[mirror_ubuntu-bionic-kernel.git] / drivers / cpuidle / cpuidle-powernv.c

// SPDX-License-Identifier: GPL-2.0
/*
 *  cpuidle-powernv - idle state cpuidle driver.
 *  Adapted from drivers/cpuidle/cpuidle-pseries
 *
 */

#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/init.h>
#include <linux/moduleparam.h>
#include <linux/cpuidle.h>
#include <linux/cpu.h>
#include <linux/notifier.h>
#include <linux/clockchips.h>
#include <linux/of.h>
#include <linux/slab.h>

#include <asm/machdep.h>
#include <asm/firmware.h>
#include <asm/opal.h>
#include <asm/runlatch.h>
#include <asm/cpuidle.h>

/*
 * Expose only those Hardware idle states via the cpuidle framework
 * that have latency value below POWERNV_THRESHOLD_LATENCY_NS.
 */
#define POWERNV_THRESHOLD_LATENCY_NS 200000

static struct cpuidle_driver powernv_idle_driver = {
	.name             = "powernv_idle",
	.owner            = THIS_MODULE,
};

static int max_idle_state __read_mostly;
static struct cpuidle_state *cpuidle_state_table __read_mostly;

struct stop_psscr_table {
	u64 val;
	u64 mask;
};

static struct stop_psscr_table stop_psscr_table[CPUIDLE_STATE_MAX] __read_mostly;

static u64 default_snooze_timeout __read_mostly;
static bool snooze_timeout_en __read_mostly;

static u64 get_snooze_timeout(struct cpuidle_device *dev,
			      struct cpuidle_driver *drv,
			      int index)
{
	int i;

	if (unlikely(!snooze_timeout_en))
		return default_snooze_timeout;

	for (i = index + 1; i < drv->state_count; i++) {
		struct cpuidle_state *s = &drv->states[i];
		struct cpuidle_state_usage *su = &dev->states_usage[i];

		if (s->disabled || su->disable)
			continue;

		return s->target_residency * tb_ticks_per_usec;
	}

	return default_snooze_timeout;
}

static int snooze_loop(struct cpuidle_device *dev,
			struct cpuidle_driver *drv,
			int index)
{
	u64 snooze_exit_time;

	set_thread_flag(TIF_POLLING_NRFLAG);

	local_irq_enable();

	snooze_exit_time = get_tb() + get_snooze_timeout(dev, drv, index);
	ppc64_runlatch_off();
	HMT_very_low();
	while (!need_resched()) {
		if (likely(snooze_timeout_en) && get_tb() > snooze_exit_time) {
			/*
			 * Task has not woken up but we are exiting the polling
			 * loop anyway. Require a barrier after polling is
			 * cleared to order subsequent test of need_resched().
			 */
			clear_thread_flag(TIF_POLLING_NRFLAG);
			smp_mb();
			break;
		}
	}

	HMT_medium();
	ppc64_runlatch_on();
	clear_thread_flag(TIF_POLLING_NRFLAG);

	return index;
}

static int nap_loop(struct cpuidle_device *dev,
			struct cpuidle_driver *drv,
			int index)
{
	power7_idle_type(PNV_THREAD_NAP);

	return index;
}

/* Register for fastsleep only in oneshot mode of broadcast */
#ifdef CONFIG_TICK_ONESHOT
static int fastsleep_loop(struct cpuidle_device *dev,
				struct cpuidle_driver *drv,
				int index)
{
	unsigned long old_lpcr = mfspr(SPRN_LPCR);
	unsigned long new_lpcr;

	if (unlikely(system_state < SYSTEM_RUNNING))
		return index;

	new_lpcr = old_lpcr;
	/* Do not exit powersave upon decrementer as we've setup the timer
	 * offload.
	 */
	new_lpcr &= ~LPCR_PECE1;

	mtspr(SPRN_LPCR, new_lpcr);

	power7_idle_type(PNV_THREAD_SLEEP);

	mtspr(SPRN_LPCR, old_lpcr);

	return index;
}
#endif

static int stop_loop(struct cpuidle_device *dev,
		     struct cpuidle_driver *drv,
		     int index)
{
	power9_idle_type(stop_psscr_table[index].val,
			 stop_psscr_table[index].mask);
	return index;
}

/*
 * States for dedicated partition case.
 */
static struct cpuidle_state powernv_states[CPUIDLE_STATE_MAX] = {
	{ /* Snooze */
		.name = "snooze",
		.desc = "snooze",
		.exit_latency = 0,
		.target_residency = 0,
		.enter = snooze_loop },
};

static int powernv_cpuidle_cpu_online(unsigned int cpu)
{
	struct cpuidle_device *dev = per_cpu(cpuidle_devices, cpu);

	if (dev && cpuidle_get_driver()) {
		cpuidle_pause_and_lock();
		cpuidle_enable_device(dev);
		cpuidle_resume_and_unlock();
	}
	return 0;
}

static int powernv_cpuidle_cpu_dead(unsigned int cpu)
{
	struct cpuidle_device *dev = per_cpu(cpuidle_devices, cpu);

	if (dev && cpuidle_get_driver()) {
		cpuidle_pause_and_lock();
		cpuidle_disable_device(dev);
		cpuidle_resume_and_unlock();
	}
	return 0;
}

