[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)
{
	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;
}

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