Contributors: 5
Author Tokens Token Proportion Commits Commit Proportion
Huang Rui 2230 81.66% 7 36.84%
Jinzhou.Su 304 11.13% 2 10.53%
Perry Yuan 167 6.11% 8 42.11%
Wyes Karny 27 0.99% 1 5.26%
Meng Li 3 0.11% 1 5.26%
Total 2731 19 


// SPDX-License-Identifier: GPL-2.0-or-later
/*
 * amd-pstate.c - AMD Processor P-state Frequency Driver
 *
 * Copyright (C) 2021 Advanced Micro Devices, Inc. All Rights Reserved.
 *
 * Author: Huang Rui <ray.huang@amd.com>
 *
 * AMD P-State introduces a new CPU performance scaling design for AMD
 * processors using the ACPI Collaborative Performance and Power Control (CPPC)
 * feature which works with the AMD SMU firmware providing a finer grained
 * frequency control range. It is to replace the legacy ACPI P-States control,
 * allows a flexible, low-latency interface for the Linux kernel to directly
 * communicate the performance hints to hardware.
 *
 * AMD P-State is supported on recent AMD Zen base CPU series include some of
 * Zen2 and Zen3 processors. _CPC needs to be present in the ACPI tables of AMD
 * P-State supported system. And there are two types of hardware implementations
 * for AMD P-State: 1) Full MSR Solution and 2) Shared Memory Solution.
 * X86_FEATURE_CPPC CPU feature flag is used to distinguish the different types.
 */

#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

#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/compiler.h>
#include <linux/dmi.h>
#include <linux/slab.h>
#include <linux/acpi.h>
#include <linux/io.h>
#include <linux/delay.h>
#include <linux/uaccess.h>
#include <linux/static_call.h>
#include <linux/amd-pstate.h>

#include <acpi/processor.h>
#include <acpi/cppc_acpi.h>

#include <asm/msr.h>
#include <asm/processor.h>
#include <asm/cpufeature.h>
#include <asm/cpu_device_id.h>
#include "amd-pstate-trace.h"

#define AMD_PSTATE_TRANSITION_LATENCY	20000
#define AMD_PSTATE_TRANSITION_DELAY	1000

/*
 * TODO: We need more time to fine tune processors with shared memory solution
 * with community together.
 *
 * There are some performance drops on the CPU benchmarks which reports from
 * Suse. We are co-working with them to fine tune the shared memory solution. So
 * we disable it by default to go acpi-cpufreq on these processors and add a
 * module parameter to be able to enable it manually for debugging.
 */
static struct cpufreq_driver amd_pstate_driver;
static int cppc_load __initdata;

static inline int pstate_enable(bool enable)
{
	return wrmsrl_safe(MSR_AMD_CPPC_ENABLE, enable);
}

static int cppc_enable(bool enable)
{
	int cpu, ret = 0;

	for_each_present_cpu(cpu) {
		ret = cppc_set_enable(cpu, enable);
		if (ret)
			return ret;
	}

	return ret;
}

DEFINE_STATIC_CALL(amd_pstate_enable, pstate_enable);

static inline int amd_pstate_enable(bool enable)
{
	return static_call(amd_pstate_enable)(enable);
}

static int pstate_init_perf(struct amd_cpudata *cpudata)
{
	u64 cap1;
	u32 highest_perf;

	int ret = rdmsrl_safe_on_cpu(cpudata->cpu, MSR_AMD_CPPC_CAP1,
				     &cap1);
	if (ret)
		return ret;

	/*
	 * TODO: Introduce AMD specific power feature.
	 *
	 * CPPC entry doesn't indicate the highest performance in some ASICs.
	 */
	highest_perf = amd_get_highest_perf();
	if (highest_perf > AMD_CPPC_HIGHEST_PERF(cap1))
		highest_perf = AMD_CPPC_HIGHEST_PERF(cap1);

	WRITE_ONCE(cpudata->highest_perf, highest_perf);

