_DEFER)
+			pr_warn("cpu clk for cpu%d not ready, retry.\n", cpu);
+		else
+			pr_err("failed to get cpu clk for cpu%d\n", cpu);
+
+		ret = PTR_ERR(cpu_clk);
+		return ret;
+	}
+
+	inter_clk = clk_get(cpu_dev, "intermediate");
+	if (IS_ERR(inter_clk)) {
+		if (PTR_ERR(inter_clk) == -EPROBE_DEFER)
+			pr_warn("intermediate clk for cpu%d not ready, retry.\n",
+				cpu);
+		else
+			pr_err("failed to get intermediate clk for cpu%d\n",
+			       cpu);
+
+		ret = PTR_ERR(inter_clk);
+		goto out_free_resources;
+	}
+
+	proc_reg = regulator_get_exclusive(cpu_dev, "proc");
+	if (IS_ERR(proc_reg)) {
+		if (PTR_ERR(proc_reg) == -EPROBE_DEFER)
+			pr_warn("proc regulator for cpu%d not ready, retry.\n",
+				cpu);
+		else
+			pr_err("failed to get proc regulator for cpu%d\n",
+			       cpu);
+
+		ret = PTR_ERR(proc_reg);
+		goto out_free_resources;
+	}
+
+	/* Both presence and absence of sram regulator are valid cases. */
+	sram_reg = regulator_get_exclusive(cpu_dev, "sram");
+
+	ret = of_init_opp_table(cpu_dev);
+	if (ret) {
+		pr_warn("no OPP table for cpu%d\n", cpu);
+		goto out_free_resources;
+	}
+
+	/* Search a safe voltage for intermediate frequency. */
+	rate = clk_get_rate(inter_clk);
+	rcu_read_lock();
+	opp = dev_pm_opp_find_freq_ceil(cpu_dev, &rate);
+	if (IS_ERR(opp)) {
+		rcu_read_unlock();
+		pr_err("failed to get intermediate opp for cpu%d\n", cpu);
+		ret = PTR_ERR(opp);
+		goto out_free_opp_table;
+	}
+	info->intermediate_voltage = dev_pm_opp_get_voltage(opp);
+	rcu_read_unlock();
+
+	info->cpu_dev = cpu_dev;
+	info->proc_reg = proc_reg;
+	info->sram_reg = IS_ERR(sram_reg) ? NULL : sram_reg;
+	info->cpu_clk = cpu_clk;
+	info->inter_clk = inter_clk;
+
+	/*
+	 * If SRAM regulator is present, software "voltage tracking" is needed
+	 * for this CPU power domain.
+	 */
+	info->need_voltage_tracking = !IS_ERR(sram_reg);
+
+	return 0;
+
+out_free_opp_table:
+	of_free_opp_table(cpu_dev);
+
+out_free_resources:
+	if (!IS_ERR(proc_reg))
+		regulator_put(proc_reg);
+	if (!IS_ERR(sram_reg))
+		regulator_put(sram_reg);
+	if (!IS_ERR(cpu_clk))
+		clk_put(cpu_clk);
+	if (!IS_ERR(inter_clk))
+		clk_put(inter_clk);
+
+	return ret;
+}
+
+static void mtk_cpu_dvfs_info_release(struct mtk_cpu_dvfs_info *info)
+{
+	if (!IS_ERR(info->proc_reg))
+		regulator_put(info->proc_reg);
+	if (!IS_ERR(info->sram_reg))
+		regulator_put(info->sram_reg);
+	if (!IS_ERR(info->cpu_clk))
+		clk_put(info->cpu_clk);
+	if (!IS_ERR(info->inter_clk))
+		clk_put(info->inter_clk);
+
+	of_free_opp_table(info->cpu_dev);
+}
+
+static int mtk_cpufreq_init(struct cpufreq_policy *policy)
+{
+	struct mtk_cpu_dvfs_info *info;
+	struct cpufreq_frequency_table *freq_table;
+	int ret;
+
+	info = kzalloc(sizeof(*info), GFP_KERNEL);
+	if (!info)
+		return -ENOMEM;
+
