cpufreq.c 42.0 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6
/*
 *  linux/drivers/cpufreq/cpufreq.c
 *
 *  Copyright (C) 2001 Russell King
 *            (C) 2002 - 2003 Dominik Brodowski <linux@brodo.de>
 *
7
 *  Oct 2005 - Ashok Raj <ashok.raj@intel.com>
8
 *	Added handling for CPU hotplug
9 10
 *  Feb 2006 - Jacob Shin <jacob.shin@amd.com>
 *	Fix handling for CPU hotplug -- affected CPUs
11
 *
L
Linus Torvalds 已提交
12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
 * This program is free software; you can redistribute it and/or modify
 * it under the terms of the GNU General Public License version 2 as
 * published by the Free Software Foundation.
 *
 */

#include <linux/config.h>
#include <linux/kernel.h>
#include <linux/module.h>
#include <linux/init.h>
#include <linux/notifier.h>
#include <linux/cpufreq.h>
#include <linux/delay.h>
#include <linux/interrupt.h>
#include <linux/spinlock.h>
#include <linux/device.h>
#include <linux/slab.h>
#include <linux/cpu.h>
#include <linux/completion.h>
31
#include <linux/mutex.h>
L
Linus Torvalds 已提交
32 33 34 35 36 37 38 39

#define dprintk(msg...) cpufreq_debug_printk(CPUFREQ_DEBUG_CORE, "cpufreq-core", msg)

/**
 * The "cpufreq driver" - the arch- or hardware-dependend low
 * level driver of CPUFreq support, and its spinlock. This lock
 * also protects the cpufreq_cpu_data array.
 */
40 41
static struct cpufreq_driver *cpufreq_driver;
static struct cpufreq_policy *cpufreq_cpu_data[NR_CPUS];
L
Linus Torvalds 已提交
42 43 44 45 46 47 48
static DEFINE_SPINLOCK(cpufreq_driver_lock);

/* internal prototypes */
static int __cpufreq_governor(struct cpufreq_policy *policy, unsigned int event);
static void handle_update(void *data);

/**
49 50
 * Two notifier lists: the "policy" list is involved in the
 * validation process for a new CPU frequency policy; the
L
Linus Torvalds 已提交
51 52 53 54
 * "transition" list for kernel code that needs to handle
 * changes to devices when the CPU clock speed changes.
 * The mutex locks both lists.
 */
55 56
static BLOCKING_NOTIFIER_HEAD(cpufreq_policy_notifier_list);
static BLOCKING_NOTIFIER_HEAD(cpufreq_transition_notifier_list);
L
Linus Torvalds 已提交
57 58 59


static LIST_HEAD(cpufreq_governor_list);
60
static DEFINE_MUTEX (cpufreq_governor_mutex);
L
Linus Torvalds 已提交
61

62
struct cpufreq_policy *cpufreq_cpu_get(unsigned int cpu)
L
Linus Torvalds 已提交
63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91
{
	struct cpufreq_policy *data;
	unsigned long flags;

	if (cpu >= NR_CPUS)
		goto err_out;

	/* get the cpufreq driver */
	spin_lock_irqsave(&cpufreq_driver_lock, flags);

	if (!cpufreq_driver)
		goto err_out_unlock;

	if (!try_module_get(cpufreq_driver->owner))
		goto err_out_unlock;


	/* get the CPU */
	data = cpufreq_cpu_data[cpu];

	if (!data)
		goto err_out_put_module;

	if (!kobject_get(&data->kobj))
		goto err_out_put_module;

	spin_unlock_irqrestore(&cpufreq_driver_lock, flags);
	return data;

92
err_out_put_module:
L
Linus Torvalds 已提交
93
	module_put(cpufreq_driver->owner);
94
err_out_unlock:
L
Linus Torvalds 已提交
95
	spin_unlock_irqrestore(&cpufreq_driver_lock, flags);
96
err_out:
L
Linus Torvalds 已提交
97 98 99 100
	return NULL;
}
EXPORT_SYMBOL_GPL(cpufreq_cpu_get);

101

L
Linus Torvalds 已提交
102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129
void cpufreq_cpu_put(struct cpufreq_policy *data)
{
	kobject_put(&data->kobj);
	module_put(cpufreq_driver->owner);
}
EXPORT_SYMBOL_GPL(cpufreq_cpu_put);


