Merge branch 'pm-cpufreq'

Merge cpufreq updates for 5.19-rc1:

 - Fix cpufreq governor clean up code to avoid using kfree() directly
   to free kobject-based items (Kevin Hao).

 - Prepare cpufreq for powerpc's asm/prom.h cleanup (Christophe Leroy).

 - Make intel_pstate notify frequency invariance code when no_turbo is
   turned on and off (Chen Yu).

 - Add Sapphire Rapids OOB mode support to intel_pstate (Srinivas
   Pandruvada).

 - Make cpufreq avoid unnecessary frequency updates due to mismatch
   between hardware and the frequency table (Viresh Kumar).

 - Make remove_cpu_dev_symlink() clear the real_cpus mask to simplify
   code (Viresh Kumar).

 - Rearrange cpufreq_offline() and cpufreq_remove_dev() to make the
   calling convention for some driver callbacks consistent (Rafael
   Wysocki).

 - Avoid accessing half-initialized cpufreq policies from the show()
   and store() sysfs functions (Schspa Shi).

 - Rearrange cpufreq_offline() to make the calling convention for some
   driver callbacks consistent (Schspa Shi).

 - Update CPPC handling in cpufreq (Pierre Gondois):

   * Add per_cpu efficiency_class to the CPPC driver.
   * Make the CPPC driver Register EM based on efficiency class
     information.
   * Adjust _OSC for flexible address space in the ACPI platform
     initialization code and always set CPPC _OSC bits if CPPC_LIB is
     supported.
   * Assume no transition latency if no PCCT in the CPPC driver.
   * Add fast_switch and dvfs_possible_from_any_cpu support to the CPPC
     driver.

* pm-cpufreq:
  cpufreq: CPPC: Enable dvfs_possible_from_any_cpu
  cpufreq: CPPC: Enable fast_switch
  ACPI: CPPC: Assume no transition latency if no PCCT
  ACPI: bus: Set CPPC _OSC bits for all and when CPPC_LIB is supported
  ACPI: CPPC: Check _OSC for flexible address space
  cpufreq: make interface functions and lock holding state clear
  cpufreq: Abort show()/store() for half-initialized policies
  cpufreq: Rearrange locking in cpufreq_remove_dev()
  cpufreq: Split cpufreq_offline()
  cpufreq: Reorganize checks in cpufreq_offline()
  cpufreq: Clear real_cpus mask from remove_cpu_dev_symlink()
  cpufreq: intel_pstate: Support Sapphire Rapids OOB mode
  Revert "cpufreq: Fix possible race in cpufreq online error path"
  cpufreq: CPPC: Register EM based on efficiency class information
  cpufreq: CPPC: Add per_cpu efficiency_class
  cpufreq: Avoid unnecessary frequency updates due to mismatch
  cpufreq: Fix possible race in cpufreq online error path
  cpufreq: intel_pstate: Handle no_turbo in frequency invariance
  cpufreq: Prepare cleanup of powerpc's asm/prom.h
  cpufreq: governor: Use kobject release() method to free dbs_data

12 files changed, 368 insertions, 64 deletions

diff --git a/drivers/acpi/bus.c b/drivers/acpi/bus.c
index 3e58b613a2c4..1fc24f4fbcb4 100644
--- a/drivers/acpi/bus.c
+++ b/drivers/acpi/bus.c
@@ -279,6 +279,20 @@ bool osc_pc_lpi_support_confirmed;
 EXPORT_SYMBOL_GPL(osc_pc_lpi_support_confirmed);
 
 /*
+ * ACPI 6.2 Section 6.2.11.2 'Platform-Wide OSPM Capabilities':
+ *   Starting with ACPI Specification 6.2, all _CPC registers can be in
+ *   PCC, System Memory, System IO, or Functional Fixed Hardware address
+ *   spaces. OSPM support for this more flexible register space scheme is
+ *   indicated by the “Flexible Address Space for CPPC Registers” _OSC bit.
+ *
+ * Otherwise (cf ACPI 6.1, s8.4.7.1.1.X), _CPC registers must be in:
+ * - PCC or Functional Fixed Hardware address space if defined
+ * - SystemMemory address space (NULL register) if not defined
+ */
+bool osc_cpc_flexible_adr_space_confirmed;
+EXPORT_SYMBOL_GPL(osc_cpc_flexible_adr_space_confirmed);
+
+/*
  * ACPI 6.4 Operating System Capabilities for USB.
  */
 bool osc_sb_native_usb4_support_confirmed;
@@ -315,12 +329,15 @@ static void acpi_bus_osc_negotiate_platform_control(void)
 #endif
 #ifdef CONFIG_X86
 	capbuf[OSC_SUPPORT_DWORD] |= OSC_SB_GENERIC_INITIATOR_SUPPORT;
-	if (boot_cpu_has(X86_FEATURE_HWP)) {
-		capbuf[OSC_SUPPORT_DWORD] |= OSC_SB_CPC_SUPPORT;
-		capbuf[OSC_SUPPORT_DWORD] |= OSC_SB_CPCV2_SUPPORT;
-	}
 #endif
 