/*
 * powernv_cpuidle_driver_init()
 */
static int powernv_cpuidle_driver_init(void)
{
	int idle_state;
	struct cpuidle_driver *drv = &powernv_idle_driver;

	drv->state_count = 0;

	for (idle_state = 0; idle_state < max_idle_state; ++idle_state) {
		/* Is the state not enabled? */
		if (cpuidle_state_table[idle_state].enter == NULL)
			continue;

		drv->states[drv->state_count] =	/* structure copy */
			cpuidle_state_table[idle_state];

		drv->state_count += 1;
	}

	/*
	 * On the PowerNV platform cpu_present may be less than cpu_possible in
	 * cases when firmware detects the CPU, but it is not available to the
	 * OS.  If CONFIG_HOTPLUG_CPU=n, then such CPUs are not hotplugable at
	 * run time and hence cpu_devices are not created for those CPUs by the
	 * generic topology_init().
	 *
	 * drv->cpumask defaults to cpu_possible_mask in
	 * __cpuidle_driver_init().  This breaks cpuidle on PowerNV where
	 * cpu_devices are not created for CPUs in cpu_possible_mask that
	 * cannot be hot-added later at run time.
	 *
	 * Trying cpuidle_register_device() on a CPU without a cpu_device is
	 * incorrect, so pass a correct CPU mask to the generic cpuidle driver.
	 */

	drv->cpumask = (struct cpumask *)cpu_present_mask;

	return 0;
}

static inline void add_powernv_state(int index, const char *name,
				     unsigned int flags,
				     int (*idle_fn)(struct cpuidle_device *,
						    struct cpuidle_driver *,
						    int),
				     unsigned int target_residency,
				     unsigned int exit_latency,
				     u64 psscr_val, u64 psscr_mask)
{
	strlcpy(powernv_states[index].name, name, CPUIDLE_NAME_LEN);
	strlcpy(powernv_states[index].desc, name, CPUIDLE_NAME_LEN);
	powernv_states[index].flags = flags;
	powernv_states[index].target_residency = target_residency;
	powernv_states[index].exit_latency = exit_latency;
	powernv_states[index].enter = idle_fn;
	stop_psscr_table[index].val = psscr_val;
	stop_psscr_table[index].mask = psscr_mask;
}

/*
 * Returns 0 if prop1_len == prop2_len. Else returns -1
 */
static inline int validate_dt_prop_sizes(const char *prop1, int prop1_len,
					 const char *prop2, int prop2_len)
{
	if (prop1_len == prop2_len)
		return 0;

	pr_warn("cpuidle-powernv: array sizes don't match for %s and %s\n",
		prop1, prop2);
	return -1;
}

extern u32 pnv_get_supported_cpuidle_states(void);
static int powernv_add_idle_states(void)
{
	struct device_node *power_mgt;
	int nr_idle_states = 1; /* Snooze */
	int dt_idle_states, count;
	u32 latency_ns[CPUIDLE_STATE_MAX];
	u32 residency_ns[CPUIDLE_STATE_MAX];
	u32 flags[CPUIDLE_STATE_MAX];
	u64 psscr_val[CPUIDLE_STATE_MAX];
	u64 psscr_mask[CPUIDLE_STATE_MAX];
	const char *names[CPUIDLE_STATE_MAX];
	u32 has_stop_states = 0;
	int i, rc;
	u32 supported_flags = pnv_get_supported_cpuidle_states();


	/* Currently we have snooze statically defined */

	power_mgt = of_find_node_by_path("/ibm,opal/power-mgt");
	if (!power_mgt) {
		pr_warn("opal: PowerMgmt Node not found\n");
		goto out;
	}

	/* Read values of any property to determine the num of idle states */
	dt_idle_states = of_property_count_u32_elems(power_mgt, "ibm,cpu-idle-state-flags");
	if (dt_idle_states < 0) {
		pr_warn("cpuidle-powernv: no idle states found in the DT\n");
		goto out;
	}

	count = of_property_count_u32_elems(power_mgt,
					    "ibm,cpu-idle-state-latencies-ns");

	if (validate_dt_prop_sizes("ibm,cpu-idle-state-flags", dt_idle_states,
				   "ibm,cpu-idle-state-latencies-ns",
				   count) != 0)
		goto out;

	count = of_property_count_strings(power_mgt,
					  "ibm,cpu-idle-state-names");
	if (validate_dt_prop_sizes("ibm,cpu-idle-state-flags", dt_idle_states,
				   "ibm,cpu-idle-state-names",
				   count) != 0)
		goto out;

	/*
	 * Since snooze is used as first idle state, max idle states allowed is
	 * CPUIDLE_STATE_MAX -1
	 */
	if (dt_idle_states > CPUIDLE_STATE_MAX - 1) {
		pr_warn("cpuidle-powernv: discovered idle states more than allowed");
		dt_idle_states = CPUIDLE_STATE_MAX - 1;
	}

	if (of_property_read_u32_array(power_mgt,
			"ibm,cpu-idle-state-flags", flags, dt_idle_states)) {
		pr_warn("cpuidle-powernv : missing ibm,cpu-idle-state-flags in DT\n");
		goto out;
	}

	if (of_property_read_u32_array(power_mgt,
		"ibm,cpu-idle-state-latencies-ns", latency_ns,
		dt_idle_states)) {
		pr_warn("cpuidle-powernv: missing ibm,cpu-idle-state-latencies-ns in DT\n");
		goto out;
	}
	if (of_property_read_string_array(power_mgt,
		"ibm,cpu-idle-state-names", names, dt_idle_states) < 0) {
		pr_warn("cpuidle-powernv: missing ibm,cpu-idle-state-names in DT\n");
		goto out;
	}

