[linux.git] / drivers / cpuidle / cpuidle-powernv.c

/*
 *  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;
static struct cpuidle_state *cpuidle_state_table;

struct stop_psscr_table {
	u64 val;
	u64 mask;
};

static struct stop_psscr_table stop_psscr_table[CPUIDLE_STATE_MAX];

static u64 snooze_timeout;
static bool snooze_timeout_en;

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

	local_irq_enable();
	set_thread_flag(TIF_POLLING_NRFLAG);

	snooze_exit_time = get_tb() + snooze_timeout;
	ppc64_runlatch_off();
	HMT_very_low();
	while (!need_resched()) {
		if (likely(snooze_timeout_en) && get_tb() > snooze_exit_time)
			break;
	}

	HMT_medium();
	ppc64_runlatch_on();
	clear_thread_flag(TIF_POLLING_NRFLAG);
	smp_mb();
	return index;
}

static int nap_loop(struct cpuidle_device *dev,
			struct cpuidle_driver *drv,
			int index)
{
	ppc64_runlatch_off();
	power7_idle();
	ppc64_runlatch_on();
	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_sleep();

	mtspr(SPRN_LPCR, old_lpcr);

	return index;
}
#endif

static int stop_loop(struct cpuidle_device *dev,
		     struct cpuidle_driver *drv,
		     int index)
{
	ppc64_runlatch_off();
	power9_idle_stop(stop_psscr_table[index].val,
			 stop_psscr_table[index].mask);
	ppc64_runlatch_on();
	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;
}

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;

	/* 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;
		/*
		 * 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 = latency_ns[i] / 1000;
		if (!rc)
			target_residency = 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;
			}
		}

