1 files changed, 504 insertions, 309 deletions

diff --git a/arch/i386/kernel/cpu/cpufreq/acpi-cpufreq.c b/arch/i386/kernel/cpu/cpufreq/acpi-cpufreq.c
index 57c880bf0bd..18f4715c655 100644
--- a/arch/i386/kernel/cpu/cpufreq/acpi-cpufreq.c
+++ b/arch/i386/kernel/cpu/cpufreq/acpi-cpufreq.c
@@ -1,9 +1,10 @@
 /*
- * acpi-cpufreq.c - ACPI Processor P-States Driver ($Revision: 1.3 $)
+ * acpi-cpufreq.c - ACPI Processor P-States Driver ($Revision: 1.4 $)
  *
  *  Copyright (C) 2001, 2002 Andy Grover <andrew.grover@intel.com>
  *  Copyright (C) 2001, 2002 Paul Diefenbaugh <paul.s.diefenbaugh@intel.com>
  *  Copyright (C) 2002 - 2004 Dominik Brodowski <linux@brodo.de>
+ *  Copyright (C) 2006       Denis Sadykov <denis.m.sadykov@intel.com>
  *
  * ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
  *
@@ -27,202 +28,387 @@
 #include <linux/kernel.h>
 #include <linux/module.h>
 #include <linux/init.h>
+#include <linux/smp.h>
+#include <linux/sched.h>
 #include <linux/cpufreq.h>
-#include <linux/proc_fs.h>
-#include <linux/seq_file.h>
 #include <linux/compiler.h>
-#include <linux/sched.h>	/* current */
 #include <linux/dmi.h>
-#include <asm/io.h>
-#include <asm/delay.h>
-#include <asm/uaccess.h>
 
 #include <linux/acpi.h>
 #include <acpi/processor.h>
 
+#include <asm/io.h>
+#include <asm/msr.h>
+#include <asm/processor.h>
+#include <asm/cpufeature.h>
+#include <asm/delay.h>
+#include <asm/uaccess.h>
+
 #define dprintk(msg...) cpufreq_debug_printk(CPUFREQ_DEBUG_DRIVER, "acpi-cpufreq", msg)
 
 MODULE_AUTHOR("Paul Diefenbaugh, Dominik Brodowski");
 MODULE_DESCRIPTION("ACPI Processor P-States Driver");
 MODULE_LICENSE("GPL");
 
+enum {
+	UNDEFINED_CAPABLE = 0,
+	SYSTEM_INTEL_MSR_CAPABLE,
+	SYSTEM_IO_CAPABLE,
+};
+
+#define INTEL_MSR_RANGE		(0xffff)
+#define CPUID_6_ECX_APERFMPERF_CAPABILITY	(0x1)
 
-struct cpufreq_acpi_io {
-	struct acpi_processor_performance	*acpi_data;
-	struct cpufreq_frequency_table		*freq_table;
-	unsigned int				resume;
+struct acpi_cpufreq_data {
+	struct acpi_processor_performance *acpi_data;
+	struct cpufreq_frequency_table *freq_table;
+	unsigned int max_freq;
+	unsigned int resume;
+	unsigned int cpu_feature;
 };
 
-static struct cpufreq_acpi_io	*acpi_io_data[NR_CPUS];
-static struct acpi_processor_performance	*acpi_perf_data[NR_CPUS];
+static struct acpi_cpufreq_data *drv_data[NR_CPUS];
+static struct acpi_processor_performance *acpi_perf_data[NR_CPUS];
 
 static struct cpufreq_driver acpi_cpufreq_driver;
 
 static unsigned int acpi_pstate_strict;
 
-static int
-acpi_processor_write_port(
-	u16	port,
-	u8	bit_width,
-	u32	value)
+static int check_est_cpu(unsigned int cpuid)
+{
+	struct cpuinfo_x86 *cpu = &cpu_data[cpuid];
+
+	if (cpu->x86_vendor != X86_VENDOR_INTEL ||
+	    !cpu_has(cpu, X86_FEATURE_EST))
+		return 0;
+
+	return 1;
+}
+
+static unsigned extract_io(u32 value, struct acpi_cpufreq_data *data)
+{
+	struct acpi_processor_performance *perf;
+	int i;
+
+	perf = data->acpi_data;
+
+	for (i=0; i<perf->state_count; i++) {
+		if (value == perf->states[i].status)
+			return data->freq_table[i].frequency;
+	}
+	return 0;
+}
+
+static unsigned extract_msr(u32 msr, struct acpi_cpufreq_data *data)
 {
-	if (bit_width <= 8) {
+	int i;
+	struct acpi_processor_performance *perf;
+
+	msr &= INTEL_MSR_RANGE;
+	perf = data->acpi_data;
+
+	for (i=0; data->freq_table[i].frequency != CPUFREQ_TABLE_END; i++) {
+		if (msr == perf->states[data->freq_table[i].index].status)
+			return data->freq_table[i].frequency;
+	}
+	return data->freq_table[0].frequency;
+}
+
+static unsigned extract_freq(u32 val, struct acpi_cpufreq_data *data)
+{
+	switch (data->cpu_feature) {
+	case SYSTEM_INTEL_MSR_CAPABLE:
+		return extract_msr(val, data);
+	case SYSTEM_IO_CAPABLE:
+		return extract_io(val, data);
+	default:
+		return 0;
+	}
+}
+
+static void wrport(u16 port, u8 bit_width, u32 value)
+{
+	if (bit_width <= 8)
 		outb(value, port);
-	} else if (bit_width <= 16) {
+	else if (bit_width <= 16)
 		outw(value, port);
-	} else if (bit_width <= 32) {
+	else if (bit_width <= 32)
 		outl(value, port);
-	} else {
-		return -ENODEV;
-	}
-	return 0;
 }
 