+#ifdef CONFIG_ACPI_CPPC_LIB
+	capbuf[OSC_SUPPORT_DWORD] |= OSC_SB_CPC_SUPPORT;
+	capbuf[OSC_SUPPORT_DWORD] |= OSC_SB_CPCV2_SUPPORT;
+#endif
+
+	capbuf[OSC_SUPPORT_DWORD] |= OSC_SB_CPC_FLEXIBLE_ADR_SPACE;
+
 	if (IS_ENABLED(CONFIG_SCHED_MC_PRIO))
 		capbuf[OSC_SUPPORT_DWORD] |= OSC_SB_CPC_DIVERSE_HIGH_SUPPORT;
 
@@ -341,10 +358,9 @@ static void acpi_bus_osc_negotiate_platform_control(void)
 		return;
 	}
 
-#ifdef CONFIG_X86
-	if (boot_cpu_has(X86_FEATURE_HWP))
-		osc_sb_cppc_not_supported = !(capbuf_ret[OSC_SUPPORT_DWORD] &
-				(OSC_SB_CPC_SUPPORT | OSC_SB_CPCV2_SUPPORT));
+#ifdef CONFIG_ACPI_CPPC_LIB
+	osc_sb_cppc_not_supported = !(capbuf_ret[OSC_SUPPORT_DWORD] &
+			(OSC_SB_CPC_SUPPORT | OSC_SB_CPCV2_SUPPORT));
 #endif
 
 	/*
@@ -366,6 +382,8 @@ static void acpi_bus_osc_negotiate_platform_control(void)
 			capbuf_ret[OSC_SUPPORT_DWORD] & OSC_SB_PCLPI_SUPPORT;
 		osc_sb_native_usb4_support_confirmed =
 			capbuf_ret[OSC_SUPPORT_DWORD] & OSC_SB_NATIVE_USB4_SUPPORT;
+		osc_cpc_flexible_adr_space_confirmed =
+			capbuf_ret[OSC_SUPPORT_DWORD] & OSC_SB_CPC_FLEXIBLE_ADR_SPACE;
 	}
 
 	kfree(context.ret.pointer);
diff --git a/drivers/acpi/cppc_acpi.c b/drivers/acpi/cppc_acpi.c
index bc1454789a06..3b299b28a8af 100644
--- a/drivers/acpi/cppc_acpi.c
+++ b/drivers/acpi/cppc_acpi.c
@@ -100,6 +100,16 @@ static DEFINE_PER_CPU(struct cpc_desc *, cpc_desc_ptr);
 				(cpc)->cpc_entry.reg.space_id ==	\
 				ACPI_ADR_SPACE_PLATFORM_COMM)
 
+/* Check if a CPC register is in SystemMemory */
+#define CPC_IN_SYSTEM_MEMORY(cpc) ((cpc)->type == ACPI_TYPE_BUFFER &&	\
+				(cpc)->cpc_entry.reg.space_id ==	\
+				ACPI_ADR_SPACE_SYSTEM_MEMORY)
+
+/* Check if a CPC register is in SystemIo */
+#define CPC_IN_SYSTEM_IO(cpc) ((cpc)->type == ACPI_TYPE_BUFFER &&	\
+				(cpc)->cpc_entry.reg.space_id ==	\
+				ACPI_ADR_SPACE_SYSTEM_IO)
+
 /* Evaluates to True if reg is a NULL register descriptor */
 #define IS_NULL_REG(reg) ((reg)->space_id ==  ACPI_ADR_SPACE_SYSTEM_MEMORY && \
 				(reg)->address == 0 &&			\
@@ -424,6 +434,24 @@ bool acpi_cpc_valid(void)
 }
 EXPORT_SYMBOL_GPL(acpi_cpc_valid);
 
+bool cppc_allow_fast_switch(void)
+{
+	struct cpc_register_resource *desired_reg;
+	struct cpc_desc *cpc_ptr;
+	int cpu;
+
+	for_each_possible_cpu(cpu) {
+		cpc_ptr = per_cpu(cpc_desc_ptr, cpu);
+		desired_reg = &cpc_ptr->cpc_regs[DESIRED_PERF];
+		if (!CPC_IN_SYSTEM_MEMORY(desired_reg) &&
+				!CPC_IN_SYSTEM_IO(desired_reg))
+			return false;
+	}
+
+	return true;
+}
+EXPORT_SYMBOL_GPL(cppc_allow_fast_switch);
+
 /**
  * acpi_get_psd_map - Map the CPUs in the freq domain of a given cpu
  * @cpu: Find all CPUs that share a domain with cpu.
@@ -736,6 +764,11 @@ int acpi_cppc_processor_probe(struct acpi_processor *pr)
 				if (gas_t->address) {
 					void __iomem *addr;
 