+	ret = mtk_cpu_dvfs_info_init(info, policy->cpu);
+	if (ret) {
+		pr_err("%s failed to initialize dvfs info for cpu%d\n",
+		       __func__, policy->cpu);
+		goto out_free_dvfs_info;
+	}
+
+	ret = dev_pm_opp_init_cpufreq_table(info->cpu_dev, &freq_table);
+	if (ret) {
+		pr_err("failed to init cpufreq table for cpu%d: %d\n",
+		       policy->cpu, ret);
+		goto out_release_dvfs_info;
+	}
+
+	ret = cpufreq_table_validate_and_show(policy, freq_table);
+	if (ret) {
+		pr_err("%s: invalid frequency table: %d\n", __func__, ret);
+		goto out_free_cpufreq_table;
+	}
+
+	/* CPUs in the same cluster share a clock and power domain. */
+	cpumask_copy(policy->cpus, &cpu_topology[policy->cpu].core_sibling);
+	policy->driver_data = info;
+	policy->clk = info->cpu_clk;
+
+	return 0;
+
+out_free_cpufreq_table:
+	dev_pm_opp_free_cpufreq_table(info->cpu_dev, &freq_table);
+
+out_release_dvfs_info:
+	mtk_cpu_dvfs_info_release(info);
+
+out_free_dvfs_info:
+	kfree(info);
+
+	return ret;
+}
+
+static int mtk_cpufreq_exit(struct cpufreq_policy *policy)
+{
+	struct mtk_cpu_dvfs_info *info = policy->driver_data;
+
+	cpufreq_cooling_unregister(info->cdev);
+	dev_pm_opp_free_cpufreq_table(info->cpu_dev, &policy->freq_table);
+	mtk_cpu_dvfs_info_release(info);
+	kfree(info);
+
+	return 0;
+}
+
+static struct cpufreq_driver mt8173_cpufreq_driver = {
+	.flags = CPUFREQ_STICKY | CPUFREQ_NEED_INITIAL_FREQ_CHECK,
+	.verify = cpufreq_generic_frequency_table_verify,
+	.target_index = mtk_cpufreq_set_target,
+	.get = cpufreq_generic_get,
+	.init = mtk_cpufreq_init,
+	.exit = mtk_cpufreq_exit,
+	.ready = mtk_cpufreq_ready,
+	.name = "mtk-cpufreq",
+	.attr = cpufreq_generic_attr,
+};
+
+static int mt8173_cpufreq_probe(struct platform_device *pdev)
+{
+	int ret;
+
+	ret = cpufreq_register_driver(&mt8173_cpufreq_driver);
+	if (ret)
+		pr_err("failed to register mtk cpufreq driver\n");
+
+	return ret;
+}
+
+static struct platform_driver mt8173_cpufreq_platdrv = {
+	.driver = {
+		.name	= "mt8173-cpufreq",
+	},
+	.probe		= mt8173_cpufreq_probe,
+};
+
+static int mt8173_cpufreq_driver_init(void)
+{
+	struct platform_device *pdev;
+	int err;
+
+	if (!of_machine_is_compatible("mediatek,mt8173"))
+		return -ENODEV;
+
+	err = platform_driver_register(&mt8173_cpufreq_platdrv);
+	if (err)
+		return err;
+
+	/*
+	 * Since there's no place to hold device registration code and no
+	 * device tree based way to match cpufreq driver yet, both the driver
+	 * and the device registration codes are put here to handle defer
+	 * probing.
+	 */
+	pdev = platform_device_register_simple("mt8173-cpufreq", -1, NULL, 0);
+	if (IS_ERR(pdev)) {
+		pr_err("failed to register mtk-cpufreq platform device\n");
+		return PTR_ERR(pdev);
+	}
+
+	return 0;
+}
+device_initcall(mt8173_cpufreq_driver_init);