-static int
-acpi_processor_read_port(
-	u16	port,
-	u8	bit_width,
-	u32	*ret)
+static void rdport(u16 port, u8 bit_width, u32 * ret)
 {
 	*ret = 0;
-	if (bit_width <= 8) {
+	if (bit_width <= 8)
 		*ret = inb(port);
-	} else if (bit_width <= 16) {
+	else if (bit_width <= 16)
 		*ret = inw(port);
-	} else if (bit_width <= 32) {
+	else if (bit_width <= 32)
 		*ret = inl(port);
-	} else {
-		return -ENODEV;
+}
+
+struct msr_addr {
+	u32 reg;
+};
+
+struct io_addr {
+	u16 port;
+	u8 bit_width;
+};
+
+typedef union {
+	struct msr_addr msr;
+	struct io_addr io;
+} drv_addr_union;
+
+struct drv_cmd {
+	unsigned int type;
+	cpumask_t mask;
+	drv_addr_union addr;
+	u32 val;
+};
+
+static void do_drv_read(struct drv_cmd *cmd)
+{
+	u32 h;
+
+	switch (cmd->type) {
+	case SYSTEM_INTEL_MSR_CAPABLE:
+		rdmsr(cmd->addr.msr.reg, cmd->val, h);
+		break;
+	case SYSTEM_IO_CAPABLE:
+		rdport(cmd->addr.io.port, cmd->addr.io.bit_width, &cmd->val);
+		break;
+	default:
+		break;
 	}
-	return 0;
 }
 
-static int
-acpi_processor_set_performance (
-	struct cpufreq_acpi_io	*data,
-	unsigned int		cpu,
-	int			state)
+static void do_drv_write(struct drv_cmd *cmd)
 {
-	u16			port = 0;
-	u8			bit_width = 0;
-	int			i = 0;
-	int			ret = 0;
-	u32			value = 0;
-	int			retval;
-	struct acpi_processor_performance	*perf;
-
-	dprintk("acpi_processor_set_performance\n");
-
-	retval = 0;
-	perf = data->acpi_data;	
-	if (state == perf->state) {
-		if (unlikely(data->resume)) {
-			dprintk("Called after resume, resetting to P%d\n", state);
-			data->resume = 0;
-		} else {
-			dprintk("Already at target state (P%d)\n", state);
-			return (retval);
-		}
+	u32 h = 0;
+
+	switch (cmd->type) {
+	case SYSTEM_INTEL_MSR_CAPABLE:
+		wrmsr(cmd->addr.msr.reg, cmd->val, h);
+		break;
+	case SYSTEM_IO_CAPABLE:
+		wrport(cmd->addr.io.port, cmd->addr.io.bit_width, cmd->val);
+		break;
+	default:
+		break;
 	}
+}
 
-	dprintk("Transitioning from P%d to P%d\n", perf->state, state);
+static void drv_read(struct drv_cmd *cmd)
+{
+	cpumask_t saved_mask = current->cpus_allowed;
+	cmd->val = 0;
 
-	/*
-	 * First we write the target state's 'control' value to the
-	 * control_register.
-	 */
+	set_cpus_allowed(current, cmd->mask);
+	do_drv_read(cmd);
+	set_cpus_allowed(current, saved_mask);
+}
+
+static void drv_write(struct drv_cmd *cmd)
+{
+	cpumask_t saved_mask = current->cpus_allowed;
+	unsigned int i;
+
+	for_each_cpu_mask(i, cmd->mask) {
+		set_cpus_allowed(current, cpumask_of_cpu(i));
+		do_drv_write(cmd);
+	}
+
+	set_cpus_allowed(current, saved_mask);
+	return;
+}
+
+static u32 get_cur_val(cpumask_t mask)
+{
+	struct acpi_processor_performance *perf;
+	struct drv_cmd cmd;
+
+	if (unlikely(cpus_empty(mask)))
+		return 0;
+
+	switch (drv_data[first_cpu(mask)]->cpu_feature) {
+	case SYSTEM_INTEL_MSR_CAPABLE:
+		cmd.type = SYSTEM_INTEL_MSR_CAPABLE;
+		cmd.addr.msr.reg = MSR_IA32_PERF_STATUS;
+		break;
+	case SYSTEM_IO_CAPABLE:
+		cmd.type = SYSTEM_IO_CAPABLE;
+		perf = drv_data[first_cpu(mask)]->acpi_data;
+		cmd.addr.io.port = perf->control_register.address;
+		cmd.addr.io.bit_width = perf->control_register.bit_width;
+		break;
+	default:
+		return 0;
+	}
+
+	cmd.mask = mask;
 