+					if (!osc_cpc_flexible_adr_space_confirmed) {
+						pr_debug("Flexible address space capability not supported\n");
+						goto out_free;
+					}
+
 					addr = ioremap(gas_t->address, gas_t->bit_width/8);
 					if (!addr)
 						goto out_free;
@@ -758,6 +791,10 @@ int acpi_cppc_processor_probe(struct acpi_processor *pr)
 						 gas_t->address);
 					goto out_free;
 				}
+				if (!osc_cpc_flexible_adr_space_confirmed) {
+					pr_debug("Flexible address space capability not supported\n");
+					goto out_free;
+				}
 			} else {
 				if (gas_t->space_id != ACPI_ADR_SPACE_FIXED_HARDWARE || !cpc_ffh_supported()) {
 					/* Support only PCC, SystemMemory, SystemIO, and FFH type regs. */
@@ -1447,6 +1484,9 @@ EXPORT_SYMBOL_GPL(cppc_set_perf);
  * transition latency for performance change requests. The closest we have
  * is the timing information from the PCCT tables which provides the info
  * on the number and frequency of PCC commands the platform can handle.
+ *
+ * If desired_reg is in the SystemMemory or SystemIo ACPI address space,
+ * then assume there is no latency.
  */
 unsigned int cppc_get_transition_latency(int cpu_num)
 {
@@ -1472,7 +1512,9 @@ unsigned int cppc_get_transition_latency(int cpu_num)
 		return CPUFREQ_ETERNAL;
 
 	desired_reg = &cpc_desc->cpc_regs[DESIRED_PERF];
-	if (!CPC_IN_PCC(desired_reg))
+	if (CPC_IN_SYSTEM_MEMORY(desired_reg) || CPC_IN_SYSTEM_IO(desired_reg))
+		return 0;
+	else if (!CPC_IN_PCC(desired_reg))
 		return CPUFREQ_ETERNAL;
 
 	if (pcc_ss_id < 0)
diff --git a/drivers/cpufreq/cppc_cpufreq.c b/drivers/cpufreq/cppc_cpufreq.c
index 82d370ae6a4a..d092c9bb4ba3 100644
--- a/drivers/cpufreq/cppc_cpufreq.c
+++ b/drivers/cpufreq/cppc_cpufreq.c
@@ -389,6 +389,27 @@ static int cppc_cpufreq_set_target(struct cpufreq_policy *policy,
 	return ret;
 }
 
+static unsigned int cppc_cpufreq_fast_switch(struct cpufreq_policy *policy,
+					      unsigned int target_freq)
+{
+	struct cppc_cpudata *cpu_data = policy->driver_data;
+	unsigned int cpu = policy->cpu;
+	u32 desired_perf;
+	int ret;
+
+	desired_perf = cppc_cpufreq_khz_to_perf(cpu_data, target_freq);
+	cpu_data->perf_ctrls.desired_perf = desired_perf;
+	ret = cppc_set_perf(cpu, &cpu_data->perf_ctrls);
+
+	if (ret) {
+		pr_debug("Failed to set target on CPU:%d. ret:%d\n",
+			 cpu, ret);
+		return 0;
+	}
+
+	return target_freq;
+}
+
 static int cppc_verify_policy(struct cpufreq_policy_data *policy)
 {
 	cpufreq_verify_within_cpu_limits(policy);
@@ -420,12 +441,197 @@ static unsigned int cppc_cpufreq_get_transition_delay_us(unsigned int cpu)
 	return cppc_get_transition_latency(cpu) / NSEC_PER_USEC;
 }
 