/*********************************************************************
 *                     UNIFIED DEBUG HELPERS                         *
 *********************************************************************/
#ifdef CONFIG_CPU_FREQ_DEBUG

/* what part(s) of the CPUfreq subsystem are debugged? */
static unsigned int debug;

/* is the debug output ratelimit'ed using printk_ratelimit? User can
 * set or modify this value.
 */
static unsigned int debug_ratelimit = 1;

/* is the printk_ratelimit'ing enabled? It's enabled after a successful
 * loading of a cpufreq driver, temporarily disabled when a new policy
 * is set, and disabled upon cpufreq driver removal
 */
static unsigned int disable_ratelimit = 1;
static DEFINE_SPINLOCK(disable_ratelimit_lock);

130
static void cpufreq_debug_enable_ratelimit(void)
L
Linus Torvalds 已提交
131 132 133 134 135 136 137 138 139
{
	unsigned long flags;

	spin_lock_irqsave(&disable_ratelimit_lock, flags);
	if (disable_ratelimit)
		disable_ratelimit--;
	spin_unlock_irqrestore(&disable_ratelimit_lock, flags);
}

140
static void cpufreq_debug_disable_ratelimit(void)
L
Linus Torvalds 已提交
141 142 143 144 145 146 147 148 149 150 151 152 153 154
{
	unsigned long flags;

	spin_lock_irqsave(&disable_ratelimit_lock, flags);
	disable_ratelimit++;
	spin_unlock_irqrestore(&disable_ratelimit_lock, flags);
}

void cpufreq_debug_printk(unsigned int type, const char *prefix, const char *fmt, ...)
{
	char s[256];
	va_list args;
	unsigned int len;
	unsigned long flags;
155

L
Linus Torvalds 已提交
156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201
	WARN_ON(!prefix);
	if (type & debug) {
		spin_lock_irqsave(&disable_ratelimit_lock, flags);
		if (!disable_ratelimit && debug_ratelimit && !printk_ratelimit()) {
			spin_unlock_irqrestore(&disable_ratelimit_lock, flags);
			return;
		}
		spin_unlock_irqrestore(&disable_ratelimit_lock, flags);

		len = snprintf(s, 256, KERN_DEBUG "%s: ", prefix);

		va_start(args, fmt);
		len += vsnprintf(&s[len], (256 - len), fmt, args);
		va_end(args);

		printk(s);

		WARN_ON(len < 5);
	}
}
EXPORT_SYMBOL(cpufreq_debug_printk);


module_param(debug, uint, 0644);
MODULE_PARM_DESC(debug, "CPUfreq debugging: add 1 to debug core, 2 to debug drivers, and 4 to debug governors.");

module_param(debug_ratelimit, uint, 0644);
MODULE_PARM_DESC(debug_ratelimit, "CPUfreq debugging: set to 0 to disable ratelimiting.");

#else /* !CONFIG_CPU_FREQ_DEBUG */

static inline void cpufreq_debug_enable_ratelimit(void) { return; }
static inline void cpufreq_debug_disable_ratelimit(void) { return; }

#endif /* CONFIG_CPU_FREQ_DEBUG */


/*********************************************************************
 *            EXTERNALLY AFFECTING FREQUENCY CHANGES                 *
 *********************************************************************/

/**
 * adjust_jiffies - adjust the system "loops_per_jiffy"
 *
 * This function alters the system "loops_per_jiffy" for the clock
 * speed change. Note that loops_per_jiffy cannot be updated on SMP
202
 * systems as each CPU might be scaled differently. So, use the arch
L
Linus Torvalds 已提交
203 204 205 206 207 208
 * per-CPU loops_per_jiffy value wherever possible.
 */
#ifndef CONFIG_SMP
static unsigned long l_p_j_ref;
static unsigned int  l_p_j_ref_freq;