	/*
	 * If the idle states use stop instruction, probe for psscr values
	 * and psscr mask which are necessary to specify required stop level.
	 */
	has_stop_states = (flags[0] &
			   (OPAL_PM_STOP_INST_FAST | OPAL_PM_STOP_INST_DEEP));
	if (has_stop_states) {
		count = of_property_count_u64_elems(power_mgt,
						    "ibm,cpu-idle-state-psscr");
		if (validate_dt_prop_sizes("ibm,cpu-idle-state-flags",
					   dt_idle_states,
					   "ibm,cpu-idle-state-psscr",
					   count) != 0)
			goto out;

		count = of_property_count_u64_elems(power_mgt,
						    "ibm,cpu-idle-state-psscr-mask");
		if (validate_dt_prop_sizes("ibm,cpu-idle-state-flags",
					   dt_idle_states,
					   "ibm,cpu-idle-state-psscr-mask",
					   count) != 0)
			goto out;

		if (of_property_read_u64_array(power_mgt,
		    "ibm,cpu-idle-state-psscr", psscr_val, dt_idle_states)) {
			pr_warn("cpuidle-powernv: missing ibm,cpu-idle-state-psscr in DT\n");
			goto out;
		}

		if (of_property_read_u64_array(power_mgt,
					       "ibm,cpu-idle-state-psscr-mask",
						psscr_mask, dt_idle_states)) {
			pr_warn("cpuidle-powernv:Missing ibm,cpu-idle-state-psscr-mask in DT\n");
			goto out;
		}
	}

	count = of_property_count_u32_elems(power_mgt,
					    "ibm,cpu-idle-state-residency-ns");

	if (count < 0) {
		rc = count;
	} else if (validate_dt_prop_sizes("ibm,cpu-idle-state-flags",
					  dt_idle_states,
					  "ibm,cpu-idle-state-residency-ns",
					  count) != 0) {
		goto out;
	} else {
		rc = of_property_read_u32_array(power_mgt,
						"ibm,cpu-idle-state-residency-ns",
						residency_ns, dt_idle_states);
	}

	for (i = 0; i < dt_idle_states; i++) {
		unsigned int exit_latency, target_residency;
		bool stops_timebase = false;

		/*
		 * Skip the platform idle state whose flag isn't in
		 * the supported_cpuidle_states flag mask.
		 */
		if ((flags[i] & supported_flags) != flags[i])
			continue;
		/*
		 * If an idle state has exit latency beyond
		 * POWERNV_THRESHOLD_LATENCY_NS then don't use it
		 * in cpu-idle.
		 */
		if (latency_ns[i] > POWERNV_THRESHOLD_LATENCY_NS)
			continue;
		/*
		 * Firmware passes residency and latency values in ns.
		 * cpuidle expects it in us.
		 */
		exit_latency = DIV_ROUND_UP(latency_ns[i], 1000);
		if (!rc)
			target_residency = DIV_ROUND_UP(residency_ns[i], 1000);
		else
			target_residency = 0;

		if (has_stop_states) {
			int err = validate_psscr_val_mask(&psscr_val[i],
							  &psscr_mask[i],
							  flags[i]);
			if (err) {
				report_invalid_psscr_val(psscr_val[i], err);
				continue;
			}
		}

		if (flags[i] & OPAL_PM_TIMEBASE_STOP)
			stops_timebase = true;

		/*
		 * For nap and fastsleep, use default target_residency
		 * values if f/w does not expose it.
		 */
		if (flags[i] & OPAL_PM_NAP_ENABLED) {
			if (!rc)
				target_residency = 100;
			/* Add NAP state */
			add_powernv_state(nr_idle_states, "Nap",
					  CPUIDLE_FLAG_NONE, nap_loop,
					  target_residency, exit_latency, 0, 0);
		} else if (has_stop_states && !stops_timebase) {
			add_powernv_state(nr_idle_states, names[i],
					  CPUIDLE_FLAG_NONE, stop_loop,
					  target_residency, exit_latency,
					  psscr_val[i], psscr_mask[i]);
		}

		/*
		 * All cpuidle states with CPUIDLE_FLAG_TIMER_STOP set must come
		 * within this config dependency check.
		 */
#ifdef CONFIG_TICK_ONESHOT
		else if (flags[i] & OPAL_PM_SLEEP_ENABLED ||
			 flags[i] & OPAL_PM_SLEEP_ENABLED_ER1) {
			if (!rc)
				target_residency = 300000;
			/* Add FASTSLEEP state */
			add_powernv_state(nr_idle_states, "FastSleep",
					  CPUIDLE_FLAG_TIMER_STOP,
					  fastsleep_loop,
					  target_residency, exit_latency, 0, 0);
		} else if (has_stop_states && stops_timebase) {
			add_powernv_state(nr_idle_states, names[i],
					  CPUIDLE_FLAG_TIMER_STOP, stop_loop,
					  target_residency, exit_latency,
					  psscr_val[i], psscr_mask[i]);
		}
#endif
		else
			continue;
		nr_idle_states++;
	}
out:
	return nr_idle_states;
}