		/*
		 * 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 ((flags[i] & OPAL_PM_STOP_INST_FAST) &&
				!(flags[i] & OPAL_PM_TIMEBASE_STOP)) {
			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
		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 ((flags[i] & OPAL_PM_STOP_INST_DEEP) &&
				(flags[i] & OPAL_PM_TIMEBASE_STOP)) {
			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
		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();
		if (max_idle_state > 1) {
			snooze_timeout_en = true;
			snooze_timeout = powernv_states[1].target_residency *
					 tb_ticks_per_usec;
		}
 	} 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
2c2e6ecf DD	1	/*
	2	* cpuidle-powernv - idle state cpuidle driver.
	3	* Adapted from drivers/cpuidle/cpuidle-pseries
	4	*
	5	*/
	6
	7	#include <linux/kernel.h>
	8	#include <linux/module.h>
	9	#include <linux/init.h>
	10	#include <linux/moduleparam.h>
	11	#include <linux/cpuidle.h>
	12	#include <linux/cpu.h>
	13	#include <linux/notifier.h>
0d948730	14	#include <linux/clockchips.h>
0888839c	15	#include <linux/of.h>
92c83ff5	16	#include <linux/slab.h>
2c2e6ecf DD	17
	18	#include <asm/machdep.h>
	19	#include <asm/firmware.h>
8eb8ac89	20	#include <asm/opal.h>
591ac0cb	21	#include <asm/runlatch.h>
09206b60	22	#include <asm/cpuidle.h>
2c2e6ecf	23
9e9fc6f0 GS	24	/*
	25	* Expose only those Hardware idle states via the cpuidle framework
	26	* that have latency value below POWERNV_THRESHOLD_LATENCY_NS.
	27	*/
3005c597 SP	28	#define POWERNV_THRESHOLD_LATENCY_NS 200000
3005c597 SP	29
ed61390b	30	static struct cpuidle_driver powernv_idle_driver = {
2c2e6ecf DD	31	.name = "powernv_idle",
	32	.owner = THIS_MODULE,
	33	};
	34
	35	static int max_idle_state;
	36	static struct cpuidle_state *cpuidle_state_table;
3005c597	37
09206b60 GS	38	struct stop_psscr_table {
	39	u64 val;
	40	u64 mask;
	41	};
	42
	43	static struct stop_psscr_table stop_psscr_table[CPUIDLE_STATE_MAX];
3005c597	44
78eaa10f SB	45	static u64 snooze_timeout;
78eaa10f SB	46	static bool snooze_timeout_en;
2c2e6ecf DD	47
	48	static int snooze_loop(struct cpuidle_device *dev,
	49	struct cpuidle_driver *drv,
	50	int index)
	51	{
78eaa10f SB	52	u64 snooze_exit_time;
78eaa10f SB	53
2c2e6ecf DD	54	local_irq_enable();
	55	set_thread_flag(TIF_POLLING_NRFLAG);
	56
78eaa10f	57	snooze_exit_time = get_tb() + snooze_timeout;
591ac0cb	58	ppc64_runlatch_off();
26eb48a9	59	HMT_very_low();
2c2e6ecf	60	while (!need_resched()) {
0baa91cb	61	if (likely(snooze_timeout_en) && get_tb() > snooze_exit_time)
78eaa10f	62	break;
2c2e6ecf DD	63	}
	64
	65	HMT_medium();
591ac0cb	66	ppc64_runlatch_on();
2c2e6ecf DD	67	clear_thread_flag(TIF_POLLING_NRFLAG);
	68	smp_mb();
	69	return index;
	70	}
	71
	72	static int nap_loop(struct cpuidle_device *dev,
	73	struct cpuidle_driver *drv,
	74	int index)
	75	{
591ac0cb	76	ppc64_runlatch_off();
2c2e6ecf	77	power7_idle();
591ac0cb	78	ppc64_runlatch_on();
2c2e6ecf DD	79	return index;
	80	}
	81
cc5a2f7b	82	/* Register for fastsleep only in oneshot mode of broadcast */
cc5a2f7b	83	#ifdef CONFIG_TICK_ONESHOT
0d948730 PM	84	static int fastsleep_loop(struct cpuidle_device *dev,
	85	struct cpuidle_driver *drv,
	86	int index)
	87	{
	88	unsigned long old_lpcr = mfspr(SPRN_LPCR);
	89	unsigned long new_lpcr;
	90