	WRITE_ONCE(cpudata->nominal_perf, AMD_CPPC_NOMINAL_PERF(cap1));
	WRITE_ONCE(cpudata->lowest_nonlinear_perf, AMD_CPPC_LOWNONLIN_PERF(cap1));
	WRITE_ONCE(cpudata->lowest_perf, AMD_CPPC_LOWEST_PERF(cap1));

	return 0;
}

static int cppc_init_perf(struct amd_cpudata *cpudata)
{
	struct cppc_perf_caps cppc_perf;
	u32 highest_perf;

	int ret = cppc_get_perf_caps(cpudata->cpu, &cppc_perf);
	if (ret)
		return ret;

	highest_perf = amd_get_highest_perf();
	if (highest_perf > cppc_perf.highest_perf)
		highest_perf = cppc_perf.highest_perf;

	WRITE_ONCE(cpudata->highest_perf, highest_perf);

	WRITE_ONCE(cpudata->nominal_perf, cppc_perf.nominal_perf);
	WRITE_ONCE(cpudata->lowest_nonlinear_perf,
		   cppc_perf.lowest_nonlinear_perf);
	WRITE_ONCE(cpudata->lowest_perf, cppc_perf.lowest_perf);

	return 0;
}

DEFINE_STATIC_CALL(amd_pstate_init_perf, pstate_init_perf);

static inline int amd_pstate_init_perf(struct amd_cpudata *cpudata)
{
	return static_call(amd_pstate_init_perf)(cpudata);
}

static void pstate_update_perf(struct amd_cpudata *cpudata, u32 min_perf,
			       u32 des_perf, u32 max_perf, bool fast_switch)
{
	if (fast_switch)
		wrmsrl(MSR_AMD_CPPC_REQ, READ_ONCE(cpudata->cppc_req_cached));
	else
		wrmsrl_on_cpu(cpudata->cpu, MSR_AMD_CPPC_REQ,
			      READ_ONCE(cpudata->cppc_req_cached));
}

static void cppc_update_perf(struct amd_cpudata *cpudata,
			     u32 min_perf, u32 des_perf,
			     u32 max_perf, bool fast_switch)
{
	struct cppc_perf_ctrls perf_ctrls;

	perf_ctrls.max_perf = max_perf;
	perf_ctrls.min_perf = min_perf;
	perf_ctrls.desired_perf = des_perf;

	cppc_set_perf(cpudata->cpu, &perf_ctrls);
}

DEFINE_STATIC_CALL(amd_pstate_update_perf, pstate_update_perf);

static inline void amd_pstate_update_perf(struct amd_cpudata *cpudata,
					  u32 min_perf, u32 des_perf,
					  u32 max_perf, bool fast_switch)
{
	static_call(amd_pstate_update_perf)(cpudata, min_perf, des_perf,
					    max_perf, fast_switch);
}

static inline bool amd_pstate_sample(struct amd_cpudata *cpudata)
{
	u64 aperf, mperf, tsc;
	unsigned long flags;

	local_irq_save(flags);
	rdmsrl(MSR_IA32_APERF, aperf);
	rdmsrl(MSR_IA32_MPERF, mperf);
	tsc = rdtsc();

	if (cpudata->prev.mperf == mperf || cpudata->prev.tsc == tsc) {
		local_irq_restore(flags);
		return false;
	}

	local_irq_restore(flags);

	cpudata->cur.aperf = aperf;
	cpudata->cur.mperf = mperf;
	cpudata->cur.tsc =  tsc;
	cpudata->cur.aperf -= cpudata->prev.aperf;
	cpudata->cur.mperf -= cpudata->prev.mperf;
	cpudata->cur.tsc -= cpudata->prev.tsc;

	cpudata->prev.aperf = aperf;
	cpudata->prev.mperf = mperf;
	cpudata->prev.tsc = tsc;

	cpudata->freq = div64_u64((cpudata->cur.aperf * cpu_khz), cpudata->cur.mperf);