/*
 * powernv_idle_probe()
 * Choose state table for shared versus dedicated partition
 */
static int powernv_idle_probe(void)
{
	if (cpuidle_disable != IDLE_NO_OVERRIDE)
		return -ENODEV;

	if (firmware_has_feature(FW_FEATURE_OPAL)) {
		cpuidle_state_table = powernv_states;
		/* Device tree can indicate more idle states */
		max_idle_state = powernv_add_idle_states();
		default_snooze_timeout = TICK_USEC * tb_ticks_per_usec;
		if (max_idle_state > 1)
			snooze_timeout_en = true;
 	} else
 		return -ENODEV;

	return 0;
}

static int __init powernv_processor_idle_init(void)
{
	int retval;

	retval = powernv_idle_probe();
	if (retval)
		return retval;

	powernv_cpuidle_driver_init();
	retval = cpuidle_register(&powernv_idle_driver, NULL);
	if (retval) {
		printk(KERN_DEBUG "Registration of powernv driver failed.\n");
		return retval;
	}

	retval = cpuhp_setup_state_nocalls(CPUHP_AP_ONLINE_DYN,
					   "cpuidle/powernv:online",
					   powernv_cpuidle_cpu_online, NULL);
	WARN_ON(retval < 0);
	retval = cpuhp_setup_state_nocalls(CPUHP_CPUIDLE_DEAD,
					   "cpuidle/powernv:dead", NULL,
					   powernv_cpuidle_cpu_dead);
	WARN_ON(retval < 0);
	printk(KERN_DEBUG "powernv_idle_driver registered\n");
	return 0;
}