209
static void adjust_jiffies(unsigned long val, struct cpufreq_freqs *ci)
L
Linus Torvalds 已提交
210 211 212 213 214 215 216 217 218 219 220
{
	if (ci->flags & CPUFREQ_CONST_LOOPS)
		return;

	if (!l_p_j_ref_freq) {
		l_p_j_ref = loops_per_jiffy;
		l_p_j_ref_freq = ci->old;
		dprintk("saving %lu as reference value for loops_per_jiffy; freq is %u kHz\n", l_p_j_ref, l_p_j_ref_freq);
	}
	if ((val == CPUFREQ_PRECHANGE  && ci->old < ci->new) ||
	    (val == CPUFREQ_POSTCHANGE && ci->old > ci->new) ||
221
	    (val == CPUFREQ_RESUMECHANGE || val == CPUFREQ_SUSPENDCHANGE)) {
L
Linus Torvalds 已提交
222 223 224 225 226 227 228 229 230 231
		loops_per_jiffy = cpufreq_scale(l_p_j_ref, l_p_j_ref_freq, ci->new);
		dprintk("scaling loops_per_jiffy to %lu for frequency %u kHz\n", loops_per_jiffy, ci->new);
	}
}
#else
static inline void adjust_jiffies(unsigned long val, struct cpufreq_freqs *ci) { return; }
#endif


/**
232 233
 * cpufreq_notify_transition - call notifier chain and adjust_jiffies
 * on frequency transition.
L
Linus Torvalds 已提交
234
 *
235 236
 * This function calls the transition notifiers and the "adjust_jiffies"
 * function. It is called twice on all CPU frequency changes that have
237
 * external effects.
L
Linus Torvalds 已提交
238 239 240
 */
void cpufreq_notify_transition(struct cpufreq_freqs *freqs, unsigned int state)
{
241 242
	struct cpufreq_policy *policy;

L
Linus Torvalds 已提交
243 244 245
	BUG_ON(irqs_disabled());

	freqs->flags = cpufreq_driver->flags;
246 247
	dprintk("notification %u of frequency transition to %u kHz\n",
		state, freqs->new);
L
Linus Torvalds 已提交
248

249
	policy = cpufreq_cpu_data[freqs->cpu];
L
Linus Torvalds 已提交
250
	switch (state) {
251

L
Linus Torvalds 已提交
252
	case CPUFREQ_PRECHANGE:
253
		/* detect if the driver reported a value as "old frequency"
254 255
		 * which is not equal to what the cpufreq core thinks is
		 * "old frequency".
L
Linus Torvalds 已提交
256 257
		 */
		if (!(cpufreq_driver->flags & CPUFREQ_CONST_LOOPS)) {
258 259 260 261 262 263
			if ((policy) && (policy->cpu == freqs->cpu) &&
			    (policy->cur) && (policy->cur != freqs->old)) {
				dprintk(KERN_WARNING "Warning: CPU frequency is"
					" %u, cpufreq assumed %u kHz.\n",
					freqs->old, policy->cur);
				freqs->old = policy->cur;
L
Linus Torvalds 已提交
264 265
			}
		}
266 267
		blocking_notifier_call_chain(&cpufreq_transition_notifier_list,
				CPUFREQ_PRECHANGE, freqs);
L
Linus Torvalds 已提交
268 269
		adjust_jiffies(CPUFREQ_PRECHANGE, freqs);
		break;
270

L
Linus Torvalds 已提交
271 272
	case CPUFREQ_POSTCHANGE:
		adjust_jiffies(CPUFREQ_POSTCHANGE, freqs);
273 274
		blocking_notifier_call_chain(&cpufreq_transition_notifier_list,
				CPUFREQ_POSTCHANGE, freqs);
275 276
		if (likely(policy) && likely(policy->cpu == freqs->cpu))
			policy->cur = freqs->new;
L
Linus Torvalds 已提交
277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306
		break;
	}
}
EXPORT_SYMBOL_GPL(cpufreq_notify_transition);



/*********************************************************************
 *                          SYSFS INTERFACE                          *
 *********************************************************************/