	return true;
}

static void amd_pstate_update(struct amd_cpudata *cpudata, u32 min_perf,
			      u32 des_perf, u32 max_perf, bool fast_switch)
{
	u64 prev = READ_ONCE(cpudata->cppc_req_cached);
	u64 value = prev;

	des_perf = clamp_t(unsigned long, des_perf, min_perf, max_perf);
	value &= ~AMD_CPPC_MIN_PERF(~0L);
	value |= AMD_CPPC_MIN_PERF(min_perf);

	value &= ~AMD_CPPC_DES_PERF(~0L);
	value |= AMD_CPPC_DES_PERF(des_perf);

	value &= ~AMD_CPPC_MAX_PERF(~0L);
	value |= AMD_CPPC_MAX_PERF(max_perf);

	if (trace_amd_pstate_perf_enabled() && amd_pstate_sample(cpudata)) {
		trace_amd_pstate_perf(min_perf, des_perf, max_perf, cpudata->freq,
			cpudata->cur.mperf, cpudata->cur.aperf, cpudata->cur.tsc,
				cpudata->cpu, (value != prev), fast_switch);
	}

	if (value == prev)
		return;

	WRITE_ONCE(cpudata->cppc_req_cached, value);

	amd_pstate_update_perf(cpudata, min_perf, des_perf,
			       max_perf, fast_switch);
}

static int amd_pstate_verify(struct cpufreq_policy_data *policy)
{
	cpufreq_verify_within_cpu_limits(policy);

	return 0;
}

static int amd_pstate_target(struct cpufreq_policy *policy,
			     unsigned int target_freq,
			     unsigned int relation)
{
	struct cpufreq_freqs freqs;
	struct amd_cpudata *cpudata = policy->driver_data;
	unsigned long max_perf, min_perf, des_perf, cap_perf;

	if (!cpudata->max_freq)
		return -ENODEV;

	cap_perf = READ_ONCE(cpudata->highest_perf);
	min_perf = READ_ONCE(cpudata->lowest_perf);
	max_perf = cap_perf;

	freqs.old = policy->cur;
	freqs.new = target_freq;

	des_perf = DIV_ROUND_CLOSEST(target_freq * cap_perf,
				     cpudata->max_freq);

	cpufreq_freq_transition_begin(policy, &freqs);
	amd_pstate_update(cpudata, min_perf, des_perf,
			  max_perf, false);
	cpufreq_freq_transition_end(policy, &freqs, false);

	return 0;
}

static void amd_pstate_adjust_perf(unsigned int cpu,
				   unsigned long _min_perf,
				   unsigned long target_perf,
				   unsigned long capacity)
{
	unsigned long max_perf, min_perf, des_perf,
		      cap_perf, lowest_nonlinear_perf;
	struct cpufreq_policy *policy = cpufreq_cpu_get(cpu);
	struct amd_cpudata *cpudata = policy->driver_data;

	cap_perf = READ_ONCE(cpudata->highest_perf);
	lowest_nonlinear_perf = READ_ONCE(cpudata->lowest_nonlinear_perf);

	des_perf = cap_perf;
	if (target_perf < capacity)
		des_perf = DIV_ROUND_UP(cap_perf * target_perf, capacity);

	min_perf = READ_ONCE(cpudata->highest_perf);
	if (_min_perf < capacity)
		min_perf = DIV_ROUND_UP(cap_perf * _min_perf, capacity);

	if (min_perf < lowest_nonlinear_perf)
		min_perf = lowest_nonlinear_perf;

	max_perf = cap_perf;
	if (max_perf < min_perf)
		max_perf = min_perf;

	amd_pstate_update(cpudata, min_perf, des_perf, max_perf, true);
}

static int amd_get_min_freq(struct amd_cpudata *cpudata)
{
	struct cppc_perf_caps cppc_perf;

	int ret = cppc_get_perf_caps(cpudata->cpu, &cppc_perf);
	if (ret)
		return ret;