	91	if (unlikely(system_state < SYSTEM_RUNNING))
	92	return index;
	93
	94	new_lpcr = old_lpcr;
9b6a68d9 MN	95	/* Do not exit powersave upon decrementer as we've setup the timer
9b6a68d9 MN	96	* offload.
0d948730	97	*/
9b6a68d9	98	new_lpcr &= ~LPCR_PECE1;
0d948730 PM	99
	100	mtspr(SPRN_LPCR, new_lpcr);
	101	power7_sleep();
	102
	103	mtspr(SPRN_LPCR, old_lpcr);
	104
	105	return index;
	106	}
cc5a2f7b	107	#endif
3005c597 SP	108
	109	static int stop_loop(struct cpuidle_device *dev,
	110	struct cpuidle_driver *drv,
	111	int index)
	112	{
	113	ppc64_runlatch_off();
09206b60 GS	114	power9_idle_stop(stop_psscr_table[index].val,
09206b60 GS	115	stop_psscr_table[index].mask);
3005c597 SP	116	ppc64_runlatch_on();
	117	return index;
	118	}
	119
2c2e6ecf DD	120	/*
	121	* States for dedicated partition case.
	122	*/
169f3fae	123	static struct cpuidle_state powernv_states[CPUIDLE_STATE_MAX] = {
2c2e6ecf DD	124	{ /* Snooze */
	125	.name = "snooze",
	126	.desc = "snooze",
2c2e6ecf DD	127	.exit_latency = 0,
2c2e6ecf DD	128	.target_residency = 0,
957efced	129	.enter = snooze_loop },
2c2e6ecf DD	130	};
2c2e6ecf DD	131
10fcca9d	132	static int powernv_cpuidle_cpu_online(unsigned int cpu)
2c2e6ecf	133	{
10fcca9d	134	struct cpuidle_device *dev = per_cpu(cpuidle_devices, cpu);
2c2e6ecf DD	135
2c2e6ecf DD	136	if (dev && cpuidle_get_driver()) {
10fcca9d SAS	137	cpuidle_pause_and_lock();
	138	cpuidle_enable_device(dev);
	139	cpuidle_resume_and_unlock();
	140	}
	141	return 0;
	142	}
2c2e6ecf	143
10fcca9d SAS	144	static int powernv_cpuidle_cpu_dead(unsigned int cpu)
	145	{
	146	struct cpuidle_device *dev = per_cpu(cpuidle_devices, cpu);
2c2e6ecf	147
10fcca9d SAS	148	if (dev && cpuidle_get_driver()) {
	149	cpuidle_pause_and_lock();
	150	cpuidle_disable_device(dev);
	151	cpuidle_resume_and_unlock();
2c2e6ecf	152	}
10fcca9d	153	return 0;
2c2e6ecf DD	154	}
2c2e6ecf DD	155
2c2e6ecf DD	156	/*
	157	* powernv_cpuidle_driver_init()
	158	*/
	159	static int powernv_cpuidle_driver_init(void)
	160	{
	161	int idle_state;
	162	struct cpuidle_driver *drv = &powernv_idle_driver;
	163
	164	drv->state_count = 0;
	165
	166	for (idle_state = 0; idle_state < max_idle_state; ++idle_state) {
	167	/* Is the state not enabled? */
	168	if (cpuidle_state_table[idle_state].enter == NULL)
	169	continue;
	170
	171	drv->states[drv->state_count] = /* structure copy */
	172	cpuidle_state_table[idle_state];
	173
	174	drv->state_count += 1;
	175	}
	176
293d264f VS	177	/*
	178	* On the PowerNV platform cpu_present may be less than cpu_possible in
	179	* cases when firmware detects the CPU, but it is not available to the
	180	* OS. If CONFIG_HOTPLUG_CPU=n, then such CPUs are not hotplugable at
	181	* run time and hence cpu_devices are not created for those CPUs by the
	182	* generic topology_init().
	183	*
	184	* drv->cpumask defaults to cpu_possible_mask in
	185	* __cpuidle_driver_init(). This breaks cpuidle on PowerNV where
	186	* cpu_devices are not created for CPUs in cpu_possible_mask that
	187	* cannot be hot-added later at run time.
	188	*
	189	* Trying cpuidle_register_device() on a CPU without a cpu_device is
	190	* incorrect, so pass a correct CPU mask to the generic cpuidle driver.
	191	*/
	192
	193	drv->cpumask = (struct cpumask *)cpu_present_mask;
	194
2c2e6ecf DD	195	return 0;
	196	}