device_initcall(powernv_processor_idle_init);
Commit	Line	Data
b2441318	1	// SPDX-License-Identifier: GPL-2.0
2c2e6ecf DD	2	/*
	3	* cpuidle-powernv - idle state cpuidle driver.
	4	* Adapted from drivers/cpuidle/cpuidle-pseries
	5	*
	6	*/
	7
	8	#include <linux/kernel.h>
	9	#include <linux/module.h>
	10	#include <linux/init.h>
	11	#include <linux/moduleparam.h>
	12	#include <linux/cpuidle.h>
	13	#include <linux/cpu.h>
	14	#include <linux/notifier.h>
0d948730	15	#include <linux/clockchips.h>
0888839c	16	#include <linux/of.h>
92c83ff5	17	#include <linux/slab.h>
2c2e6ecf DD	18
	19	#include <asm/machdep.h>
	20	#include <asm/firmware.h>
8eb8ac89	21	#include <asm/opal.h>
591ac0cb	22	#include <asm/runlatch.h>
09206b60	23	#include <asm/cpuidle.h>
2c2e6ecf	24
9e9fc6f0 GS	25	/*
	26	* Expose only those Hardware idle states via the cpuidle framework
	27	* that have latency value below POWERNV_THRESHOLD_LATENCY_NS.
	28	*/
3005c597 SP	29	#define POWERNV_THRESHOLD_LATENCY_NS 200000
3005c597 SP	30
ed61390b	31	static struct cpuidle_driver powernv_idle_driver = {
2c2e6ecf DD	32	.name = "powernv_idle",
	33	.owner = THIS_MODULE,
	34	};
	35
624e46d0 NP	36	static int max_idle_state __read_mostly;
624e46d0 NP	37	static struct cpuidle_state *cpuidle_state_table __read_mostly;
3005c597	38
09206b60 GS	39	struct stop_psscr_table {
	40	u64 val;
	41	u64 mask;
	42	};
	43
624e46d0	44	static struct stop_psscr_table stop_psscr_table[CPUIDLE_STATE_MAX] __read_mostly;
3005c597	45
a11240db	46	static u64 default_snooze_timeout __read_mostly;
624e46d0	47	static bool snooze_timeout_en __read_mostly;
2c2e6ecf	48
a11240db GS	49	static u64 get_snooze_timeout(struct cpuidle_device *dev,
	50	struct cpuidle_driver *drv,
	51	int index)
	52	{
	53	int i;
	54
	55	if (unlikely(!snooze_timeout_en))
	56	return default_snooze_timeout;
	57
	58	for (i = index + 1; i < drv->state_count; i++) {
	59	struct cpuidle_state *s = &drv->states[i];
	60	struct cpuidle_state_usage *su = &dev->states_usage[i];
	61
	62	if (s->disabled \|\| su->disable)
	63	continue;
	64
	65	return s->target_residency * tb_ticks_per_usec;
	66	}
	67
	68	return default_snooze_timeout;
	69	}
	70
2c2e6ecf DD	71	static int snooze_loop(struct cpuidle_device *dev,
	72	struct cpuidle_driver *drv,
	73	int index)
	74	{
78eaa10f SB	75	u64 snooze_exit_time;
78eaa10f SB	76
2c2e6ecf DD	77	set_thread_flag(TIF_POLLING_NRFLAG);
2c2e6ecf DD	78
3fc5ee92 NP	79	local_irq_enable();
3fc5ee92 NP	80
a11240db	81	snooze_exit_time = get_tb() + get_snooze_timeout(dev, drv, index);
591ac0cb	82	ppc64_runlatch_off();
26eb48a9	83	HMT_very_low();
2c2e6ecf	84	while (!need_resched()) {
7ded4291 NP	85	if (likely(snooze_timeout_en) && get_tb() > snooze_exit_time) {
	86	/*
	87	* Task has not woken up but we are exiting the polling
	88	* loop anyway. Require a barrier after polling is
	89	* cleared to order subsequent test of need_resched().
	90	*/
	91	clear_thread_flag(TIF_POLLING_NRFLAG);
	92	smp_mb();
78eaa10f	93	break;
7ded4291	94	}
2c2e6ecf DD	95	}
	96
	97	HMT_medium();
591ac0cb	98	ppc64_runlatch_on();
2c2e6ecf	99	clear_thread_flag(TIF_POLLING_NRFLAG);
3fc5ee92	100
2c2e6ecf DD	101	return index;
	102	}
	103
	104	static int nap_loop(struct cpuidle_device *dev,
	105	struct cpuidle_driver *drv,
	106	int index)
	107	{
2201f994 NP	108	power7_idle_type(PNV_THREAD_NAP);
2201f994 NP	109
2c2e6ecf DD	110	return index;
	111	}
	112
cc5a2f7b	113	/* Register for fastsleep only in oneshot mode of broadcast */
cc5a2f7b	114	#ifdef CONFIG_TICK_ONESHOT
0d948730 PM	115	static int fastsleep_loop(struct cpuidle_device *dev,
	116	struct cpuidle_driver *drv,
	117	int index)
	118	{
	119	unsigned long old_lpcr = mfspr(SPRN_LPCR);
	120	unsigned long new_lpcr;
	121
	122	if (unlikely(system_state < SYSTEM_RUNNING))
	123	return index;
	124
	125	new_lpcr = old_lpcr;
9b6a68d9 MN	126	/* Do not exit powersave upon decrementer as we've setup the timer
9b6a68d9 MN	127	* offload.
0d948730	128	*/
9b6a68d9	129	new_lpcr &= ~LPCR_PECE1;
0d948730 PM	130
0d948730 PM	131	mtspr(SPRN_LPCR, new_lpcr);
2201f994 NP	132
2201f994 NP	133	power7_idle_type(PNV_THREAD_SLEEP);
0d948730 PM	134
	135	mtspr(SPRN_LPCR, old_lpcr);
	136
	137	return index;
	138	}
cc5a2f7b	139	#endif
3005c597 SP	140
	141	static int stop_loop(struct cpuidle_device *dev,
	142	struct cpuidle_driver *drv,
	143	int index)
	144	{