	/* Switch to khz */
	return cppc_perf.lowest_freq * 1000;
}

static int amd_get_max_freq(struct amd_cpudata *cpudata)
{
	struct cppc_perf_caps cppc_perf;
	u32 max_perf, max_freq, nominal_freq, nominal_perf;
	u64 boost_ratio;

	int ret = cppc_get_perf_caps(cpudata->cpu, &cppc_perf);
	if (ret)
		return ret;

	nominal_freq = cppc_perf.nominal_freq;
	nominal_perf = READ_ONCE(cpudata->nominal_perf);
	max_perf = READ_ONCE(cpudata->highest_perf);

	boost_ratio = div_u64(max_perf << SCHED_CAPACITY_SHIFT,
			      nominal_perf);

	max_freq = nominal_freq * boost_ratio >> SCHED_CAPACITY_SHIFT;

	/* Switch to khz */
	return max_freq * 1000;
}

static int amd_get_nominal_freq(struct amd_cpudata *cpudata)
{
	struct cppc_perf_caps cppc_perf;

	int ret = cppc_get_perf_caps(cpudata->cpu, &cppc_perf);
	if (ret)
		return ret;

	/* Switch to khz */
	return cppc_perf.nominal_freq * 1000;
}

static int amd_get_lowest_nonlinear_freq(struct amd_cpudata *cpudata)
{
	struct cppc_perf_caps cppc_perf;
	u32 lowest_nonlinear_freq, lowest_nonlinear_perf,
	    nominal_freq, nominal_perf;
	u64 lowest_nonlinear_ratio;

	int ret = cppc_get_perf_caps(cpudata->cpu, &cppc_perf);
	if (ret)
		return ret;

	nominal_freq = cppc_perf.nominal_freq;
	nominal_perf = READ_ONCE(cpudata->nominal_perf);

	lowest_nonlinear_perf = cppc_perf.lowest_nonlinear_perf;

	lowest_nonlinear_ratio = div_u64(lowest_nonlinear_perf << SCHED_CAPACITY_SHIFT,
					 nominal_perf);

	lowest_nonlinear_freq = nominal_freq * lowest_nonlinear_ratio >> SCHED_CAPACITY_SHIFT;

	/* Switch to khz */
	return lowest_nonlinear_freq * 1000;
}

static int amd_pstate_set_boost(struct cpufreq_policy *policy, int state)
{
	struct amd_cpudata *cpudata = policy->driver_data;
	int ret;

	if (!cpudata->boost_supported) {
		pr_err("Boost mode is not supported by this processor or SBIOS\n");
		return -EINVAL;
	}

	if (state)
		policy->cpuinfo.max_freq = cpudata->max_freq;
	else
		policy->cpuinfo.max_freq = cpudata->nominal_freq;

	policy->max = policy->cpuinfo.max_freq;

	ret = freq_qos_update_request(&cpudata->req[1],
				      policy->cpuinfo.max_freq);
	if (ret < 0)
		return ret;

	return 0;
}

static void amd_pstate_boost_init(struct amd_cpudata *cpudata)
{
	u32 highest_perf, nominal_perf;

	highest_perf = READ_ONCE(cpudata->highest_perf);
	nominal_perf = READ_ONCE(cpudata->nominal_perf);

	if (highest_perf <= nominal_perf)
		return;

	cpudata->boost_supported = true;
	amd_pstate_driver.boost_enabled = true;
}

static void amd_perf_ctl_reset(unsigned int cpu)
{
	wrmsrl_on_cpu(cpu, MSR_AMD_PERF_CTL, 0);
}

static int amd_pstate_cpu_init(struct cpufreq_policy *policy)
{
	int min_freq, max_freq, nominal_freq, lowest_nonlinear_freq, ret;
	struct device *dev;
	struct amd_cpudata *cpudata;