	197
9e9fc6f0 GS	198	static inline void add_powernv_state(int index, const char *name,
	199	unsigned int flags,
	200	int (idle_fn)(struct cpuidle_device ,
	201	struct cpuidle_driver *,
	202	int),
	203	unsigned int target_residency,
	204	unsigned int exit_latency,
09206b60	205	u64 psscr_val, u64 psscr_mask)
9e9fc6f0 GS	206	{
	207	strlcpy(powernv_states[index].name, name, CPUIDLE_NAME_LEN);
	208	strlcpy(powernv_states[index].desc, name, CPUIDLE_NAME_LEN);
	209	powernv_states[index].flags = flags;
	210	powernv_states[index].target_residency = target_residency;
	211	powernv_states[index].exit_latency = exit_latency;
	212	powernv_states[index].enter = idle_fn;
09206b60 GS	213	stop_psscr_table[index].val = psscr_val;
09206b60 GS	214	stop_psscr_table[index].mask = psscr_mask;
9e9fc6f0 GS	215	}
9e9fc6f0 GS	216
ecad4502 GS	217	/*
	218	* Returns 0 if prop1_len == prop2_len. Else returns -1
	219	*/
	220	static inline int validate_dt_prop_sizes(const char *prop1, int prop1_len,
	221	const char *prop2, int prop2_len)
	222	{
	223	if (prop1_len == prop2_len)
	224	return 0;
	225
	226	pr_warn("cpuidle-powernv: array sizes don't match for %s and %s\n",
	227	prop1, prop2);
	228	return -1;
	229	}
	230
0888839c PM	231	static int powernv_add_idle_states(void)
	232	{
	233	struct device_node *power_mgt;
0888839c	234	int nr_idle_states = 1; /* Snooze */
ecad4502	235	int dt_idle_states, count;
957efced SP	236	u32 latency_ns[CPUIDLE_STATE_MAX];
	237	u32 residency_ns[CPUIDLE_STATE_MAX];
	238	u32 flags[CPUIDLE_STATE_MAX];
3005c597	239	u64 psscr_val[CPUIDLE_STATE_MAX];
09206b60	240	u64 psscr_mask[CPUIDLE_STATE_MAX];
3005c597	241	const char *names[CPUIDLE_STATE_MAX];
09206b60	242	u32 has_stop_states = 0;
92c83ff5	243	int i, rc;
0888839c PM	244
	245	/* Currently we have snooze statically defined */
	246
	247	power_mgt = of_find_node_by_path("/ibm,opal/power-mgt");
	248	if (!power_mgt) {
	249	pr_warn("opal: PowerMgmt Node not found\n");
92c83ff5	250	goto out;
0888839c PM	251	}
0888839c PM	252
70734a78 PM	253	/* Read values of any property to determine the num of idle states */
	254	dt_idle_states = of_property_count_u32_elems(power_mgt, "ibm,cpu-idle-state-flags");
	255	if (dt_idle_states < 0) {
	256	pr_warn("cpuidle-powernv: no idle states found in the DT\n");
92c83ff5	257	goto out;
0888839c PM	258	}
0888839c PM	259
ecad4502 GS	260	count = of_property_count_u32_elems(power_mgt,
	261	"ibm,cpu-idle-state-latencies-ns");
	262
	263	if (validate_dt_prop_sizes("ibm,cpu-idle-state-flags", dt_idle_states,
	264	"ibm,cpu-idle-state-latencies-ns",
	265	count) != 0)
	266	goto out;
	267
	268	count = of_property_count_strings(power_mgt,
	269	"ibm,cpu-idle-state-names");
	270	if (validate_dt_prop_sizes("ibm,cpu-idle-state-flags", dt_idle_states,
	271	"ibm,cpu-idle-state-names",
	272	count) != 0)
	273	goto out;
	274
957efced SP	275	/*
	276	* Since snooze is used as first idle state, max idle states allowed is
	277	* CPUIDLE_STATE_MAX -1
	278	*/
	279	if (dt_idle_states > CPUIDLE_STATE_MAX - 1) {
	280	pr_warn("cpuidle-powernv: discovered idle states more than allowed");
	281	dt_idle_states = CPUIDLE_STATE_MAX - 1;
	282	}
	283
70734a78 PM	284	if (of_property_read_u32_array(power_mgt,
	285	"ibm,cpu-idle-state-flags", flags, dt_idle_states)) {