drivers/cpufreq/powernow-k7.c

@@ -421,7 +421,7 @@ static int powernow_acpi_init(void)
 	return 0;
 
 err2:
-	acpi_processor_unregister_performance(acpi_processor_perf, 0);
+	acpi_processor_unregister_performance(0);
 err1:
 	free_cpumask_var(acpi_processor_perf->shared_cpu_map);
 err05:
@@ -661,7 +661,7 @@ static int powernow_cpu_exit(struct cpufreq_policy *policy)
 {
 #ifdef CONFIG_X86_POWERNOW_K7_ACPI
 	if (acpi_processor_perf) {
-		acpi_processor_unregister_performance(acpi_processor_perf, 0);
+		acpi_processor_unregister_performance(0);
 		free_cpumask_var(acpi_processor_perf->shared_cpu_map);
 		kfree(acpi_processor_perf);
 	}

drivers/cpufreq/powernow-k8.c

@@ -795,7 +795,7 @@ err_out_mem:
 	kfree(powernow_table);
 
 err_out:
-	acpi_processor_unregister_performance(&data->acpi_data, data->cpu);
+	acpi_processor_unregister_performance(data->cpu);
 
 	/* data->acpi_data.state_count informs us at ->exit()
 	 * whether ACPI was used */
@@ -863,8 +863,7 @@ static int fill_powernow_table_fidvid(struct powernow_k8_data *data,
 static void powernow_k8_cpu_exit_acpi(struct powernow_k8_data *data)
 {
 	if (data->acpi_data.state_count)
-		acpi_processor_unregister_performance(&data->acpi_data,
+		acpi_processor_unregister_performance(data->cpu);
-				data->cpu);
 	free_cpumask_var(data->acpi_data.shared_cpu_map);
 }
 

drivers/cpufreq/powernv-cpufreq.c

@@ -27,20 +27,31 @@
 #include <linux/smp.h>
 #include <linux/of.h>
 #include <linux/reboot.h>
+#include <linux/slab.h>
 
 #include <asm/cputhreads.h>
 #include <asm/firmware.h>
 #include <asm/reg.h>
 #include <asm/smp.h> /* Required for cpu_sibling_mask() in UP configs */
+#include <asm/opal.h>
 
 #define POWERNV_MAX_PSTATES	256
 #define PMSR_PSAFE_ENABLE	(1UL << 30)
 #define PMSR_SPR_EM_DISABLE	(1UL << 31)
 #define PMSR_MAX(x)		((x >> 32) & 0xFF)
-#define PMSR_LP(x)		((x >> 48) & 0xFF)
 
 static struct cpufreq_frequency_table powernv_freqs[POWERNV_MAX_PSTATES+1];
-static bool rebooting, throttled;
+static bool rebooting, throttled, occ_reset;
+
+static struct chip {
+	unsigned int id;
+	bool throttled;
+	cpumask_t mask;
+	struct work_struct throttle;
+	bool restore;
+} *chips;
+
+static int nr_chips;
 
 /*
  * Note: The set of pstates consists of contiguous integers, the
@@ -298,28 +309,35 @@ static inline unsigned int get_nominal_index(void)
 	return powernv_pstate_info.max - powernv_pstate_info.nominal;
 }
 
-static void powernv_cpufreq_throttle_check(unsigned int cpu)
+static void powernv_cpufreq_throttle_check(void *data)
 {
+	unsigned int cpu = smp_processor_id();
 	unsigned long pmsr;
-	int pmsr_pmax, pmsr_lp;
+	int pmsr_pmax, i;
 
 	pmsr = get_pmspr(SPRN_PMSR);
 