-	port = perf->control_register.address;
-	bit_width = perf->control_register.bit_width;
-	value = (u32) perf->states[state].control;
+	drv_read(&cmd);
 
-	dprintk("Writing 0x%08x to port 0x%04x\n", value, port);
+	dprintk("get_cur_val = %u\n", cmd.val);
 
-	ret = acpi_processor_write_port(port, bit_width, value);
-	if (ret) {
-		dprintk("Invalid port width 0x%04x\n", bit_width);
-		return (ret);
+	return cmd.val;
+}
+
+/*
+ * Return the measured active (C0) frequency on this CPU since last call
+ * to this function.
+ * Input: cpu number
+ * Return: Average CPU frequency in terms of max frequency (zero on error)
+ *
+ * We use IA32_MPERF and IA32_APERF MSRs to get the measured performance
+ * over a period of time, while CPU is in C0 state.
+ * IA32_MPERF counts at the rate of max advertised frequency
+ * IA32_APERF counts at the rate of actual CPU frequency
+ * Only IA32_APERF/IA32_MPERF ratio is architecturally defined and
+ * no meaning should be associated with absolute values of these MSRs.
+ */
+static unsigned int get_measured_perf(unsigned int cpu)
+{
+	union {
+		struct {
+			u32 lo;
+			u32 hi;
+		} split;
+		u64 whole;
+	} aperf_cur, mperf_cur;
+
+	cpumask_t saved_mask;
+	unsigned int perf_percent;
+	unsigned int retval;
+
+	saved_mask = current->cpus_allowed;
+	set_cpus_allowed(current, cpumask_of_cpu(cpu));
+	if (get_cpu() != cpu) {
+		/* We were not able to run on requested processor */
+		put_cpu();
+		return 0;
 	}
 
+	rdmsr(MSR_IA32_APERF, aperf_cur.split.lo, aperf_cur.split.hi);
+	rdmsr(MSR_IA32_MPERF, mperf_cur.split.lo, mperf_cur.split.hi);
+
+	wrmsr(MSR_IA32_APERF, 0,0);
+	wrmsr(MSR_IA32_MPERF, 0,0);
+
+#ifdef __i386__
 	/*
-	 * Assume the write went through when acpi_pstate_strict is not used.
-	 * As read status_register is an expensive operation and there 
-	 * are no specific error cases where an IO port write will fail.
+	 * We dont want to do 64 bit divide with 32 bit kernel
+	 * Get an approximate value. Return failure in case we cannot get
+	 * an approximate value.
 	 */
-	if (acpi_pstate_strict) {
-		/* Then we read the 'status_register' and compare the value 
-		 * with the target state's 'status' to make sure the 
-		 * transition was successful.
-		 * Note that we'll poll for up to 1ms (100 cycles of 10us) 
-		 * before giving up.
-		 */
-
-		port = perf->status_register.address;
-		bit_width = perf->status_register.bit_width;
-
-		dprintk("Looking for 0x%08x from port 0x%04x\n",
-			(u32) perf->states[state].status, port);
-
-		for (i = 0; i < 100; i++) {
-			ret = acpi_processor_read_port(port, bit_width, &value);
-			if (ret) {	
-				dprintk("Invalid port width 0x%04x\n", bit_width);
-				return (ret);
-			}
-			if (value == (u32) perf->states[state].status)
-				break;
-			udelay(10);
-		}
-	} else {
-		value = (u32) perf->states[state].status;
+	if (unlikely(aperf_cur.split.hi || mperf_cur.split.hi)) {
+		int shift_count;
+		u32 h;
+
+		h = max_t(u32, aperf_cur.split.hi, mperf_cur.split.hi);
+		shift_count = fls(h);
+
+		aperf_cur.whole >>= shift_count;
+		mperf_cur.whole >>= shift_count;
+	}
+
+	if (((unsigned long)(-1) / 100) < aperf_cur.split.lo) {
+		int shift_count = 7;
+		aperf_cur.split.lo >>= shift_count;
+		mperf_cur.split.lo >>= shift_count;
+	}
+
+	if (aperf_cur.split.lo && mperf_cur.split.lo)
+		perf_percent = (aperf_cur.split.lo * 100) / mperf_cur.split.lo;
+	else
+		perf_percent = 0;
+
+#else
+	if (unlikely(((unsigned long)(-1) / 100) < aperf_cur.whole)) {
+		int shift_count = 7;
+		aperf_cur.whole >>= shift_count;
+		mperf_cur.whole >>= shift_count;
 	}
 
