path: root/drivers/cpufreq/powernv-cpufreq.c

// SPDX-License-Identifier: GPL-2.0-or-later
/*
 * POWERNV cpufreq driver for the IBM POWER processors
 *
 * (C) Copyright IBM 2014
 *
 * Author: Vaidyanathan Srinivasan <svaidy at linux.vnet.ibm.com>
 */

#define pr_fmt(fmt)	"powernv-cpufreq: " fmt

#include <linux/kernel.h>
#include <linux/sysfs.h>
#include <linux/cpumask.h>
#include <linux/module.h>
#include <linux/cpufreq.h>
#include <linux/smp.h>
#include <linux/of.h>
#include <linux/reboot.h>
#include <linux/slab.h>
#include <linux/cpu.h>
#include <linux/hashtable.h>
#include <trace/events/power.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>
#include <linux/timer.h>

#define POWERNV_MAX_PSTATES_ORDER  8
#define POWERNV_MAX_PSTATES	(1UL << (POWERNV_MAX_PSTATES_ORDER))
#define PMSR_PSAFE_ENABLE	(1UL << 30)
#define PMSR_SPR_EM_DISABLE	(1UL << 31)
#define MAX_PSTATE_SHIFT	32
#define LPSTATE_SHIFT		48
#define GPSTATE_SHIFT		56
#define MAX_NR_CHIPS		32

#define MAX_RAMP_DOWN_TIME				5120
/*
 * On an idle system we want the global pstate to ramp-down from max value to
 * min over a span of ~5 secs. Also we want it to initially ramp-down slowly and
 * then ramp-down rapidly later on.
 *
 * This gives a percentage rampdown for time elapsed in milliseconds.
 * ramp_down_percentage = ((ms * ms) >> 18)
 *			~= 3.8 * (sec * sec)
 *
 * At 0 ms	ramp_down_percent = 0
 * At 5120 ms	ramp_down_percent = 100
 */
#define ramp_down_percent(time)		((time * time) >> 18)

/* Interval after which the timer is queued to bring down global pstate */
#define GPSTATE_TIMER_INTERVAL				2000

/**
 * struct global_pstate_info -	Per policy data structure to maintain history of
 *				global pstates
 * @highest_lpstate_idx:	The local pstate index from which we are
 *				ramping down
 * @elapsed_time:		Time in ms spent in ramping down from
 *				highest_lpstate_idx
 * @last_sampled_time:		Time from boot in ms when global pstates were
 *				last set
 * @last_lpstate_idx:		Last set value of local pstate and global
 * @last_gpstate_idx:		pstate in terms of cpufreq table index
 * @timer:			Is used for ramping down if cpu goes idle for
 *				a long time with global pstate held high
 * @gpstate_lock:		A spinlock to maintain synchronization between
 *				routines called by the timer handler and
 *				governer's target_index calls
 * @policy:			Associated CPUFreq policy
 */
struct global_pstate_info {
	int highest_lpstate_idx;
	unsigned int elapsed_time;
	unsigned int last_sampled_time;
	int last_lpstate_idx;
	int last_gpstate_idx;
	spinlock_t gpstate_lock;
	struct timer_list timer;
	struct cpufreq_policy *policy;
};

static struct cpufreq_frequency_table powernv_freqs[POWERNV_MAX_PSTATES+1];

static DEFINE_HASHTABLE(pstate_revmap, POWERNV_MAX_PSTATES_ORDER);
/**
 * struct pstate_idx_revmap_data: Entry in the hashmap pstate_revmap
 *				  indexed by a function of pstate id.
 *
 * @pstate_id: pstate id for this entry.
 *
 * @cpufreq_table_idx: Index into the powernv_freqs
 *		       cpufreq_frequency_table for frequency
 *		       corresponding to pstate_id.
 *
 * @hentry: hlist_node that hooks this entry into the pstate_revmap
 *	    hashtable
 */
struct pstate_idx_revmap_data {
	u8 pstate_id;
	unsigned int cpufreq_table_idx;
	struct hlist_node hentry;
};

static bool rebooting, throttled, occ_reset;

static const char * const throttle_reason[] = {
	"No throttling",
	"Power Cap",
	"Processor Over Temperature",
	"Power Supply Failure",
	"Over Current",
	"OCC Reset"
};

enum throttle_reason_type {
	NO_THROTTLE = 0,
	POWERCAP,
	CPU_OVERTEMP,
	POWER_SUPPLY_FAILURE,
	OVERCURRENT,
	OCC_RESET_THROTTLE,
	OCC_MAX_REASON
};

static struct chip {
	unsigned int id;
	bool throttled;
	bool restore;
	u8 throttle_reason;
	cpumask_t mask;
	struct work_struct throttle;
	int throttle_turbo;
	int throttle_sub_turbo;
	int reason[OCC_MAX_REASON];
} *chips;

static int nr_chips;
static DEFINE_PER_CPU(struct chip *, chip_info);

/*
 * Note:
 * The set of pstates consists of contiguous integers.
 * powernv_pstate_info stores the index of the frequency table for
 * max, min and nominal frequencies. It also stores number of
 * available frequencies.
 *
 * powernv_pstate_info.nominal indicates the index to the highest
 * non-turbo frequency.
 */
static struct powernv_pstate_info {
	unsigned int min;
	unsigned int max;
	unsigned int nominal;
	unsigned int nr_pstates;
	bool wof_enabled;
} powernv_pstate_info;

static inline u8 extract_pstate(u64 pmsr_val, unsigned int shift)
{
	return ((pmsr_val >> shift) & 0xFF);
}

#define extract_local_pstate(x) extract_pstate(x, LPSTATE_SHIFT)
#define extract_global_pstate(x) extract_pstate(x, GPSTATE_SHIFT)
#define extract_max_pstate(x)  extract_pstate(x, MAX_PSTATE_SHIFT)

/* Use following functions for conversions between pstate_id and index */

/*
 * idx_to_pstate : Returns the pstate id corresponding to the
 *		   frequency in the cpufreq frequency table
 *		   powernv_freqs indexed by @i.
 *
 *		   If @i is out of bound, this will return the pstate
 *		   corresponding to the nominal frequency.
 */
static inline u8 idx_to_pstate(unsigned int i)
{
	if (unlikely(i >= powernv_pstate_info.nr_pstates)) {
		pr_warn_once("idx_to_pstate: index %u is out of bound\n", i);
		return powernv_freqs[powernv_pstate_info.nominal].driver_data;
	}

	return powernv_freqs[i].driver_data;
}

/*
 * pstate_to_idx : Returns the index in the cpufreq frequencytable
 *		   powernv_freqs for the frequency whose corresponding
 *		   pstate id is @pstate.
 *
 *		   If no frequency corresponding to @pstate is found,
 *		   this will return the index of the nominal
 *		   frequency.
 */
static unsigned int pstate_to_idx(u8 pstate)
{
	unsigned int key = pstate % POWERNV_MAX_PSTATES;
	struct pstate_idx_revmap_data *revmap_data;

	hash_for_each_possible(pstate_revmap, revmap_data