+	for (i = 0; i < nr_chips; i++)
+		if (chips[i].id == cpu_to_chip_id(cpu))
+			break;
+
 	/* Check for Pmax Capping */
 	pmsr_pmax = (s8)PMSR_MAX(pmsr);
 	if (pmsr_pmax != powernv_pstate_info.max) {
-		throttled = true;
+		if (chips[i].throttled)
-		pr_info("CPU %d Pmax is reduced to %d\n", cpu, pmsr_pmax);
+			goto next;
-		pr_info("Max allowed Pstate is capped\n");
+		chips[i].throttled = true;
+		pr_info("CPU %d on Chip %u has Pmax reduced to %d\n", cpu,
+			chips[i].id, pmsr_pmax);
+	} else if (chips[i].throttled) {
+		chips[i].throttled = false;
+		pr_info("CPU %d on Chip %u has Pmax restored to %d\n", cpu,
+			chips[i].id, pmsr_pmax);
 	}
 
-	/*
+	/* Check if Psafe_mode_active is set in PMSR. */
-	 * Check for Psafe by reading LocalPstate
+next:
-	 * or check if Psafe_mode_active is set in PMSR.
+	if (pmsr & PMSR_PSAFE_ENABLE) {
-	 */
-	pmsr_lp = (s8)PMSR_LP(pmsr);
-	if ((pmsr_lp < powernv_pstate_info.min) ||
-				(pmsr & PMSR_PSAFE_ENABLE)) {
 		throttled = true;
 		pr_info("Pstate set to safe frequency\n");
 	}
@@ -350,7 +368,7 @@ static int powernv_cpufreq_target_index(struct cpufreq_policy *policy,
 		return 0;
 
 	if (!throttled)
-		powernv_cpufreq_throttle_check(smp_processor_id());
+		powernv_cpufreq_throttle_check(NULL);
 
 	freq_data.pstate_id = powernv_freqs[new_index].driver_data;
 
@@ -395,6 +413,119 @@ static struct notifier_block powernv_cpufreq_reboot_nb = {
 	.notifier_call = powernv_cpufreq_reboot_notifier,
 };
 