-	if (unlikely(value != (u32) perf->states[state].status)) {
-		printk(KERN_WARNING "acpi-cpufreq: Transition failed\n");
-		retval = -ENODEV;
-		return (retval);
+	if (aperf_cur.whole && mperf_cur.whole)
+		perf_percent = (aperf_cur.whole * 100) / mperf_cur.whole;
+	else
+		perf_percent = 0;
+
+#endif
+
+	retval = drv_data[cpu]->max_freq * perf_percent / 100;
+
+	put_cpu();
+	set_cpus_allowed(current, saved_mask);
+
+	dprintk("cpu %d: performance percent %d\n", cpu, perf_percent);
+	return retval;
+}
+
+static unsigned int get_cur_freq_on_cpu(unsigned int cpu)
+{
+	struct acpi_cpufreq_data *data = drv_data[cpu];
+	unsigned int freq;
+
+	dprintk("get_cur_freq_on_cpu (%d)\n", cpu);
+
+	if (unlikely(data == NULL ||
+		     data->acpi_data == NULL || data->freq_table == NULL)) {
+		return 0;
 	}
 
-	dprintk("Transition successful after %d microseconds\n", i * 10);
+	freq = extract_freq(get_cur_val(cpumask_of_cpu(cpu)), data);
+	dprintk("cur freq = %u\n", freq);
 
-	perf->state = state;
-	return (retval);
+	return freq;
 }
 
+static unsigned int check_freqs(cpumask_t mask, unsigned int freq,
+				struct acpi_cpufreq_data *data)
+{
+	unsigned int cur_freq;
+	unsigned int i;
+
+	for (i=0; i<100; i++) {
+		cur_freq = extract_freq(get_cur_val(mask), data);
+		if (cur_freq == freq)
+			return 1;
+		udelay(10);
+	}
+	return 0;
+}
 
-static int
-acpi_cpufreq_target (
-	struct cpufreq_policy   *policy,
-	unsigned int target_freq,
-	unsigned int relation)
+static int acpi_cpufreq_target(struct cpufreq_policy *policy,
+			       unsigned int target_freq, unsigned int relation)
 {
-	struct cpufreq_acpi_io *data = acpi_io_data[policy->cpu];
+	struct acpi_cpufreq_data *data = drv_data[policy->cpu];
 	struct acpi_processor_performance *perf;
 	struct cpufreq_freqs freqs;
 	cpumask_t online_policy_cpus;
-	cpumask_t saved_mask;
-	cpumask_t set_mask;
-	cpumask_t covered_cpus;
-	unsigned int cur_state = 0;
+	struct drv_cmd cmd;
+	unsigned int msr;
 	unsigned int next_state = 0;
-	unsigned int result = 0;
-	unsigned int j;
-	unsigned int tmp;
+	unsigned int next_perf_state = 0;
+	unsigned int i;
+	int result = 0;
 
-	dprintk("acpi_cpufreq_setpolicy\n");
+	dprintk("acpi_cpufreq_target %d (%d)\n", target_freq, policy->cpu);
 
-	result = cpufreq_frequency_table_target(policy,
-			data->freq_table,
-			target_freq,
-			relation,
-			&next_state);
-	if (unlikely(result))
-		return (result);
+	if (unlikely(data == NULL ||
+	     data->acpi_data == NULL || data->freq_table == NULL)) {
+		return -ENODEV;
+	}
 
 	perf = data->acpi_data;
-	cur_state = perf->state;
-	freqs.old = data->freq_table[cur_state].frequency;
-	freqs.new = data->freq_table[next_state].frequency;
+	result = cpufreq_frequency_table_target(policy,
+						data->freq_table,
+						target_freq,
+						relation, &next_state);
+	if (unlikely(result))
+		return -ENODEV;
 
 #ifdef CONFIG_HOTPLUG_CPU
 	/* cpufreq holds the hotplug lock, so we are safe from here on */
@@ -231,106 +417,84 @@ acpi_cpufreq_target (
 	online_policy_cpus = policy->cpus;
 #endif
 
-	for_each_cpu_mask(j, online_policy_cpus) {
-		freqs.cpu = j;
-		cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE);
+	next_perf_state = data->freq_table[next_state].index;
+	if (perf->state == next_perf_state) {
+		if (unlikely(data->resume)) {
+			dprintk("Called after resume, resetting to P%d\n",
+				next_perf_state);
+			data->resume = 0;
+		} else {
+			dprintk("Already at target state (P%d)\n",
+				next_perf_state);
+			return 0;
+		}
 	}
 