+static DEFINE_PER_CPU(unsigned int, efficiency_class);
+static void cppc_cpufreq_register_em(struct cpufreq_policy *policy);
+
+/* Create an artificial performance state every CPPC_EM_CAP_STEP capacity unit. */
+#define CPPC_EM_CAP_STEP	(20)
+/* Increase the cost value by CPPC_EM_COST_STEP every performance state. */
+#define CPPC_EM_COST_STEP	(1)
+/* Add a cost gap correspnding to the energy of 4 CPUs. */
+#define CPPC_EM_COST_GAP	(4 * SCHED_CAPACITY_SCALE * CPPC_EM_COST_STEP \
+				/ CPPC_EM_CAP_STEP)
+
+static unsigned int get_perf_level_count(struct cpufreq_policy *policy)
+{
+	struct cppc_perf_caps *perf_caps;
+	unsigned int min_cap, max_cap;
+	struct cppc_cpudata *cpu_data;
+	int cpu = policy->cpu;
+
+	cpu_data = policy->driver_data;
+	perf_caps = &cpu_data->perf_caps;
+	max_cap = arch_scale_cpu_capacity(cpu);
+	min_cap = div_u64(max_cap * perf_caps->lowest_perf, perf_caps->highest_perf);
+	if ((min_cap == 0) || (max_cap < min_cap))
+		return 0;
+	return 1 + max_cap / CPPC_EM_CAP_STEP - min_cap / CPPC_EM_CAP_STEP;
+}
+
+/*
+ * The cost is defined as:
+ *   cost = power * max_frequency / frequency
+ */
+static inline unsigned long compute_cost(int cpu, int step)
+{
+	return CPPC_EM_COST_GAP * per_cpu(efficiency_class, cpu) +
+			step * CPPC_EM_COST_STEP;
+}
+
+static int cppc_get_cpu_power(struct device *cpu_dev,
+		unsigned long *power, unsigned long *KHz)
+{
+	unsigned long perf_step, perf_prev, perf, perf_check;
+	unsigned int min_step, max_step, step, step_check;
+	unsigned long prev_freq = *KHz;
+	unsigned int min_cap, max_cap;
+	struct cpufreq_policy *policy;
+
+	struct cppc_perf_caps *perf_caps;
+	struct cppc_cpudata *cpu_data;
+
+	policy = cpufreq_cpu_get_raw(cpu_dev->id);
+	cpu_data = policy->driver_data;
+	perf_caps = &cpu_data->perf_caps;
+	max_cap = arch_scale_cpu_capacity(cpu_dev->id);
+	min_cap = div_u64(max_cap * perf_caps->lowest_perf,
+			perf_caps->highest_perf);
+
+	perf_step = CPPC_EM_CAP_STEP * perf_caps->highest_perf / max_cap;
+	min_step = min_cap / CPPC_EM_CAP_STEP;
+	max_step = max_cap / CPPC_EM_CAP_STEP;
+
+	perf_prev = cppc_cpufreq_khz_to_perf(cpu_data, *KHz);
+	step = perf_prev / perf_step;
+
+	if (step > max_step)
+		return -EINVAL;
+
+	if (min_step == max_step) {
+		step = max_step;
+		perf = perf_caps->highest_perf;
+	} else if (step < min_step) {
+		step = min_step;
+		perf = perf_caps->lowest_perf;
+	} else {
+		step++;
+		if (step == max_step)
+			perf = perf_caps->highest_perf;
+		else
+			perf = step * perf_step;
+	}
+
+	*KHz = cppc_cpufreq_perf_to_khz(cpu_data, perf);
+	perf_check = cppc_cpufreq_khz_to_perf(cpu_data, *KHz);
+	step_check = perf_check / perf_step;
+
+	/*
+	 * To avoid bad integer approximation, check that new frequency value
+	 * increased and that the new frequency will be converted to the
+	 * desired step value.
+	 */
+	while ((*KHz == prev_freq) || (step_check != step)) {
+		perf++;
+		*KHz = cppc_cpufreq_perf_to_khz(cpu_data, perf);
+		perf_check = cppc_cpufreq_khz_to_perf(cpu_data, *KHz);
+		step_check = perf_check / perf_step;
+	}
+
+	/*
+	 * With an artificial EM, only the cost value is used. Still the power
+	 * is populated such as 0 < power < EM_MAX_POWER. This allows to add
+	 * more sense to the artificial performance states.
+	 */
+	*power = compute_cost(cpu_dev->id, step);
+
+	return 0;
+}
+
+static int cppc_get_cpu_cost(struct device *cpu_dev, unsigned long KHz,
+		unsigned long *cost)
+{
+	unsigned long perf_step, perf_prev;
+	struct cppc_perf_caps *perf_caps;
+	struct cpufreq_policy *policy;
+	struct cppc_cpudata *cpu_data;
+	unsigned int max_cap;
+	int step;
+
+	policy = cpufreq_cpu_get_raw(cpu_dev->id);
+	cpu_data = policy->driver_data;
+	perf_caps = &cpu_data->perf_caps;
+	max_cap = arch_scale_cpu_capacity(cpu_dev->id);
+
+	perf_prev = cppc_cpufreq_khz_to_perf(cpu_data, KHz);
+	perf_step = CPPC_EM_CAP_STEP * perf_caps->highest_perf / max_cap;
+	step = perf_prev / perf_step;
+
+	*cost = compute_cost(cpu_dev->id, step);
+
+	return 0;
+}
+
+static int populate_efficiency_class(void)
+{
+	struct acpi_madt_generic_interrupt *gicc;
+	DECLARE_BITMAP(used_classes, 256) = {};
+	int class, cpu, index;
+
+	for_each_possible_cpu(cpu) {
+		gicc = acpi_cpu_get_madt_gicc(cpu);
+		class = gicc->efficiency_class;
+		bitmap_set(used_classes, class, 1);
+	}
+
+	if (bitmap_weight(used_classes, 256) <= 1) {
+		pr_debug("Efficiency classes are all equal (=%d). "
+			"No EM registered", class);
+		return -EINVAL;
+	}
+
+	/*
+	 * Squeeze efficiency class values on [0:#efficiency_class-1].
+	 * Values are per spec in [0:255].
+	 */
+	index = 0;
+	for_each_set_bit(class, used_classes, 256) {
+		for_each_possible_cpu(cpu) {
+			gicc = acpi_cpu_get_madt_gicc(cpu);
+			if (gicc->efficiency_class == class)
+				per_cpu(efficiency_class, cpu) = index;
+		}
+		index++;
+	}
+	cppc_cpufreq_driver.register_em = cppc_cpufreq_register_em;
+
+	return 0;
+}
+
+static void cppc_cpufreq_register_em(struct cpufreq_policy *policy)
+{
+	struct cppc_cpudata *cpu_data;
+	struct em_data_callback em_cb =
+		EM_ADV_DATA_CB(cppc_get_cpu_power, cppc_get_cpu_cost);
+
+	cpu_data = policy->driver_data;
+	em_dev_register_perf_domain(get_cpu_device(policy->cpu),
+			get_perf_level_count(policy), &em_cb,
+			cpu_data->shared_cpu_map, 0);
+}
+
 #else
 