2201f994	145	power9_idle_type(stop_psscr_table[index].val,
09206b60	146	stop_psscr_table[index].mask);
3005c597 SP	147	return index;
	148	}
	149
2c2e6ecf DD	150	/*
	151	* States for dedicated partition case.
	152	*/
169f3fae	153	static struct cpuidle_state powernv_states[CPUIDLE_STATE_MAX] = {
2c2e6ecf DD	154	{ /* Snooze */
	155	.name = "snooze",
	156	.desc = "snooze",
2c2e6ecf DD	157	.exit_latency = 0,
2c2e6ecf DD	158	.target_residency = 0,
957efced	159	.enter = snooze_loop },
2c2e6ecf DD	160	};
2c2e6ecf DD	161
10fcca9d	162	static int powernv_cpuidle_cpu_online(unsigned int cpu)
2c2e6ecf	163	{
10fcca9d	164	struct cpuidle_device *dev = per_cpu(cpuidle_devices, cpu);
2c2e6ecf DD	165
2c2e6ecf DD	166	if (dev && cpuidle_get_driver()) {
10fcca9d SAS	167	cpuidle_pause_and_lock();
	168	cpuidle_enable_device(dev);
	169	cpuidle_resume_and_unlock();
	170	}
	171	return 0;
	172	}
2c2e6ecf	173
10fcca9d SAS	174	static int powernv_cpuidle_cpu_dead(unsigned int cpu)
	175	{
	176	struct cpuidle_device *dev = per_cpu(cpuidle_devices, cpu);
2c2e6ecf	177
10fcca9d SAS	178	if (dev && cpuidle_get_driver()) {
	179	cpuidle_pause_and_lock();
	180	cpuidle_disable_device(dev);
	181	cpuidle_resume_and_unlock();
2c2e6ecf	182	}
10fcca9d	183	return 0;
2c2e6ecf DD	184	}
2c2e6ecf DD	185
2c2e6ecf DD	186	/*
	187	* powernv_cpuidle_driver_init()
	188	*/
	189	static int powernv_cpuidle_driver_init(void)
	190	{
	191	int idle_state;
	192	struct cpuidle_driver *drv = &powernv_idle_driver;
	193
	194	drv->state_count = 0;
	195
	196	for (idle_state = 0; idle_state < max_idle_state; ++idle_state) {
	197	/* Is the state not enabled? */
	198	if (cpuidle_state_table[idle_state].enter == NULL)
	199	continue;
	200
	201	drv->states[drv->state_count] = /* structure copy */
	202	cpuidle_state_table[idle_state];
	203
	204	drv->state_count += 1;
	205	}
	206
293d264f VS	207	/*
	208	* On the PowerNV platform cpu_present may be less than cpu_possible in
	209	* cases when firmware detects the CPU, but it is not available to the
	210	* OS. If CONFIG_HOTPLUG_CPU=n, then such CPUs are not hotplugable at
	211	* run time and hence cpu_devices are not created for those CPUs by the
	212	* generic topology_init().
	213	*
	214	* drv->cpumask defaults to cpu_possible_mask in
	215	* __cpuidle_driver_init(). This breaks cpuidle on PowerNV where
	216	* cpu_devices are not created for CPUs in cpu_possible_mask that
	217	* cannot be hot-added later at run time.
	218	*
	219	* Trying cpuidle_register_device() on a CPU without a cpu_device is
	220	* incorrect, so pass a correct CPU mask to the generic cpuidle driver.
	221	*/
	222
	223	drv->cpumask = (struct cpumask *)cpu_present_mask;
	224
2c2e6ecf DD	225	return 0;
	226	}
	227
9e9fc6f0 GS	228	static inline void add_powernv_state(int index, const char *name,
	229	unsigned int flags,
	230	int (idle_fn)(struct cpuidle_device ,
	231	struct cpuidle_driver *,
	232	int),
	233	unsigned int target_residency,
	234	unsigned int exit_latency,
09206b60	235	u64 psscr_val, u64 psscr_mask)
9e9fc6f0 GS	236	{
	237	strlcpy(powernv_states[index].name, name, CPUIDLE_NAME_LEN);
	238	strlcpy(powernv_states[index].desc, name, CPUIDLE_NAME_LEN);
	239	powernv_states[index].flags = flags;
	240	powernv_states[index].target_residency = target_residency;
	241	powernv_states[index].exit_latency = exit_latency;
	242	powernv_states[index].enter = idle_fn;
09206b60 GS	243	stop_psscr_table[index].val = psscr_val;
09206b60 GS	244	stop_psscr_table[index].mask = psscr_mask;
9e9fc6f0 GS	245	}
9e9fc6f0 GS	246
ecad4502 GS	247	/*
	248	* Returns 0 if prop1_len == prop2_len. Else returns -1
	249	*/
	250	static inline int validate_dt_prop_sizes(const char *prop1, int prop1_len,
	251	const char *prop2, int prop2_len)
	252	{
	253	if (prop1_len == prop2_len)
	254	return 0;
	255
	256	pr_warn("cpuidle-powernv: array sizes don't match for %s and %s\n",
	257	prop1, prop2);
	258	return -1;
	259	}
	260
785a12af	261	extern u32 pnv_get_supported_cpuidle_states(void);
0888839c PM	262	static int powernv_add_idle_states(void)
	263	{
	264	struct device_node *power_mgt;
0888839c	265	int nr_idle_states = 1; /* Snooze */
ecad4502	266	int dt_idle_states, count;
957efced SP	267	u32 latency_ns[CPUIDLE_STATE_MAX];
	268	u32 residency_ns[CPUIDLE_STATE_MAX];
	269	u32 flags[CPUIDLE_STATE_MAX];
3005c597	270	u64 psscr_val[CPUIDLE_STATE_MAX];
09206b60	271	u64 psscr_mask[CPUIDLE_STATE_MAX];