	286	pr_warn("cpuidle-powernv : missing ibm,cpu-idle-state-flags in DT\n");
957efced	287	goto out;
70734a78	288	}
92c83ff5	289
957efced SP	290	if (of_property_read_u32_array(power_mgt,
	291	"ibm,cpu-idle-state-latencies-ns", latency_ns,
	292	dt_idle_states)) {
92c83ff5	293	pr_warn("cpuidle-powernv: missing ibm,cpu-idle-state-latencies-ns in DT\n");
957efced	294	goto out;
74aa51b5	295	}
3005c597 SP	296	if (of_property_read_string_array(power_mgt,
	297	"ibm,cpu-idle-state-names", names, dt_idle_states) < 0) {
	298	pr_warn("cpuidle-powernv: missing ibm,cpu-idle-state-names in DT\n");
	299	goto out;
	300	}
	301
	302	/*
	303	* If the idle states use stop instruction, probe for psscr values
09206b60	304	* and psscr mask which are necessary to specify required stop level.
3005c597	305	*/
09206b60 GS	306	has_stop_states = (flags[0] &
	307	(OPAL_PM_STOP_INST_FAST \| OPAL_PM_STOP_INST_DEEP));
	308	if (has_stop_states) {
ecad4502 GS	309	count = of_property_count_u64_elems(power_mgt,
	310	"ibm,cpu-idle-state-psscr");
	311	if (validate_dt_prop_sizes("ibm,cpu-idle-state-flags",
	312	dt_idle_states,
	313	"ibm,cpu-idle-state-psscr",
	314	count) != 0)
	315	goto out;
	316
	317	count = of_property_count_u64_elems(power_mgt,
	318	"ibm,cpu-idle-state-psscr-mask");
	319	if (validate_dt_prop_sizes("ibm,cpu-idle-state-flags",
	320	dt_idle_states,
	321	"ibm,cpu-idle-state-psscr-mask",
	322	count) != 0)
	323	goto out;
	324
3005c597 SP	325	if (of_property_read_u64_array(power_mgt,
3005c597 SP	326	"ibm,cpu-idle-state-psscr", psscr_val, dt_idle_states)) {
09206b60	327	pr_warn("cpuidle-powernv: missing ibm,cpu-idle-state-psscr in DT\n");
3005c597 SP	328	goto out;
3005c597 SP	329	}
74aa51b5	330
09206b60 GS	331	if (of_property_read_u64_array(power_mgt,
	332	"ibm,cpu-idle-state-psscr-mask",
	333	psscr_mask, dt_idle_states)) {
	334	pr_warn("cpuidle-powernv:Missing ibm,cpu-idle-state-psscr-mask in DT\n");
	335	goto out;
	336	}
	337	}
	338
ecad4502 GS	339	count = of_property_count_u32_elems(power_mgt,
	340	"ibm,cpu-idle-state-residency-ns");
	341
	342	if (count < 0) {
	343	rc = count;
	344	} else if (validate_dt_prop_sizes("ibm,cpu-idle-state-flags",
	345	dt_idle_states,
	346	"ibm,cpu-idle-state-residency-ns",
	347	count) != 0) {
	348	goto out;
	349	} else {
	350	rc = of_property_read_u32_array(power_mgt,
	351	"ibm,cpu-idle-state-residency-ns",
	352	residency_ns, dt_idle_states);
	353	}
0888839c PM	354
0888839c PM	355	for (i = 0; i < dt_idle_states; i++) {
9e9fc6f0	356	unsigned int exit_latency, target_residency;
3005c597 SP	357	/*
	358	* If an idle state has exit latency beyond
	359	* POWERNV_THRESHOLD_LATENCY_NS then don't use it
	360	* in cpu-idle.
	361	*/
	362	if (latency_ns[i] > POWERNV_THRESHOLD_LATENCY_NS)
	363	continue;
9e9fc6f0 GS	364	/*
	365	* Firmware passes residency and latency values in ns.
	366	* cpuidle expects it in us.
	367	*/
	368	exit_latency = latency_ns[i] / 1000;
	369	if (!rc)
	370	target_residency = residency_ns[i] / 1000;
	371	else
	372	target_residency = 0;
0888839c	373
09206b60 GS	374	if (has_stop_states) {
	375	int err = validate_psscr_val_mask(&psscr_val[i],
	376	&psscr_mask[i],
	377	flags[i]);
	378	if (err) {
	379	report_invalid_psscr_val(psscr_val[i], err);
	380	continue;
	381	}
	382	}
	383
92c83ff5	384	/*
9e9fc6f0 GS	385	* For nap and fastsleep, use default target_residency