-	/*
-	 * We need to call driver->target() on all or any CPU in
-	 * policy->cpus, depending on policy->shared_type.
-	 */
-	saved_mask = current->cpus_allowed;
-	cpus_clear(covered_cpus);
-	for_each_cpu_mask(j, online_policy_cpus) {
-		/*
-		 * Support for SMP systems.
-		 * Make sure we are running on CPU that wants to change freq
-		 */
-		cpus_clear(set_mask);
-		if (policy->shared_type == CPUFREQ_SHARED_TYPE_ANY)
-			cpus_or(set_mask, set_mask, online_policy_cpus);
-		else
-			cpu_set(j, set_mask);
-
-		set_cpus_allowed(current, set_mask);
-		if (unlikely(!cpu_isset(smp_processor_id(), set_mask))) {
-			dprintk("couldn't limit to CPUs in this domain\n");
-			result = -EAGAIN;
-			break;
-		}
+	switch (data->cpu_feature) {
+	case SYSTEM_INTEL_MSR_CAPABLE:
+		cmd.type = SYSTEM_INTEL_MSR_CAPABLE;
+		cmd.addr.msr.reg = MSR_IA32_PERF_CTL;
+		msr =
+		    (u32) perf->states[next_perf_state].
+		    control & INTEL_MSR_RANGE;
+		cmd.val = (cmd.val & ~INTEL_MSR_RANGE) | msr;
+		break;
+	case SYSTEM_IO_CAPABLE:
+		cmd.type = SYSTEM_IO_CAPABLE;
+		cmd.addr.io.port = perf->control_register.address;
+		cmd.addr.io.bit_width = perf->control_register.bit_width;
+		cmd.val = (u32) perf->states[next_perf_state].control;
+		break;
+	default:
+		return -ENODEV;
+	}
 
-		result = acpi_processor_set_performance (data, j, next_state);
-		if (result) {
-			result = -EAGAIN;
-			break;
-		}
+	cpus_clear(cmd.mask);
 
-		if (policy->shared_type == CPUFREQ_SHARED_TYPE_ANY)
-			break;
- 
-		cpu_set(j, covered_cpus);
-	}
+	if (policy->shared_type != CPUFREQ_SHARED_TYPE_ANY)
+		cmd.mask = online_policy_cpus;
+	else
+		cpu_set(policy->cpu, cmd.mask);
 
-	for_each_cpu_mask(j, online_policy_cpus) {
-		freqs.cpu = j;
-		cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE);
+	freqs.old = data->freq_table[perf->state].frequency;
+	freqs.new = data->freq_table[next_perf_state].frequency;
+	for_each_cpu_mask(i, cmd.mask) {
+		freqs.cpu = i;
+		cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE);
 	}
 
-	if (unlikely(result)) {
-		/*
-		 * We have failed halfway through the frequency change.
-		 * We have sent callbacks to online_policy_cpus and
-		 * acpi_processor_set_performance() has been called on 
-		 * coverd_cpus. Best effort undo..
-		 */
-
-		if (!cpus_empty(covered_cpus)) {
-			for_each_cpu_mask(j, covered_cpus) {
-				policy->cpu = j;
-				acpi_processor_set_performance (data, 
-						j, 
-						cur_state);
-			}
-		}
+	drv_write(&cmd);
 
-		tmp = freqs.new;
-		freqs.new = freqs.old;
-		freqs.old = tmp;
-		for_each_cpu_mask(j, online_policy_cpus) {
-			freqs.cpu = j;
-			cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE);
-			cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE);
+	if (acpi_pstate_strict) {
+		if (!check_freqs(cmd.mask, freqs.new, data)) {
+			dprintk("acpi_cpufreq_target failed (%d)\n",
+				policy->cpu);
+			return -EAGAIN;
 		}
 	}
 
-	set_cpus_allowed(current, saved_mask);
-	return (result);
-}
+	for_each_cpu_mask(i, cmd.mask) {
+		freqs.cpu = i;
+		cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE);
+	}
+	perf->state = next_perf_state;
 
+	return result;
+}
 
-static int
-acpi_cpufreq_verify (
-	struct cpufreq_policy   *policy)
+static int acpi_cpufreq_verify(struct cpufreq_policy *policy)
 {
-	unsigned int result = 0;
-	struct cpufreq_acpi_io *data = acpi_io_data[policy->cpu];
+	struct acpi_cpufreq_data *data = drv_data[policy->cpu];
 
 	dprintk("acpi_cpufreq_verify\n");
 
-	result = cpufreq_frequency_table_verify(policy, 
-			data->freq_table);
-
-	return (result);
+	return cpufreq_frequency_table_verify(policy, data->freq_table);
 }
 
-
 static unsigned long
-acpi_cpufreq_guess_freq (
-	struct cpufreq_acpi_io	*data,
-	unsigned int		cpu)
+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 = data->acpi_data;
 
 	if (cpu_khz) {
 		/* search the closest match to cpu_khz */
@@ -338,16 +502,16 @@ acpi_cpufreq_guess_freq (
 		unsigned long freq;
 		unsigned long freqn = perf->states[0].core_frequency * 1000;
 