/**
 * cpufreq_parse_governor - parse a governor string
 */
static int cpufreq_parse_governor (char *str_governor, unsigned int *policy,
				struct cpufreq_governor **governor)
{
	if (!cpufreq_driver)
		return -EINVAL;
	if (cpufreq_driver->setpolicy) {
		if (!strnicmp(str_governor, "performance", CPUFREQ_NAME_LEN)) {
			*policy = CPUFREQ_POLICY_PERFORMANCE;
			return 0;
		} else if (!strnicmp(str_governor, "powersave", CPUFREQ_NAME_LEN)) {
			*policy = CPUFREQ_POLICY_POWERSAVE;
			return 0;
		}
		return -EINVAL;
	} else {
		struct cpufreq_governor *t;
307
		mutex_lock(&cpufreq_governor_mutex);
L
Linus Torvalds 已提交
308 309 310 311 312
		if (!cpufreq_driver || !cpufreq_driver->target)
			goto out;
		list_for_each_entry(t, &cpufreq_governor_list, governor_list) {
			if (!strnicmp(str_governor,t->name,CPUFREQ_NAME_LEN)) {
				*governor = t;
313
				mutex_unlock(&cpufreq_governor_mutex);
L
Linus Torvalds 已提交
314 315 316
				return 0;
			}
		}
317
out:
318
		mutex_unlock(&cpufreq_governor_mutex);
L
Linus Torvalds 已提交
319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335
	}
	return -EINVAL;
}
EXPORT_SYMBOL_GPL(cpufreq_parse_governor);


/* drivers/base/cpu.c */
extern struct sysdev_class cpu_sysdev_class;


/**
 * cpufreq_per_cpu_attr_read() / show_##file_name() - print out cpufreq information
 *
 * Write out information from cpufreq_driver->policy[cpu]; object must be
 * "unsigned int".
 */

336 337 338 339 340
#define show_one(file_name, object)			\
static ssize_t show_##file_name				\
(struct cpufreq_policy * policy, char *buf)		\
{							\
	return sprintf (buf, "%u\n", policy->object);	\
L
Linus Torvalds 已提交
341 342 343 344 345 346 347 348
}

show_one(cpuinfo_min_freq, cpuinfo.min_freq);
show_one(cpuinfo_max_freq, cpuinfo.max_freq);
show_one(scaling_min_freq, min);
show_one(scaling_max_freq, max);
show_one(scaling_cur_freq, cur);

349 350
static int __cpufreq_set_policy(struct cpufreq_policy *data, struct cpufreq_policy *policy);

L
Linus Torvalds 已提交
351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368
/**
 * cpufreq_per_cpu_attr_write() / store_##file_name() - sysfs write access
 */
#define store_one(file_name, object)			\
static ssize_t store_##file_name					\
(struct cpufreq_policy * policy, const char *buf, size_t count)		\
{									\
	unsigned int ret = -EINVAL;					\
	struct cpufreq_policy new_policy;				\
									\
	ret = cpufreq_get_policy(&new_policy, policy->cpu);		\
	if (ret)							\
		return -EINVAL;						\
									\
	ret = sscanf (buf, "%u", &new_policy.object);			\
	if (ret != 1)							\
		return -EINVAL;						\
									\
369 370 371 372
	mutex_lock(&policy->lock);					\
	ret = __cpufreq_set_policy(policy, &new_policy);		\
	policy->user_policy.object = policy->object;			\
	mutex_unlock(&policy->lock);					\
L
Linus Torvalds 已提交
373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409
									\
	return ret ? ret : count;					\
}

store_one(scaling_min_freq,min);
store_one(scaling_max_freq,max);

/**
 * show_cpuinfo_cur_freq - current CPU frequency as detected by hardware
 */
static ssize_t show_cpuinfo_cur_freq (struct cpufreq_policy * policy, char *buf)
{
	unsigned int cur_freq = cpufreq_get(policy->cpu);
	if (!cur_freq)
		return sprintf(buf, "<unknown>");
	return sprintf(buf, "%u\n", cur_freq);
}


/**
 * show_scaling_governor - show the current policy for the specified CPU
 */
static ssize_t show_scaling_governor (struct cpufreq_policy * policy, char *buf)
{
	if(policy->policy == CPUFREQ_POLICY_POWERSAVE)
		return sprintf(buf, "powersave\n");
	else if (policy->policy == CPUFREQ_POLICY_PERFORMANCE)
		return sprintf(buf, "performance\n");
	else if (policy->governor)
		return scnprintf(buf, CPUFREQ_NAME_LEN, "%s\n", policy->governor->name);
	return -EINVAL;
}