9e9fc6f0 GS	386	* values if f/w does not expose it.
74aa51b5	387	*/
70734a78	388	if (flags[i] & OPAL_PM_NAP_ENABLED) {
9e9fc6f0 GS	389	if (!rc)
9e9fc6f0 GS	390	target_residency = 100;
0888839c	391	/* Add NAP state */
9e9fc6f0 GS	392	add_powernv_state(nr_idle_states, "Nap",
9e9fc6f0 GS	393	CPUIDLE_FLAG_NONE, nap_loop,
09206b60	394	target_residency, exit_latency, 0, 0);
3005c597 SP	395	} else if ((flags[i] & OPAL_PM_STOP_INST_FAST) &&
3005c597 SP	396	!(flags[i] & OPAL_PM_TIMEBASE_STOP)) {
9e9fc6f0 GS	397	add_powernv_state(nr_idle_states, names[i],
	398	CPUIDLE_FLAG_NONE, stop_loop,
	399	target_residency, exit_latency,
09206b60	400	psscr_val[i], psscr_mask[i]);
cc5a2f7b	401	}
	402
	403	/*
	404	* All cpuidle states with CPUIDLE_FLAG_TIMER_STOP set must come
	405	* within this config dependency check.
	406	*/
	407	#ifdef CONFIG_TICK_ONESHOT
	408	if (flags[i] & OPAL_PM_SLEEP_ENABLED \|\|
70734a78	409	flags[i] & OPAL_PM_SLEEP_ENABLED_ER1) {
9e9fc6f0 GS	410	if (!rc)
9e9fc6f0 GS	411	target_residency = 300000;
0888839c	412	/* Add FASTSLEEP state */
9e9fc6f0 GS	413	add_powernv_state(nr_idle_states, "FastSleep",
	414	CPUIDLE_FLAG_TIMER_STOP,
	415	fastsleep_loop,
09206b60	416	target_residency, exit_latency, 0, 0);
3005c597 SP	417	} else if ((flags[i] & OPAL_PM_STOP_INST_DEEP) &&
3005c597 SP	418	(flags[i] & OPAL_PM_TIMEBASE_STOP)) {
9e9fc6f0 GS	419	add_powernv_state(nr_idle_states, names[i],
	420	CPUIDLE_FLAG_TIMER_STOP, stop_loop,
	421	target_residency, exit_latency,
09206b60	422	psscr_val[i], psscr_mask[i]);
0888839c	423	}
cc5a2f7b	424	#endif
92c83ff5	425	nr_idle_states++;
0888839c	426	}
92c83ff5	427	out:
0888839c PM	428	return nr_idle_states;
	429	}
	430
2c2e6ecf DD	431	/*
	432	* powernv_idle_probe()
	433	* Choose state table for shared versus dedicated partition
	434	*/
	435	static int powernv_idle_probe(void)
	436	{
2c2e6ecf DD	437	if (cpuidle_disable != IDLE_NO_OVERRIDE)
	438	return -ENODEV;
	439
e4d54f71	440	if (firmware_has_feature(FW_FEATURE_OPAL)) {
2c2e6ecf	441	cpuidle_state_table = powernv_states;
0888839c PM	442	/* Device tree can indicate more idle states */
0888839c PM	443	max_idle_state = powernv_add_idle_states();
78eaa10f SB	444	if (max_idle_state > 1) {
	445	snooze_timeout_en = true;
	446	snooze_timeout = powernv_states[1].target_residency *
	447	tb_ticks_per_usec;
	448	}
2c2e6ecf DD	449	} else
	450	return -ENODEV;
	451
	452	return 0;
	453	}
	454
	455	static int __init powernv_processor_idle_init(void)
	456	{
	457	int retval;
	458
	459	retval = powernv_idle_probe();
	460	if (retval)
	461	return retval;
	462
	463	powernv_cpuidle_driver_init();
	464	retval = cpuidle_register(&powernv_idle_driver, NULL);
	465	if (retval) {
	466	printk(KERN_DEBUG "Registration of powernv driver failed.\n");
	467	return retval;
	468	}
	469
10fcca9d SAS	470	retval = cpuhp_setup_state_nocalls(CPUHP_AP_ONLINE_DYN,
	471	"cpuidle/powernv:online",
	472	powernv_cpuidle_cpu_online, NULL);
	473	WARN_ON(retval < 0);
	474	retval = cpuhp_setup_state_nocalls(CPUHP_CPUIDLE_DEAD,
	475	"cpuidle/powernv:dead", NULL,
	476	powernv_cpuidle_cpu_dead);
	477	WARN_ON(retval < 0);
2c2e6ecf DD	478	printk(KERN_DEBUG "powernv_idle_driver registered\n");
	479	return 0;
	480	}
	481
	482	device_initcall(powernv_processor_idle_init);