 static unsigned int cppc_cpufreq_get_transition_delay_us(unsigned int cpu)
 {
 	return cppc_get_transition_latency(cpu) / NSEC_PER_USEC;
 }
+static int populate_efficiency_class(void)
+{
+	return 0;
+}
+static void cppc_cpufreq_register_em(struct cpufreq_policy *policy)
+{
+}
 #endif
 
 
@@ -536,6 +742,9 @@ static int cppc_cpufreq_cpu_init(struct cpufreq_policy *policy)
 		goto out;
 	}
 
+	policy->fast_switch_possible = cppc_allow_fast_switch();
+	policy->dvfs_possible_from_any_cpu = true;
+
 	/*
 	 * If 'highest_perf' is greater than 'nominal_perf', we assume CPU Boost
 	 * is supported.
@@ -681,6 +890,7 @@ static struct cpufreq_driver cppc_cpufreq_driver = {
 	.verify = cppc_verify_policy,
 	.target = cppc_cpufreq_set_target,
 	.get = cppc_cpufreq_get_rate,
+	.fast_switch = cppc_cpufreq_fast_switch,
 	.init = cppc_cpufreq_cpu_init,
 	.exit = cppc_cpufreq_cpu_exit,
 	.set_boost = cppc_cpufreq_set_boost,
@@ -742,6 +952,7 @@ static int __init cppc_cpufreq_init(void)
 
 	cppc_check_hisi_workaround();
 	cppc_freq_invariance_init();
+	populate_efficiency_class();
 
 	ret = cpufreq_register_driver(&cppc_cpufreq_driver);
 	if (ret)
diff --git a/drivers/cpufreq/cpufreq.c b/drivers/cpufreq/cpufreq.c
index 80f535cc8a75..2cad42774164 100644
--- a/drivers/cpufreq/cpufreq.c
+++ b/drivers/cpufreq/cpufreq.c
@@ -28,6 +28,7 @@
 #include <linux/suspend.h>
 #include <linux/syscore_ops.h>
 #include <linux/tick.h>
+#include <linux/units.h>
 #include <trace/events/power.h>
 
 static LIST_HEAD(cpufreq_policy_list);
@@ -947,13 +948,14 @@ static ssize_t show(struct kobject *kobj, struct attribute *attr, char *buf)
 {
 	struct cpufreq_policy *policy = to_policy(kobj);
 	struct freq_attr *fattr = to_attr(attr);
-	ssize_t ret;
+	ssize_t ret = -EBUSY;
 
 	if (!fattr->show)
 		return -EIO;
 
 	down_read(&policy->rwsem);
-	ret = fattr->show(policy, buf);
+	if (likely(!policy_is_inactive(policy)))
+		ret = fattr->show(policy, buf);
 	up_read(&policy->rwsem);
 
 	return ret;
@@ -964,7 +966,7 @@ static ssize_t store(struct kobject *kobj, struct attribute *attr,
 {
 	struct cpufreq_policy *policy = to_policy(kobj);
 	struct freq_attr *fattr = to_attr(attr);
-	ssize_t ret = -EINVAL;
+	ssize_t ret = -EBUSY;
 
 	if (!fattr->store)
 		return -EIO;
@@ -978,7 +980,8 @@ static ssize_t store(struct kobject *kobj, struct attribute *attr,
 
 	if (cpu_online(policy->cpu)) {
 		down_write(&policy->rwsem);
-		ret = fattr->store(policy, buf, count);
+		if (likely(!policy_is_inactive(policy)))
+			ret = fattr->store(policy, buf, count);
 		up_write(&policy->rwsem);
 	}
 