-		for (i = 0; i < (perf->state_count - 1); i++) {
+		for (i=0; i<(perf->state_count-1); i++) {
 			freq = freqn;
 			freqn = perf->states[i+1].core_frequency * 1000;
 			if ((2 * cpu_khz) > (freqn + freq)) {
 				perf->state = i;
-				return (freq);
+				return freq;
 			}
 		}
-		perf->state = perf->state_count - 1;
-		return (freqn);
+		perf->state = perf->state_count-1;
+		return freqn;
 	} else {
 		/* assume CPU is at P0... */
 		perf->state = 0;
@@ -355,7 +519,6 @@ acpi_cpufreq_guess_freq (
 	}
 }
 
-
 /*
  * acpi_cpufreq_early_init - initialize ACPI P-States library
  *
@@ -364,30 +527,34 @@ acpi_cpufreq_guess_freq (
  * do _PDC and _PSD and find out the processor dependency for the
  * actual init that will happen later...
  */
-static int acpi_cpufreq_early_init_acpi(void)
+static int acpi_cpufreq_early_init(void)
 {
-	struct acpi_processor_performance	*data;
-	unsigned int				i, j;
+	struct acpi_processor_performance *data;
+	cpumask_t covered;
+	unsigned int i, j;
 
 	dprintk("acpi_cpufreq_early_init\n");
 
 	for_each_possible_cpu(i) {
-		data = kzalloc(sizeof(struct acpi_processor_performance), 
-			GFP_KERNEL);
+		data = kzalloc(sizeof(struct acpi_processor_performance),
+			       GFP_KERNEL);
 		if (!data) {
-			for_each_possible_cpu(j) {
+			for_each_cpu_mask(j, covered) {
 				kfree(acpi_perf_data[j]);
 				acpi_perf_data[j] = NULL;
 			}
-			return (-ENOMEM);
+			return -ENOMEM;
 		}
 		acpi_perf_data[i] = data;
+		cpu_set(i, covered);
 	}
 
 	/* Do initialization in ACPI core */
-	return acpi_processor_preregister_performance(acpi_perf_data);
+	acpi_processor_preregister_performance(acpi_perf_data);
+	return 0;
 }
 
+#ifdef CONFIG_SMP
 /*
  * Some BIOSes do SW_ANY coordination internally, either set it up in hw
  * or do it in BIOS firmware and won't inform about it to OS. If not
@@ -414,39 +581,42 @@ static struct dmi_system_id sw_any_bug_dmi_table[] = {
 	},
 	{ }
 };
+#endif
 
-static int
-acpi_cpufreq_cpu_init (
-	struct cpufreq_policy   *policy)
+static int acpi_cpufreq_cpu_init(struct cpufreq_policy *policy)
 {
-	unsigned int		i;
-	unsigned int		cpu = policy->cpu;
-	struct cpufreq_acpi_io	*data;
-	unsigned int		result = 0;
+	unsigned int i;
+	unsigned int valid_states = 0;
+	unsigned int cpu = policy->cpu;
+	struct acpi_cpufreq_data *data;
+	unsigned int result = 0;
 	struct cpuinfo_x86 *c = &cpu_data[policy->cpu];
-	struct acpi_processor_performance	*perf;
+	struct acpi_processor_performance *perf;
 
 	dprintk("acpi_cpufreq_cpu_init\n");
 
 	if (!acpi_perf_data[cpu])
-		return (-ENODEV);
+		return -ENODEV;
 
-	data = kzalloc(sizeof(struct cpufreq_acpi_io), GFP_KERNEL);
+	data = kzalloc(sizeof(struct acpi_cpufreq_data), GFP_KERNEL);
 	if (!data)
-		return (-ENOMEM);
+		return -ENOMEM;
 
 	data->acpi_data = acpi_perf_data[cpu];
-	acpi_io_data[cpu] = data;
+	drv_data[cpu] = data;
 
-	result = acpi_processor_register_performance(data->acpi_data, cpu);
+	if (cpu_has(c, X86_FEATURE_CONSTANT_TSC))
+		acpi_cpufreq_driver.flags |= CPUFREQ_CONST_LOOPS;
 
+	result = acpi_processor_register_performance(data->acpi_data, cpu);
 	if (result)
 		goto err_free;
 
 	perf = data->acpi_data;
 	policy->shared_type = perf->shared_type;
+
 	/*
-	 * Will let policy->cpus know about dependency only when software 
+	 * Will let policy->cpus know about dependency only when software
 	 * coordination is required.
 	 */
 	if (policy->shared_type == CPUFREQ_SHARED_TYPE_ALL ||
@@ -462,10 +632,6 @@ acpi_cpufreq_cpu_init (
 	}
 #endif
 
-	if (cpu_has(c, X86_FEATURE_CONSTANT_TSC)) {
-		acpi_cpufreq_driver.flags |= CPUFREQ_CONST_LOOPS;
-	}
-
 	/* capability check */
 	if (perf->state_count <= 1) {
 		dprintk("No P-States\n");
@@ -473,17 +639,33 @@ acpi_cpufreq_cpu_init (
 		goto err_unreg;
 	}
 
