[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 __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 snooze_timeout __read_mostly;
static bool snooze_timeout_en __read_mostly;

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() + snooze_timeout;
	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 = 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;
			}
		}

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