/**
 * store_scaling_governor - store policy for the specified CPU
 */
410 411
static ssize_t store_scaling_governor (struct cpufreq_policy * policy,
				       const char *buf, size_t count)
L
Linus Torvalds 已提交
412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427
{
	unsigned int ret = -EINVAL;
	char	str_governor[16];
	struct cpufreq_policy new_policy;

	ret = cpufreq_get_policy(&new_policy, policy->cpu);
	if (ret)
		return ret;

	ret = sscanf (buf, "%15s", str_governor);
	if (ret != 1)
		return -EINVAL;

	if (cpufreq_parse_governor(str_governor, &new_policy.policy, &new_policy.governor))
		return -EINVAL;

428 429 430 431 432 433 434 435 436
	/* Do not use cpufreq_set_policy here or the user_policy.max
	   will be wrongly overridden */
	mutex_lock(&policy->lock);
	ret = __cpufreq_set_policy(policy, &new_policy);

	policy->user_policy.policy = policy->policy;
	policy->user_policy.governor = policy->governor;
	mutex_unlock(&policy->lock);

L
Linus Torvalds 已提交
437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466
	return ret ? ret : count;
}

/**
 * show_scaling_driver - show the cpufreq driver currently loaded
 */
static ssize_t show_scaling_driver (struct cpufreq_policy * policy, char *buf)
{
	return scnprintf(buf, CPUFREQ_NAME_LEN, "%s\n", cpufreq_driver->name);
}

/**
 * show_scaling_available_governors - show the available CPUfreq governors
 */
static ssize_t show_scaling_available_governors (struct cpufreq_policy * policy,
				char *buf)
{
	ssize_t i = 0;
	struct cpufreq_governor *t;

	if (!cpufreq_driver->target) {
		i += sprintf(buf, "performance powersave");
		goto out;
	}

	list_for_each_entry(t, &cpufreq_governor_list, governor_list) {
		if (i >= (ssize_t) ((PAGE_SIZE / sizeof(char)) - (CPUFREQ_NAME_LEN + 2)))
			goto out;
		i += scnprintf(&buf[i], CPUFREQ_NAME_LEN, "%s ", t->name);
	}
467
out:
L
Linus Torvalds 已提交
468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536
	i += sprintf(&buf[i], "\n");
	return i;
}
/**
 * show_affected_cpus - show the CPUs affected by each transition
 */
static ssize_t show_affected_cpus (struct cpufreq_policy * policy, char *buf)
{
	ssize_t i = 0;
	unsigned int cpu;

	for_each_cpu_mask(cpu, policy->cpus) {
		if (i)
			i += scnprintf(&buf[i], (PAGE_SIZE - i - 2), " ");
		i += scnprintf(&buf[i], (PAGE_SIZE - i - 2), "%u", cpu);
		if (i >= (PAGE_SIZE - 5))
		    break;
	}
	i += sprintf(&buf[i], "\n");
	return i;
}


#define define_one_ro(_name) \
static struct freq_attr _name = \
__ATTR(_name, 0444, show_##_name, NULL)

#define define_one_ro0400(_name) \
static struct freq_attr _name = \
__ATTR(_name, 0400, show_##_name, NULL)

#define define_one_rw(_name) \
static struct freq_attr _name = \
__ATTR(_name, 0644, show_##_name, store_##_name)

define_one_ro0400(cpuinfo_cur_freq);
define_one_ro(cpuinfo_min_freq);
define_one_ro(cpuinfo_max_freq);
define_one_ro(scaling_available_governors);
define_one_ro(scaling_driver);
define_one_ro(scaling_cur_freq);
define_one_ro(affected_cpus);
define_one_rw(scaling_min_freq);
define_one_rw(scaling_max_freq);
define_one_rw(scaling_governor);

static struct attribute * default_attrs[] = {
	&cpuinfo_min_freq.attr,
	&cpuinfo_max_freq.attr,
	&scaling_min_freq.attr,
	&scaling_max_freq.attr,
	&affected_cpus.attr,
	&scaling_governor.attr,
	&scaling_driver.attr,
	&scaling_available_governors.attr,
	NULL
};