-	if ((perf->control_register.space_id != ACPI_ADR_SPACE_SYSTEM_IO) ||
-	    (perf->status_register.space_id != ACPI_ADR_SPACE_SYSTEM_IO)) {
-		dprintk("Unsupported address space [%d, %d]\n",
-			(u32) (perf->control_register.space_id),
-			(u32) (perf->status_register.space_id));
+	if (perf->control_register.space_id != perf->status_register.space_id) {
+		result = -ENODEV;
+		goto err_unreg;
+	}
+
+	switch (perf->control_register.space_id) {
+	case ACPI_ADR_SPACE_SYSTEM_IO:
+		dprintk("SYSTEM IO addr space\n");
+		data->cpu_feature = SYSTEM_IO_CAPABLE;
+		break;
+	case ACPI_ADR_SPACE_FIXED_HARDWARE:
+		dprintk("HARDWARE addr space\n");
+		if (!check_est_cpu(cpu)) {
+			result = -ENODEV;
+			goto err_unreg;
+		}
+		data->cpu_feature = SYSTEM_INTEL_MSR_CAPABLE;
+		break;
+	default:
+		dprintk("Unknown addr space %d\n",
+			(u32) (perf->control_register.space_id));
 		result = -ENODEV;
 		goto err_unreg;
 	}
 
-	/* alloc freq_table */
-	data->freq_table = kmalloc(sizeof(struct cpufreq_frequency_table) * (perf->state_count + 1), GFP_KERNEL);
+	data->freq_table = kmalloc(sizeof(struct cpufreq_frequency_table) *
+		    (perf->state_count+1), GFP_KERNEL);
 	if (!data->freq_table) {
 		result = -ENOMEM;
 		goto err_unreg;
@@ -492,129 +674,140 @@ acpi_cpufreq_cpu_init (
 	/* detect transition latency */
 	policy->cpuinfo.transition_latency = 0;
 	for (i=0; i<perf->state_count; i++) {
-		if ((perf->states[i].transition_latency * 1000) > policy->cpuinfo.transition_latency)
-			policy->cpuinfo.transition_latency = perf->states[i].transition_latency * 1000;
+		if ((perf->states[i].transition_latency * 1000) >
+		    policy->cpuinfo.transition_latency)
+			policy->cpuinfo.transition_latency =
+			    perf->states[i].transition_latency * 1000;
 	}
 	policy->governor = CPUFREQ_DEFAULT_GOVERNOR;
 
-	/* The current speed is unknown and not detectable by ACPI...  */
-	policy->cur = acpi_cpufreq_guess_freq(data, policy->cpu);
-
+	data->max_freq = perf->states[0].core_frequency * 1000;
 	/* table init */
-	for (i=0; i<=perf->state_count; i++)
-	{
-		data->freq_table[i].index = i;
-		if (i<perf->state_count)
-			data->freq_table[i].frequency = perf->states[i].core_frequency * 1000;
-		else
-			data->freq_table[i].frequency = CPUFREQ_TABLE_END;
+	for (i=0; i<perf->state_count; i++) {
+		if (i>0 && perf->states[i].core_frequency ==
+		    perf->states[i-1].core_frequency)
+			continue;
+
+		data->freq_table[valid_states].index = i;
+		data->freq_table[valid_states].frequency =
+		    perf->states[i].core_frequency * 1000;
+		valid_states++;
 	}
+	data->freq_table[valid_states].frequency = CPUFREQ_TABLE_END;
 
 	result = cpufreq_frequency_table_cpuinfo(policy, data->freq_table);
-	if (result) {
+	if (result)
 		goto err_freqfree;
+
+	switch (data->cpu_feature) {
+	case ACPI_ADR_SPACE_SYSTEM_IO:
+		/* Current speed is unknown and not detectable by IO port */
+		policy->cur = acpi_cpufreq_guess_freq(data, policy->cpu);
+		break;
+	case ACPI_ADR_SPACE_FIXED_HARDWARE:
+		acpi_cpufreq_driver.get = get_cur_freq_on_cpu;
+		get_cur_freq_on_cpu(cpu);
+		break;
+	default:
+		break;
 	}
 
 	/* notify BIOS that we exist */
 	acpi_processor_notify_smm(THIS_MODULE);
 