@@ -1019,11 +1022,12 @@ static void add_cpu_dev_symlink(struct cpufreq_policy *policy, unsigned int cpu,
 		dev_err(dev, "cpufreq symlink creation failed\n");
 }
 
-static void remove_cpu_dev_symlink(struct cpufreq_policy *policy,
+static void remove_cpu_dev_symlink(struct cpufreq_policy *policy, int cpu,
 				   struct device *dev)
 {
 	dev_dbg(dev, "%s: Removing symlink\n", __func__);
 	sysfs_remove_link(&dev->kobj, "cpufreq");
+	cpumask_clear_cpu(cpu, policy->real_cpus);
 }
 
 static int cpufreq_add_dev_interface(struct cpufreq_policy *policy)
@@ -1337,12 +1341,12 @@ static int cpufreq_online(unsigned int cpu)
 		down_write(&policy->rwsem);
 		policy->cpu = cpu;
 		policy->governor = NULL;
-		up_write(&policy->rwsem);
 	} else {
 		new_policy = true;
 		policy = cpufreq_policy_alloc(cpu);
 		if (!policy)
 			return -ENOMEM;
+		down_write(&policy->rwsem);
 	}
 
 	if (!new_policy && cpufreq_driver->online) {
@@ -1382,7 +1386,6 @@ static int cpufreq_online(unsigned int cpu)
 		cpumask_copy(policy->related_cpus, policy->cpus);
 	}
 
-	down_write(&policy->rwsem);
 	/*
 	 * affected cpus must always be the one, which are online. We aren't
 	 * managing offline cpus here.
@@ -1531,9 +1534,9 @@ static int cpufreq_online(unsigned int cpu)
 
 out_destroy_policy:
 	for_each_cpu(j, policy->real_cpus)
-		remove_cpu_dev_symlink(policy, get_cpu_device(j));
+		remove_cpu_dev_symlink(policy, j, get_cpu_device(j));
 
-	up_write(&policy->rwsem);
+	cpumask_clear(policy->cpus);
 
 out_offline_policy:
 	if (cpufreq_driver->offline)
@@ -1544,6 +1547,8 @@ out_exit_policy:
 		cpufreq_driver->exit(policy);
 
 out_free_policy:
+	up_write(&policy->rwsem);
+
 	cpufreq_policy_free(policy);
 	return ret;
 }
@@ -1575,47 +1580,36 @@ static int cpufreq_add_dev(struct device *dev, struct subsys_interface *sif)
 	return 0;
 }
 
-static int cpufreq_offline(unsigned int cpu)
+static void __cpufreq_offline(unsigned int cpu, struct cpufreq_policy *policy)
 {
-	struct cpufreq_policy *policy;
 	int ret;
 
-	pr_debug("%s: unregistering CPU %u\n", __func__, cpu);
-
-	policy = cpufreq_cpu_get_raw(cpu);
-	if (!policy) {
-		pr_debug("%s: No cpu_data found\n", __func__);
-		return 0;
-	}
-
-	down_write(&policy->rwsem);
 	if (has_target())
 		cpufreq_stop_governor(policy);
 
 	cpumask_clear_cpu(cpu, policy->cpus);
 
-	if (policy_is_inactive(policy)) {
-		if (has_target())
-			strncpy(policy->last_governor, policy->governor->name,
-				CPUFREQ_NAME_LEN);
-		else
-			policy->last_policy = policy->policy;
-	} else if (cpu == policy->cpu) {
-		/* Nominate new CPU */
-		policy->cpu = cpumask_any(policy->cpus);
-	}
-
-	/* Start governor again for active policy */
 	if (!policy_is_inactive(policy)) {
+		/* Nominate a new CPU if necessary. */
+		if (cpu == policy->cpu)
+			policy->cpu = cpumask_any(policy->cpus);
+
+		/* Start the governor again for the active policy. */
 		if (has_target()) {
 			ret = cpufreq_start_governor(policy);
 			if (ret)
 				pr_err("%s: Failed to start governor\n", __func__);
 		}
 
-		goto unlock;
+		return;
 	}
 
+	if (has_target())
+		strncpy(policy->last_governor, policy->governor->name,
+			CPUFREQ_NAME_LEN);
+	else
+		policy->last_policy = policy->policy;
+
 	if (cpufreq_thermal_control_enabled(cpufreq_driver)) {
 		cpufreq_cooling_unregister(policy->cdev);
 		policy->cdev = NULL;
@@ -1634,8 +1628,24 @@ static int cpufreq_offline(unsigned int cpu)
 		cpufreq_driver->exit(policy);
 		policy->freq_table = NULL;
 	}
+}
+
+static int cpufreq_offline(unsigned int cpu)
+{
+	struct cpufreq_policy *policy;
+
+	pr_debug("%s: unregistering CPU %u\n", __func__, cpu);
+
+	policy = cpufreq_cpu_get_raw(cpu);
+	if (!policy) {
+		pr_debug("%s: No cpu_data found\n", __func__);
+		return 0;
+	}
+
+	down_write(&policy->rwsem);
+
+	__cpufreq_offline(cpu, policy);
 