+void powernv_cpufreq_work_fn(struct work_struct *work)
+{
+	struct chip *chip = container_of(work, struct chip, throttle);
+	unsigned int cpu;
+	cpumask_var_t mask;
+
+	smp_call_function_any(&chip->mask,
+			      powernv_cpufreq_throttle_check, NULL, 0);
+
+	if (!chip->restore)
+		return;
+
+	chip->restore = false;
+	cpumask_copy(mask, &chip->mask);
+	for_each_cpu_and(cpu, mask, cpu_online_mask) {
+		int index, tcpu;
+		struct cpufreq_policy policy;
+
+		cpufreq_get_policy(&policy, cpu);
+		cpufreq_frequency_table_target(&policy, policy.freq_table,
+					       policy.cur,
+					       CPUFREQ_RELATION_C, &index);
+		powernv_cpufreq_target_index(&policy, index);
+		for_each_cpu(tcpu, policy.cpus)
+			cpumask_clear_cpu(tcpu, mask);
+	}
+}
+
+static char throttle_reason[][30] = {
+					"No throttling",
+					"Power Cap",
+					"Processor Over Temperature",
+					"Power Supply Failure",
+					"Over Current",
+					"OCC Reset"
+				     };
+
+static int powernv_cpufreq_occ_msg(struct notifier_block *nb,
+				   unsigned long msg_type, void *_msg)
+{
+	struct opal_msg *msg = _msg;
+	struct opal_occ_msg omsg;
+	int i;
+
+	if (msg_type != OPAL_MSG_OCC)
+		return 0;
+
+	omsg.type = be64_to_cpu(msg->params[0]);
+
+	switch (omsg.type) {
+	case OCC_RESET:
+		occ_reset = true;
+		pr_info("OCC (On Chip Controller - enforces hard thermal/power limits) Resetting\n");
+		/*
+		 * powernv_cpufreq_throttle_check() is called in
+		 * target() callback which can detect the throttle state
+		 * for governors like ondemand.
+		 * But static governors will not call target() often thus
+		 * report throttling here.
+		 */
+		if (!throttled) {
+			throttled = true;
+			pr_crit("CPU frequency is throttled for duration\n");
+		}
+
+		break;
+	case OCC_LOAD:
+		pr_info("OCC Loading, CPU frequency is throttled until OCC is started\n");
+		break;
+	case OCC_THROTTLE:
+		omsg.chip = be64_to_cpu(msg->params[1]);
+		omsg.throttle_status = be64_to_cpu(msg->params[2]);
+
+		if (occ_reset) {
+			occ_reset = false;
+			throttled = false;
+			pr_info("OCC Active, CPU frequency is no longer throttled\n");
+
+			for (i = 0; i < nr_chips; i++) {
+				chips[i].restore = true;
+				schedule_work(&chips[i].throttle);
+			}
+
+			return 0;
+		}
+
+		if (omsg.throttle_status &&
+		    omsg.throttle_status <= OCC_MAX_THROTTLE_STATUS)
+			pr_info("OCC: Chip %u Pmax reduced due to %s\n",
+				(unsigned int)omsg.chip,
+				throttle_reason[omsg.throttle_status]);
+		else if (!omsg.throttle_status)
+			pr_info("OCC: Chip %u %s\n", (unsigned int)omsg.chip,
+				throttle_reason[omsg.throttle_status]);
+		else
+			return 0;
+
+		for (i = 0; i < nr_chips; i++)
+			if (chips[i].id == omsg.chip) {
+				if (!omsg.throttle_status)
+					chips[i].restore = true;
+				schedule_work(&chips[i].throttle);
+			}
+	}
+	return 0;
+}
+
+static struct notifier_block powernv_cpufreq_opal_nb = {
+	.notifier_call	= powernv_cpufreq_occ_msg,
+	.next		= NULL,
+	.priority	= 0,
+};
+
 static void powernv_cpufreq_stop_cpu(struct cpufreq_policy *policy)
 {
 	struct powernv_smp_call_data freq_data;
@@ -414,6 +545,36 @@ static struct cpufreq_driver powernv_cpufreq_driver = {
 	.attr		= powernv_cpu_freq_attr,
 };
 
+static int init_chip_info(void)
+{
+	unsigned int chip[256];
+	unsigned int cpu, i;
+	unsigned int prev_chip_id = UINT_MAX;
+
+	for_each_possible_cpu(cpu) {
+		unsigned int id = cpu_to_chip_id(cpu);
+
+		if (prev_chip_id != id) {
+			prev_chip_id = id;
+			chip[nr_chips++] = id;
+		}
+	}
+
+	chips = kmalloc_array(nr_chips, sizeof(struct chip), GFP_KERNEL);
+	if (!chips)
+		return -ENOMEM;
+
+	for (i = 0; i < nr_chips; i++) {
+		chips[i].id = chip[i];
+		chips[i].throttled = false;
+		cpumask_copy(&chips[i].mask, cpumask_of_node(chip[i]));
+		INIT_WORK(&chips[i].throttle, powernv_cpufreq_work_fn);
+		chips[i].restore = false;
+	}
+
+	return 0;
+}
+
 static int __init powernv_cpufreq_init(void)
 {
 	int rc = 0;
@@ -429,7 +590,13 @@ static int __init powernv_cpufreq_init(void)
 		return rc;
 	}
 