3005c597	272	const char *names[CPUIDLE_STATE_MAX];
09206b60	273	u32 has_stop_states = 0;
92c83ff5	274	int i, rc;
785a12af GS	275	u32 supported_flags = pnv_get_supported_cpuidle_states();
785a12af GS	276
0888839c PM	277
	278	/* Currently we have snooze statically defined */
	279
	280	power_mgt = of_find_node_by_path("/ibm,opal/power-mgt");
	281	if (!power_mgt) {
	282	pr_warn("opal: PowerMgmt Node not found\n");
92c83ff5	283	goto out;
0888839c PM	284	}
0888839c PM	285
70734a78 PM	286	/* Read values of any property to determine the num of idle states */
	287	dt_idle_states = of_property_count_u32_elems(power_mgt, "ibm,cpu-idle-state-flags");
	288	if (dt_idle_states < 0) {
	289	pr_warn("cpuidle-powernv: no idle states found in the DT\n");
92c83ff5	290	goto out;
0888839c PM	291	}
0888839c PM	292
ecad4502 GS	293	count = of_property_count_u32_elems(power_mgt,
	294	"ibm,cpu-idle-state-latencies-ns");
	295
	296	if (validate_dt_prop_sizes("ibm,cpu-idle-state-flags", dt_idle_states,
	297	"ibm,cpu-idle-state-latencies-ns",
	298	count) != 0)
	299	goto out;
	300
	301	count = of_property_count_strings(power_mgt,
	302	"ibm,cpu-idle-state-names");
	303	if (validate_dt_prop_sizes("ibm,cpu-idle-state-flags", dt_idle_states,
	304	"ibm,cpu-idle-state-names",
	305	count) != 0)
	306	goto out;
	307
957efced SP	308	/*
	309	* Since snooze is used as first idle state, max idle states allowed is
	310	* CPUIDLE_STATE_MAX -1
	311	*/
	312	if (dt_idle_states > CPUIDLE_STATE_MAX - 1) {
	313	pr_warn("cpuidle-powernv: discovered idle states more than allowed");
	314	dt_idle_states = CPUIDLE_STATE_MAX - 1;
	315	}
	316
70734a78 PM	317	if (of_property_read_u32_array(power_mgt,
	318	"ibm,cpu-idle-state-flags", flags, dt_idle_states)) {
	319	pr_warn("cpuidle-powernv : missing ibm,cpu-idle-state-flags in DT\n");
957efced	320	goto out;
70734a78	321	}
92c83ff5	322
957efced SP	323	if (of_property_read_u32_array(power_mgt,
	324	"ibm,cpu-idle-state-latencies-ns", latency_ns,
	325	dt_idle_states)) {
92c83ff5	326	pr_warn("cpuidle-powernv: missing ibm,cpu-idle-state-latencies-ns in DT\n");
957efced	327	goto out;
74aa51b5	328	}
3005c597 SP	329	if (of_property_read_string_array(power_mgt,
	330	"ibm,cpu-idle-state-names", names, dt_idle_states) < 0) {
	331	pr_warn("cpuidle-powernv: missing ibm,cpu-idle-state-names in DT\n");
	332	goto out;
	333	}
	334
	335	/*
	336	* If the idle states use stop instruction, probe for psscr values
09206b60	337	* and psscr mask which are necessary to specify required stop level.
3005c597	338	*/
09206b60 GS	339	has_stop_states = (flags[0] &
	340	(OPAL_PM_STOP_INST_FAST \| OPAL_PM_STOP_INST_DEEP));
	341	if (has_stop_states) {
ecad4502 GS	342	count = of_property_count_u64_elems(power_mgt,
	343	"ibm,cpu-idle-state-psscr");
	344	if (validate_dt_prop_sizes("ibm,cpu-idle-state-flags",
	345	dt_idle_states,
	346	"ibm,cpu-idle-state-psscr",
	347	count) != 0)
	348	goto out;
	349
	350	count = of_property_count_u64_elems(power_mgt,
	351	"ibm,cpu-idle-state-psscr-mask");
	352	if (validate_dt_prop_sizes("ibm,cpu-idle-state-flags",
	353	dt_idle_states,
	354	"ibm,cpu-idle-state-psscr-mask",
	355	count) != 0)
	356	goto out;
	357
3005c597 SP	358	if (of_property_read_u64_array(power_mgt,
3005c597 SP	359	"ibm,cpu-idle-state-psscr", psscr_val, dt_idle_states)) {
09206b60	360	pr_warn("cpuidle-powernv: missing ibm,cpu-idle-state-psscr in DT\n");
3005c597 SP	361	goto out;
3005c597 SP	362	}
74aa51b5	363
09206b60 GS	364	if (of_property_read_u64_array(power_mgt,
	365	"ibm,cpu-idle-state-psscr-mask",
	366	psscr_mask, dt_idle_states)) {
	367	pr_warn("cpuidle-powernv:Missing ibm,cpu-idle-state-psscr-mask in DT\n");
	368	goto out;
	369	}
	370	}
	371
ecad4502 GS	372	count = of_property_count_u32_elems(power_mgt,
	373	"ibm,cpu-idle-state-residency-ns");
	374
	375	if (count < 0) {
	376	rc = count;
	377	} else if (validate_dt_prop_sizes("ibm,cpu-idle-state-flags",
	378	dt_idle_states,
	379	"ibm,cpu-idle-state-residency-ns",
	380	count) != 0) {
	381	goto out;
	382	} else {
	383	rc = of_property_read_u32_array(power_mgt,
	384	"ibm,cpu-idle-state-residency-ns",
	385	residency_ns, dt_idle_states);
	386	}
0888839c PM	387
0888839c PM	388	for (i = 0; i < dt_idle_states; i++) {
9e9fc6f0	389	unsigned int exit_latency, target_residency;
f9122ee4	390	bool stops_timebase = false;
785a12af GS	391