-unlock:
 	up_write(&policy->rwsem);
 	return 0;
 }
@@ -1653,19 +1663,25 @@ static void cpufreq_remove_dev(struct device *dev, struct subsys_interface *sif)
 	if (!policy)
 		return;
 
-	if (cpu_online(cpu))
-		cpufreq_offline(cpu);
+	down_write(&policy->rwsem);
 
-	cpumask_clear_cpu(cpu, policy->real_cpus);
-	remove_cpu_dev_symlink(policy, dev);
+	if (cpu_online(cpu))
+		__cpufreq_offline(cpu, policy);
 
-	if (cpumask_empty(policy->real_cpus)) {
-		/* We did light-weight exit earlier, do full tear down now */
-		if (cpufreq_driver->offline)
-			cpufreq_driver->exit(policy);
+	remove_cpu_dev_symlink(policy, cpu, dev);
 
-		cpufreq_policy_free(policy);
+	if (!cpumask_empty(policy->real_cpus)) {
+		up_write(&policy->rwsem);
+		return;
 	}
+
+	/* We did light-weight exit earlier, do full tear down now */
+	if (cpufreq_driver->offline)
+		cpufreq_driver->exit(policy);
+
+	up_write(&policy->rwsem);
+
+	cpufreq_policy_free(policy);
 }
 
 /**
@@ -1707,6 +1723,16 @@ static unsigned int cpufreq_verify_current_freq(struct cpufreq_policy *policy, b
 		return new_freq;
 
 	if (policy->cur != new_freq) {
+		/*
+		 * For some platforms, the frequency returned by hardware may be
+		 * slightly different from what is provided in the frequency
+		 * table, for example hardware may return 499 MHz instead of 500
+		 * MHz. In such cases it is better to avoid getting into
+		 * unnecessary frequency updates.
+		 */
+		if (abs(policy->cur - new_freq) < HZ_PER_MHZ)
+			return policy->cur;
+
 		cpufreq_out_of_sync(policy, new_freq);
 		if (update)
 			schedule_work(&policy->update);
diff --git a/drivers/cpufreq/cpufreq_governor.c b/drivers/cpufreq/cpufreq_governor.c
index 0d42cf8b88d8..85da677c43d6 100644
--- a/drivers/cpufreq/cpufreq_governor.c
+++ b/drivers/cpufreq/cpufreq_governor.c
@@ -388,6 +388,15 @@ static void free_policy_dbs_info(struct policy_dbs_info *policy_dbs,
 	gov->free(policy_dbs);
 }
 
+static void cpufreq_dbs_data_release(struct kobject *kobj)
+{
+	struct dbs_data *dbs_data = to_dbs_data(to_gov_attr_set(kobj));
+	struct dbs_governor *gov = dbs_data->gov;
+
+	gov->exit(dbs_data);
+	kfree(dbs_data);
+}
+
 int cpufreq_dbs_governor_init(struct cpufreq_policy *policy)
 {
 	struct dbs_governor *gov = dbs_governor_of(policy);
@@ -425,6 +434,7 @@ int cpufreq_dbs_governor_init(struct cpufreq_policy *policy)
 		goto free_policy_dbs_info;
 	}
 
+	dbs_data->gov = gov;
 	gov_attr_set_init(&dbs_data->attr_set, &policy_dbs->list);
 
 	ret = gov->init(dbs_data);
@@ -447,6 +457,7 @@ int cpufreq_dbs_governor_init(struct cpufreq_policy *policy)
 	policy->governor_data = policy_dbs;
 
 	gov->kobj_type.sysfs_ops = &governor_sysfs_ops;
+	gov->kobj_type.release = cpufreq_dbs_data_release;
 	ret = kobject_init_and_add(&dbs_data->attr_set.kobj, &gov->kobj_type,
 				   get_governor_parent_kobj(policy),
 				   "%s", gov->gov.name);
@@ -488,13 +499,8 @@ void cpufreq_dbs_governor_exit(struct cpufreq_policy *policy)
 
 	policy->governor_data = NULL;
 