	/*
	 * Resetting PERF_CTL_MSR will put the CPU in P0 frequency,
	 * which is ideal for initialization process.
	 */
	amd_perf_ctl_reset(policy->cpu);
	dev = get_cpu_device(policy->cpu);
	if (!dev)
		return -ENODEV;

	cpudata = kzalloc(sizeof(*cpudata), GFP_KERNEL);
	if (!cpudata)
		return -ENOMEM;

	cpudata->cpu = policy->cpu;

	ret = amd_pstate_init_perf(cpudata);
	if (ret)
		goto free_cpudata1;

	min_freq = amd_get_min_freq(cpudata);
	max_freq = amd_get_max_freq(cpudata);
	nominal_freq = amd_get_nominal_freq(cpudata);
	lowest_nonlinear_freq = amd_get_lowest_nonlinear_freq(cpudata);

	if (min_freq < 0 || max_freq < 0 || min_freq > max_freq) {
		dev_err(dev, "min_freq(%d) or max_freq(%d) value is incorrect\n",
			min_freq, max_freq);
		ret = -EINVAL;
		goto free_cpudata1;
	}

	policy->cpuinfo.transition_latency = AMD_PSTATE_TRANSITION_LATENCY;
	policy->transition_delay_us = AMD_PSTATE_TRANSITION_DELAY;

	policy->min = min_freq;
	policy->max = max_freq;

	policy->cpuinfo.min_freq = min_freq;
	policy->cpuinfo.max_freq = max_freq;

	/* It will be updated by governor */
	policy->cur = policy->cpuinfo.min_freq;

	if (boot_cpu_has(X86_FEATURE_CPPC))
		policy->fast_switch_possible = true;

	ret = freq_qos_add_request(&policy->constraints, &cpudata->req[0],
				   FREQ_QOS_MIN, policy->cpuinfo.min_freq);
	if (ret < 0) {
		dev_err(dev, "Failed to add min-freq constraint (%d)\n", ret);
		goto free_cpudata1;
	}

	ret = freq_qos_add_request(&policy->constraints, &cpudata->req[1],
				   FREQ_QOS_MAX, policy->cpuinfo.max_freq);
	if (ret < 0) {
		dev_err(dev, "Failed to add max-freq constraint (%d)\n", ret);
		goto free_cpudata2;
	}

	/* Initial processor data capability frequencies */
	cpudata->max_freq = max_freq;
	cpudata->min_freq = min_freq;
	cpudata->nominal_freq = nominal_freq;
	cpudata->lowest_nonlinear_freq = lowest_nonlinear_freq;

	policy->driver_data = cpudata;

	amd_pstate_boost_init(cpudata);

	return 0;

free_cpudata2:
	freq_qos_remove_request(&cpudata->req[0]);
free_cpudata1:
	kfree(cpudata);
	return ret;
}

static int amd_pstate_cpu_exit(struct cpufreq_policy *policy)
{
	struct amd_cpudata *cpudata = policy->driver_data;

	freq_qos_remove_request(&cpudata->req[1]);
	freq_qos_remove_request(&cpudata->req[0]);
	kfree(cpudata);

	return 0;
}

static int amd_pstate_cpu_resume(struct cpufreq_policy *policy)
{
	int ret;

	ret = amd_pstate_enable(true);
	if (ret)
		pr_err("failed to enable amd-pstate during resume, return %d\n", ret);

	return ret;
}

static int amd_pstate_cpu_suspend(struct cpufreq_policy *policy)
{
	int ret;

	ret = amd_pstate_enable(false);
	if (ret)
		pr_err("failed to disable amd-pstate during suspend, return %d\n", ret);

	return ret;
}

/* Sysfs attributes */

/*
 * This frequency is to indicate the maximum hardware frequency.
 * If boost is not active but supported, the frequency will be larger than the
 * one in cpuinfo.
 */
static ssize_t show_amd_pstate_max_freq(struct cpufreq_policy *policy,
					char *buf)
{
	int max_freq;
	struct amd_cpudata *cpudata = policy->driver_data;

	max_freq = amd_get_max_freq(cpudata);
	if (max_freq < 0)
		return max_freq;

	return sprintf(&buf[0], "%u\n", max_freq);
}

static ssize_t show_amd_pstate_lowest_nonlinear_freq(struct cpufreq_policy *policy,
						     char *buf)
{
	int freq;
	struct amd_cpudata *cpudata = policy->driver_data;