	392	/*
	393	* Skip the platform idle state whose flag isn't in
	394	* the supported_cpuidle_states flag mask.
	395	*/
	396	if ((flags[i] & supported_flags) != flags[i])
	397	continue;
3005c597 SP	398	/*
	399	* If an idle state has exit latency beyond
	400	* POWERNV_THRESHOLD_LATENCY_NS then don't use it
	401	* in cpu-idle.
	402	*/
	403	if (latency_ns[i] > POWERNV_THRESHOLD_LATENCY_NS)
	404	continue;
9e9fc6f0 GS	405	/*
	406	* Firmware passes residency and latency values in ns.
	407	* cpuidle expects it in us.
	408	*/
8d4e10e9	409	exit_latency = DIV_ROUND_UP(latency_ns[i], 1000);
9e9fc6f0	410	if (!rc)
8d4e10e9	411	target_residency = DIV_ROUND_UP(residency_ns[i], 1000);
9e9fc6f0 GS	412	else
9e9fc6f0 GS	413	target_residency = 0;
0888839c	414
09206b60 GS	415	if (has_stop_states) {
	416	int err = validate_psscr_val_mask(&psscr_val[i],
	417	&psscr_mask[i],
	418	flags[i]);
	419	if (err) {
	420	report_invalid_psscr_val(psscr_val[i], err);
	421	continue;
	422	}
	423	}
	424
f9122ee4 GS	425	if (flags[i] & OPAL_PM_TIMEBASE_STOP)
	426	stops_timebase = true;
	427
92c83ff5	428	/*
9e9fc6f0 GS	429	* For nap and fastsleep, use default target_residency
9e9fc6f0 GS	430	* values if f/w does not expose it.
74aa51b5	431	*/
70734a78	432	if (flags[i] & OPAL_PM_NAP_ENABLED) {
9e9fc6f0 GS	433	if (!rc)
9e9fc6f0 GS	434	target_residency = 100;
0888839c	435	/* Add NAP state */
9e9fc6f0 GS	436	add_powernv_state(nr_idle_states, "Nap",
9e9fc6f0 GS	437	CPUIDLE_FLAG_NONE, nap_loop,
09206b60	438	target_residency, exit_latency, 0, 0);
f9122ee4	439	} else if (has_stop_states && !stops_timebase) {
9e9fc6f0 GS	440	add_powernv_state(nr_idle_states, names[i],
	441	CPUIDLE_FLAG_NONE, stop_loop,
	442	target_residency, exit_latency,
09206b60	443	psscr_val[i], psscr_mask[i]);
cc5a2f7b	444	}
	445
	446	/*
	447	* All cpuidle states with CPUIDLE_FLAG_TIMER_STOP set must come
	448	* within this config dependency check.
	449	*/
	450	#ifdef CONFIG_TICK_ONESHOT
f9122ee4 GS	451	else if (flags[i] & OPAL_PM_SLEEP_ENABLED \|\|
f9122ee4 GS	452	flags[i] & OPAL_PM_SLEEP_ENABLED_ER1) {
9e9fc6f0 GS	453	if (!rc)
9e9fc6f0 GS	454	target_residency = 300000;
0888839c	455	/* Add FASTSLEEP state */
9e9fc6f0 GS	456	add_powernv_state(nr_idle_states, "FastSleep",
	457	CPUIDLE_FLAG_TIMER_STOP,
	458	fastsleep_loop,
09206b60	459	target_residency, exit_latency, 0, 0);
f9122ee4	460	} else if (has_stop_states && stops_timebase) {
9e9fc6f0 GS	461	add_powernv_state(nr_idle_states, names[i],
	462	CPUIDLE_FLAG_TIMER_STOP, stop_loop,
	463	target_residency, exit_latency,
09206b60	464	psscr_val[i], psscr_mask[i]);
0888839c	465	}
cc5a2f7b	466	#endif
f9122ee4 GS	467	else
f9122ee4 GS	468	continue;
92c83ff5	469	nr_idle_states++;
0888839c	470	}
92c83ff5	471	out:
0888839c PM	472	return nr_idle_states;
	473	}
	474
2c2e6ecf DD	475	/*
	476	* powernv_idle_probe()
	477	* Choose state table for shared versus dedicated partition
	478	*/
	479	static int powernv_idle_probe(void)
	480	{
2c2e6ecf DD	481	if (cpuidle_disable != IDLE_NO_OVERRIDE)
	482	return -ENODEV;
	483
e4d54f71	484	if (firmware_has_feature(FW_FEATURE_OPAL)) {
2c2e6ecf	485	cpuidle_state_table = powernv_states;
0888839c PM	486	/* Device tree can indicate more idle states */
0888839c PM	487	max_idle_state = powernv_add_idle_states();
a11240db GS	488	default_snooze_timeout = TICK_USEC * tb_ticks_per_usec;
a11240db GS	489	if (max_idle_state > 1)
78eaa10f	490	snooze_timeout_en = true;
2c2e6ecf DD	491	} else
	492	return -ENODEV;
	493
	494	return 0;
	495	}
	496
	497	static int __init powernv_processor_idle_init(void)
	498	{
	499	int retval;
	500
	501	retval = powernv_idle_probe();
	502	if (retval)
	503	return retval;
	504
	505	powernv_cpuidle_driver_init();
	506	retval = cpuidle_register(&powernv_idle_driver, NULL);
	507	if (retval) {
	508	printk(KERN_DEBUG "Registration of powernv driver failed.\n");
	509	return retval;
	510	}
	511
10fcca9d SAS	512	retval = cpuhp_setup_state_nocalls(CPUHP_AP_ONLINE_DYN,
	513	"cpuidle/powernv:online",
	514	powernv_cpuidle_cpu_online, NULL);
	515	WARN_ON(retval < 0);
	516	retval = cpuhp_setup_state_nocalls(CPUHP_CPUIDLE_DEAD,
	517	"cpuidle/powernv:dead", NULL,
	518	powernv_cpuidle_cpu_dead);
	519	WARN_ON(retval < 0);
2c2e6ecf DD	520	printk(KERN_DEBUG "powernv_idle_driver registered\n");
	521	return 0;
	522	}
	523
	524	device_initcall(powernv_processor_idle_init);