+	/* Populate chip info */
+	rc = init_chip_info();
+	if (rc)
+		return rc;
+
 	register_reboot_notifier(&powernv_cpufreq_reboot_nb);
+	opal_message_notifier_register(OPAL_MSG_OCC, &powernv_cpufreq_opal_nb);
 	return cpufreq_register_driver(&powernv_cpufreq_driver);
 }
 module_init(powernv_cpufreq_init);
@@ -437,6 +604,8 @@ module_init(powernv_cpufreq_init);
 static void __exit powernv_cpufreq_exit(void)
 {
 	unregister_reboot_notifier(&powernv_cpufreq_reboot_nb);
+	opal_message_notifier_unregister(OPAL_MSG_OCC,
+					 &powernv_cpufreq_opal_nb);
 	cpufreq_unregister_driver(&powernv_cpufreq_driver);
 }
 module_exit(powernv_cpufreq_exit);

drivers/cpufreq/ppc_cbe_cpufreq_pmi.c

@@ -97,8 +97,8 @@ static int pmi_notifier(struct notifier_block *nb,
 	struct cpufreq_frequency_table *cbe_freqs;
 	u8 node;
 
-	/* Should this really be called for CPUFREQ_ADJUST, CPUFREQ_INCOMPATIBLE
+	/* Should this really be called for CPUFREQ_ADJUST and CPUFREQ_NOTIFY
-	 * and CPUFREQ_NOTIFY policy events?)
+	 * policy events?)
 	 */
 	if (event == CPUFREQ_START)
 		return 0;

drivers/cpufreq/sfi-cpufreq.c

@@ -45,12 +45,10 @@ static int sfi_parse_freq(struct sfi_table_header *table)
 	pentry = (struct sfi_freq_table_entry *)sb->pentry;
 	totallen = num_freq_table_entries * sizeof(*pentry);
 
-	sfi_cpufreq_array = kzalloc(totallen, GFP_KERNEL);
+	sfi_cpufreq_array = kmemdup(pentry, totallen, GFP_KERNEL);
 	if (!sfi_cpufreq_array)
 		return -ENOMEM;
 
-	memcpy(sfi_cpufreq_array, pentry, totallen);
-
 	return 0;
 }
 

drivers/cpufreq/speedstep-lib.c

@@ -386,7 +386,7 @@ unsigned int speedstep_get_freqs(enum speedstep_processor processor,
 	unsigned int prev_speed;
 	unsigned int ret = 0;
 	unsigned long flags;
-	struct timeval tv1, tv2;
+	ktime_t tv1, tv2;
 
 	if ((!processor) || (!low_speed) || (!high_speed) || (!set_state))
 		return -EINVAL;
@@ -415,14 +415,14 @@ unsigned int speedstep_get_freqs(enum speedstep_processor processor,
 
 	/* start latency measurement */
 	if (transition_latency)
-		do_gettimeofday(&tv1);
+		tv1 = ktime_get();
 
 	/* switch to high state */
 	set_state(SPEEDSTEP_HIGH);
 
 	/* end latency measurement */
 	if (transition_latency)
-		do_gettimeofday(&tv2);
+		tv2 = ktime_get();
 
 	*high_speed = speedstep_get_frequency(processor);
 	if (!*high_speed) {
@@ -442,8 +442,7 @@ unsigned int speedstep_get_freqs(enum speedstep_processor processor,
 		set_state(SPEEDSTEP_LOW);
 
 	if (transition_latency) {
-		*transition_latency = (tv2.tv_sec - tv1.tv_sec) * USEC_PER_SEC +
+		*transition_latency = ktime_to_us(ktime_sub(tv2, tv1));
-			tv2.tv_usec - tv1.tv_usec;
 		pr_debug("transition latency is %u uSec\n", *transition_latency);
 