	freq = amd_get_lowest_nonlinear_freq(cpudata);
	if (freq < 0)
		return freq;

	return sprintf(&buf[0], "%u\n", freq);
}

/*
 * In some of ASICs, the highest_perf is not the one in the _CPC table, so we
 * need to expose it to sysfs.
 */
static ssize_t show_amd_pstate_highest_perf(struct cpufreq_policy *policy,
					    char *buf)
{
	u32 perf;
	struct amd_cpudata *cpudata = policy->driver_data;

	perf = READ_ONCE(cpudata->highest_perf);

	return sprintf(&buf[0], "%u\n", perf);
}

cpufreq_freq_attr_ro(amd_pstate_max_freq);
cpufreq_freq_attr_ro(amd_pstate_lowest_nonlinear_freq);

cpufreq_freq_attr_ro(amd_pstate_highest_perf);

static struct freq_attr *amd_pstate_attr[] = {
	&amd_pstate_max_freq,
	&amd_pstate_lowest_nonlinear_freq,
	&amd_pstate_highest_perf,
	NULL,
};

static struct cpufreq_driver amd_pstate_driver = {
	.flags		= CPUFREQ_CONST_LOOPS | CPUFREQ_NEED_UPDATE_LIMITS,
	.verify		= amd_pstate_verify,
	.target		= amd_pstate_target,
	.init		= amd_pstate_cpu_init,
	.exit		= amd_pstate_cpu_exit,
	.suspend	= amd_pstate_cpu_suspend,
	.resume		= amd_pstate_cpu_resume,
	.set_boost	= amd_pstate_set_boost,
	.name		= "amd-pstate",
	.attr		= amd_pstate_attr,
};

static int __init amd_pstate_init(void)
{
	int ret;

	if (boot_cpu_data.x86_vendor != X86_VENDOR_AMD)
		return -ENODEV;
	/*
	 * by default the pstate driver is disabled to load
	 * enable the amd_pstate passive mode driver explicitly
	 * with amd_pstate=passive in kernel command line
	 */
	if (!cppc_load) {
		pr_debug("driver load is disabled, boot with amd_pstate=passive to enable this\n");
		return -ENODEV;
	}

	if (!acpi_cpc_valid()) {
		pr_warn_once("the _CPC object is not present in SBIOS or ACPI disabled\n");
		return -ENODEV;
	}

	/* don't keep reloading if cpufreq_driver exists */
	if (cpufreq_get_current_driver())
		return -EEXIST;

	/* capability check */
	if (boot_cpu_has(X86_FEATURE_CPPC)) {
		pr_debug("AMD CPPC MSR based functionality is supported\n");
		amd_pstate_driver.adjust_perf = amd_pstate_adjust_perf;
	} else {
		pr_debug("AMD CPPC shared memory based functionality is supported\n");
		static_call_update(amd_pstate_enable, cppc_enable);
		static_call_update(amd_pstate_init_perf, cppc_init_perf);
		static_call_update(amd_pstate_update_perf, cppc_update_perf);
	}

	/* enable amd pstate feature */
	ret = amd_pstate_enable(true);
	if (ret) {
		pr_err("failed to enable amd-pstate with return %d\n", ret);
		return ret;
	}

	ret = cpufreq_register_driver(&amd_pstate_driver);
	if (ret)
		pr_err("failed to register amd_pstate_driver with return %d\n",
		       ret);

	return ret;
}
device_initcall(amd_pstate_init);

static int __init amd_pstate_param(char *str)
{
	if (!str)
		return -EINVAL;

	if (!strcmp(str, "disable")) {
		cppc_load = 0;
		pr_info("driver is explicitly disabled\n");
	} else if (!strcmp(str, "passive"))
		cppc_load = 1;

	return 0;
}
early_param("amd_pstate", amd_pstate_param);

MODULE_AUTHOR("Huang Rui <ray.huang@amd.com>");
MODULE_DESCRIPTION("AMD Processor P-state Frequency Driver");
MODULE_LICENSE("GPL");