-	if (!count) {
-		if (!have_governor_per_policy())
-			gov->gdbs_data = NULL;
-
-		gov->exit(dbs_data);
-		kfree(dbs_data);
-	}
+	if (!count && !have_governor_per_policy())
+		gov->gdbs_data = NULL;
 
 	free_policy_dbs_info(policy_dbs, gov);
 
diff --git a/drivers/cpufreq/cpufreq_governor.h b/drivers/cpufreq/cpufreq_governor.h
index a5a0bc3cc23e..168c23fd7fca 100644
--- a/drivers/cpufreq/cpufreq_governor.h
+++ b/drivers/cpufreq/cpufreq_governor.h
@@ -37,6 +37,7 @@ enum {OD_NORMAL_SAMPLE, OD_SUB_SAMPLE};
 /* Governor demand based switching data (per-policy or global). */
 struct dbs_data {
 	struct gov_attr_set attr_set;
+	struct dbs_governor *gov;
 	void *tuners;
 	unsigned int ignore_nice_load;
 	unsigned int sampling_rate;
diff --git a/drivers/cpufreq/intel_pstate.c b/drivers/cpufreq/intel_pstate.c
index 846bb3a78788..57cdb3679885 100644
--- a/drivers/cpufreq/intel_pstate.c
+++ b/drivers/cpufreq/intel_pstate.c
@@ -1322,6 +1322,7 @@ static ssize_t store_no_turbo(struct kobject *a, struct kobj_attribute *b,
 	mutex_unlock(&intel_pstate_limits_lock);
 
 	intel_pstate_update_policies();
+	arch_set_max_freq_ratio(global.no_turbo);
 
 	mutex_unlock(&intel_pstate_driver_lock);
 
@@ -2424,6 +2425,7 @@ static const struct x86_cpu_id intel_pstate_cpu_oob_ids[] __initconst = {
 	X86_MATCH(BROADWELL_X,		core_funcs),
 	X86_MATCH(SKYLAKE_X,		core_funcs),
 	X86_MATCH(ICELAKE_X,		core_funcs),
+	X86_MATCH(SAPPHIRERAPIDS_X,	core_funcs),
 	{}
 };
 
diff --git a/drivers/cpufreq/pasemi-cpufreq.c b/drivers/cpufreq/pasemi-cpufreq.c
index 815645170c4d..039a66bbe1be 100644
--- a/drivers/cpufreq/pasemi-cpufreq.c
+++ b/drivers/cpufreq/pasemi-cpufreq.c
@@ -18,7 +18,6 @@
 
 #include <asm/hw_irq.h>
 #include <asm/io.h>
-#include <asm/prom.h>
 #include <asm/time.h>
 #include <asm/smp.h>
 
diff --git a/drivers/cpufreq/pmac32-cpufreq.c b/drivers/cpufreq/pmac32-cpufreq.c
index 4f20c6a9108d..20f64a8b0a35 100644
--- a/drivers/cpufreq/pmac32-cpufreq.c
+++ b/drivers/cpufreq/pmac32-cpufreq.c
@@ -24,7 +24,7 @@
 #include <linux/device.h>
 #include <linux/hardirq.h>
 #include <linux/of_device.h>
-#include <asm/prom.h>
+
 #include <asm/machdep.h>
 #include <asm/irq.h>
 #include <asm/pmac_feature.h>
diff --git a/drivers/cpufreq/pmac64-cpufreq.c b/drivers/cpufreq/pmac64-cpufreq.c
index d7542a106e6b..ba9c31d98bd6 100644
--- a/drivers/cpufreq/pmac64-cpufreq.c
+++ b/drivers/cpufreq/pmac64-cpufreq.c
@@ -22,7 +22,7 @@
 #include <linux/completion.h>
 #include <linux/mutex.h>
 #include <linux/of_device.h>
-#include <asm/prom.h>
+
 #include <asm/machdep.h>
 #include <asm/irq.h>
 #include <asm/sections.h>
diff --git a/drivers/cpufreq/ppc_cbe_cpufreq.c b/drivers/cpufreq/ppc_cbe_cpufreq.c
index c58abb4cca3a..e3313ce63b38 100644
--- a/drivers/cpufreq/ppc_cbe_cpufreq.c
+++ b/drivers/cpufreq/ppc_cbe_cpufreq.c
@@ -12,7 +12,6 @@
 #include <linux/of_platform.h>
 
 #include <asm/machdep.h>
-#include <asm/prom.h>
 #include <asm/cell-regs.h>
 
 #include "ppc_cbe_cpufreq.h"
diff --git a/drivers/cpufreq/ppc_cbe_cpufreq_pmi.c b/drivers/cpufreq/ppc_cbe_cpufreq_pmi.c
index 037fe23bc6ed..4fba3637b115 100644
--- a/drivers/cpufreq/ppc_cbe_cpufreq_pmi.c
+++ b/drivers/cpufreq/ppc_cbe_cpufreq_pmi.c
@@ -13,9 +13,9 @@
 #include <linux/init.h>
 #include <linux/of_platform.h>
 #include <linux/pm_qos.h>
+#include <linux/slab.h>
 
 #include <asm/processor.h>
-#include <asm/prom.h>
 #include <asm/pmi.h>
 #include <asm/cell-regs.h>