 		/* convert uSec to nSec and add 20% for safety reasons */

drivers/video/fbdev/pxafb.c

@@ -1668,7 +1668,6 @@ pxafb_freq_policy(struct notifier_block *nb, unsigned long val, void *data)
 
 	switch (val) {
 	case CPUFREQ_ADJUST:
-	case CPUFREQ_INCOMPATIBLE:
 		pr_debug("min dma period: %d ps, "
 			"new clock %d kHz\n", pxafb_display_dma_period(var),
 			policy->max);

drivers/video/fbdev/sa1100fb.c

@@ -1042,7 +1042,6 @@ sa1100fb_freq_policy(struct notifier_block *nb, unsigned long val,
 
 	switch (val) {
 	case CPUFREQ_ADJUST:
-	case CPUFREQ_INCOMPATIBLE:
 		dev_dbg(fbi->dev, "min dma period: %d ps, "
 			"new clock %d kHz\n", sa1100fb_min_dma_period(fbi),
 			policy->max);

drivers/xen/xen-acpi-processor.c

@@ -560,11 +560,9 @@ static int __init xen_acpi_processor_init(void)
 
 	return 0;
 err_unregister:
-	for_each_possible_cpu(i) {
+	for_each_possible_cpu(i)
-		struct acpi_processor_performance *perf;
+		acpi_processor_unregister_performance(i);
-		perf = per_cpu_ptr(acpi_perf_data, i);
+
-		acpi_processor_unregister_performance(perf, i);
-	}
 err_out:
 	/* Freeing a NULL pointer is OK: alloc_percpu zeroes. */
 	free_acpi_perf_data();
@@ -579,11 +577,9 @@ static void __exit xen_acpi_processor_exit(void)
 	kfree(acpi_ids_done);
 	kfree(acpi_id_present);
 	kfree(acpi_id_cst_present);
-	for_each_possible_cpu(i) {
+	for_each_possible_cpu(i)
-		struct acpi_processor_performance *perf;
+		acpi_processor_unregister_performance(i);
-		perf = per_cpu_ptr(acpi_perf_data, i);
+
-		acpi_processor_unregister_performance(perf, i);
-	}
 	free_acpi_perf_data();
 }
 

include/acpi/processor.h

@@ -228,10 +228,7 @@ extern int acpi_processor_preregister_performance(struct
 
 extern int acpi_processor_register_performance(struct acpi_processor_performance
 					       *performance, unsigned int cpu);
-extern void acpi_processor_unregister_performance(struct
+extern void acpi_processor_unregister_performance(unsigned int cpu);
-						  acpi_processor_performance
-						  *performance,
-						  unsigned int cpu);
 
 /* note: this locks both the calling module and the processor module
          if a _PPC object exists, rmmod is disallowed then */

include/linux/cpufreq.h

@@ -51,11 +51,9 @@ struct cpufreq_cpuinfo {
 	unsigned int		transition_latency;
 };
 
-struct cpufreq_real_policy {
+struct cpufreq_user_policy {
 	unsigned int		min;    /* in kHz */
 	unsigned int		max;    /* in kHz */
-	unsigned int		policy; /* see above */
-	struct cpufreq_governor	*governor; /* see below */
 };
 
 struct cpufreq_policy {
@@ -88,7 +86,7 @@ struct cpufreq_policy {
 	struct work_struct	update; /* if update_policy() needs to be
 					 * called, but you're in IRQ context */
 
-	struct cpufreq_real_policy	user_policy;
+	struct cpufreq_user_policy user_policy;
 	struct cpufreq_frequency_table	*freq_table;
 
 	struct list_head        policy_list;
@@ -369,11 +367,10 @@ static inline void cpufreq_resume(void) {}
 
 /* Policy Notifiers  */
 #define CPUFREQ_ADJUST			(0)
-#define CPUFREQ_INCOMPATIBLE		(1)
+#define CPUFREQ_NOTIFY			(1)
-#define CPUFREQ_NOTIFY			(2)
+#define CPUFREQ_START			(2)
-#define CPUFREQ_START			(3)
+#define CPUFREQ_CREATE_POLICY		(3)
-#define CPUFREQ_CREATE_POLICY		(4)
+#define CPUFREQ_REMOVE_POLICY		(4)
-#define CPUFREQ_REMOVE_POLICY		(5)
 
 #ifdef CONFIG_CPU_FREQ
 int cpufreq_register_notifier(struct notifier_block *nb, unsigned int list);