-	printk(KERN_INFO "acpi-cpufreq: CPU%u - ACPI performance management activated.\n",
-	       cpu);
+	/* Check for APERF/MPERF support in hardware */
+	if (c->x86_vendor == X86_VENDOR_INTEL && c->cpuid_level >= 6) {
+		unsigned int ecx;
+		ecx = cpuid_ecx(6);
+		if (ecx & CPUID_6_ECX_APERFMPERF_CAPABILITY)
+			acpi_cpufreq_driver.getavg = get_measured_perf;
+	}
+
+	dprintk("CPU%u - ACPI performance management activated.\n", cpu);
 	for (i = 0; i < perf->state_count; i++)
 		dprintk("     %cP%d: %d MHz, %d mW, %d uS\n",
-			(i == perf->state?'*':' '), i,
+			(i == perf->state ? '*' : ' '), i,
 			(u32) perf->states[i].core_frequency,
 			(u32) perf->states[i].power,
 			(u32) perf->states[i].transition_latency);
 
 	cpufreq_frequency_table_get_attr(data->freq_table, policy->cpu);
-	
+
 	/*
 	 * the first call to ->target() should result in us actually
 	 * writing something to the appropriate registers.
 	 */
 	data->resume = 1;
-	
-	return (result);
 
- err_freqfree:
+	return result;
+
+err_freqfree:
 	kfree(data->freq_table);
- err_unreg:
+err_unreg:
 	acpi_processor_unregister_performance(perf, cpu);
- err_free:
+err_free:
 	kfree(data);
-	acpi_io_data[cpu] = NULL;
+	drv_data[cpu] = NULL;
 
-	return (result);
+	return result;
 }
 
-
-static int
-acpi_cpufreq_cpu_exit (
-	struct cpufreq_policy   *policy)
+static int acpi_cpufreq_cpu_exit(struct cpufreq_policy *policy)
 {
-	struct cpufreq_acpi_io *data = acpi_io_data[policy->cpu];
-
+	struct acpi_cpufreq_data *data = drv_data[policy->cpu];
 
 	dprintk("acpi_cpufreq_cpu_exit\n");
 
 	if (data) {
 		cpufreq_frequency_table_put_attr(policy->cpu);
-		acpi_io_data[policy->cpu] = NULL;
-		acpi_processor_unregister_performance(data->acpi_data, policy->cpu);
+		drv_data[policy->cpu] = NULL;
+		acpi_processor_unregister_performance(data->acpi_data,
+						      policy->cpu);
 		kfree(data);
 	}
 
-	return (0);
+	return 0;
 }
 
-static int
-acpi_cpufreq_resume (
-	struct cpufreq_policy   *policy)
+static int acpi_cpufreq_resume(struct cpufreq_policy *policy)
 {
-	struct cpufreq_acpi_io *data = acpi_io_data[policy->cpu];
-
+	struct acpi_cpufreq_data *data = drv_data[policy->cpu];
 
 	dprintk("acpi_cpufreq_resume\n");
 
 	data->resume = 1;
 
-	return (0);
+	return 0;
 }
 
-
-static struct freq_attr* acpi_cpufreq_attr[] = {
+static struct freq_attr *acpi_cpufreq_attr[] = {
 	&cpufreq_freq_attr_scaling_available_freqs,
 	NULL,
 };
 
 static struct cpufreq_driver acpi_cpufreq_driver = {
-	.verify	= acpi_cpufreq_verify,
-	.target	= acpi_cpufreq_target,
-	.init	= acpi_cpufreq_cpu_init,
-	.exit	= acpi_cpufreq_cpu_exit,
-	.resume	= acpi_cpufreq_resume,
-	.name	= "acpi-cpufreq",
-	.owner	= THIS_MODULE,
-	.attr	= acpi_cpufreq_attr,
+	.verify = acpi_cpufreq_verify,
+	.target = acpi_cpufreq_target,
+	.init = acpi_cpufreq_cpu_init,
+	.exit = acpi_cpufreq_cpu_exit,
+	.resume = acpi_cpufreq_resume,
+	.name = "acpi-cpufreq",
+	.owner = THIS_MODULE,
+	.attr = acpi_cpufreq_attr,
 };
 
-
-static int __init
-acpi_cpufreq_init (void)
+static int __init acpi_cpufreq_init(void)
 {
 	dprintk("acpi_cpufreq_init\n");
 
-	acpi_cpufreq_early_init_acpi();
+	acpi_cpufreq_early_init();
 
 	return cpufreq_register_driver(&acpi_cpufreq_driver);
 }
 
-
-static void __exit
-acpi_cpufreq_exit (void)
+static void __exit acpi_cpufreq_exit(void)
 {
-	unsigned int	i;
+	unsigned int i;
 	dprintk("acpi_cpufreq_exit\n");
 
 	cpufreq_unregister_driver(&acpi_cpufreq_driver);
@@ -627,7 +820,9 @@ acpi_cpufreq_exit (void)
 }
 
 module_param(acpi_pstate_strict, uint, 0644);
-MODULE_PARM_DESC(acpi_pstate_strict, "value 0 or non-zero. non-zero -> strict ACPI checks are performed during frequency changes.");
+MODULE_PARM_DESC(acpi_pstate_strict,
+	"value 0 or non-zero. non-zero -> strict ACPI checks are "
+	"performed during frequency changes.");
 
 late_initcall(acpi_cpufreq_init);
 module_exit(acpi_cpufreq_exit);