#define to_policy(k) container_of(k,struct cpufreq_policy,kobj)
#define to_attr(a) container_of(a,struct freq_attr,attr)

static ssize_t show(struct kobject * kobj, struct attribute * attr ,char * buf)
{
	struct cpufreq_policy * policy = to_policy(kobj);
	struct freq_attr * fattr = to_attr(attr);
	ssize_t ret;
	policy = cpufreq_cpu_get(policy->cpu);
	if (!policy)
		return -EINVAL;
537
	ret = fattr->show ? fattr->show(policy,buf) : -EIO;
L
Linus Torvalds 已提交
538 539 540 541
	cpufreq_cpu_put(policy);
	return ret;
}

542
static ssize_t store(struct kobject * kobj, struct attribute * attr,
L
Linus Torvalds 已提交
543 544 545 546 547 548 549 550
		     const char * buf, size_t count)
{
	struct cpufreq_policy * policy = to_policy(kobj);
	struct freq_attr * fattr = to_attr(attr);
	ssize_t ret;
	policy = cpufreq_cpu_get(policy->cpu);
	if (!policy)
		return -EINVAL;
551
	ret = fattr->store ? fattr->store(policy,buf,count) : -EIO;
L
Linus Torvalds 已提交
552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577
	cpufreq_cpu_put(policy);
	return ret;
}

static void cpufreq_sysfs_release(struct kobject * kobj)
{
	struct cpufreq_policy * policy = to_policy(kobj);
	dprintk("last reference is dropped\n");
	complete(&policy->kobj_unregister);
}

static struct sysfs_ops sysfs_ops = {
	.show	= show,
	.store	= store,
};

static struct kobj_type ktype_cpufreq = {
	.sysfs_ops	= &sysfs_ops,
	.default_attrs	= default_attrs,
	.release	= cpufreq_sysfs_release,
};


/**
 * cpufreq_add_dev - add a CPU device
 *
578
 * Adds the cpufreq interface for a CPU device.
L
Linus Torvalds 已提交
579 580 581 582 583 584 585 586
 */
static int cpufreq_add_dev (struct sys_device * sys_dev)
{
	unsigned int cpu = sys_dev->id;
	int ret = 0;
	struct cpufreq_policy new_policy;
	struct cpufreq_policy *policy;
	struct freq_attr **drv_attr;
587
	struct sys_device *cpu_sys_dev;
L
Linus Torvalds 已提交
588 589
	unsigned long flags;
	unsigned int j;
590 591 592
#ifdef CONFIG_SMP
	struct cpufreq_policy *managed_policy;
#endif
L
Linus Torvalds 已提交
593

594 595 596
	if (cpu_is_offline(cpu))
		return 0;

L
Linus Torvalds 已提交
597 598 599 600 601 602 603 604
	cpufreq_debug_disable_ratelimit();
	dprintk("adding CPU %u\n", cpu);

#ifdef CONFIG_SMP
	/* check whether a different CPU already registered this
	 * CPU because it is in the same boat. */
	policy = cpufreq_cpu_get(cpu);
	if (unlikely(policy)) {
605
		cpufreq_cpu_put(policy);
L
Linus Torvalds 已提交
606 607 608 609 610 611 612 613 614 615
		cpufreq_debug_enable_ratelimit();
		return 0;
	}
#endif

	if (!try_module_get(cpufreq_driver->owner)) {
		ret = -EINVAL;
		goto module_out;
	}

616
	policy = kzalloc(sizeof(struct cpufreq_policy), GFP_KERNEL);
L
Linus Torvalds 已提交
617 618 619 620 621 622 623 624
	if (!policy) {
		ret = -ENOMEM;
		goto nomem_out;
	}

	policy->cpu = cpu;
	policy->cpus = cpumask_of_cpu(cpu);

625 626
	mutex_init(&policy->lock);
	mutex_lock(&policy->lock);
L
Linus Torvalds 已提交
627 628 629 630 631 632 633 634 635
	init_completion(&policy->kobj_unregister);
	INIT_WORK(&policy->update, handle_update, (void *)(long)cpu);

	/* call driver. From then on the cpufreq must be able
	 * to accept all calls to ->verify and ->setpolicy for this CPU
	 */
	ret = cpufreq_driver->init(policy);
	if (ret) {
		dprintk("initialization failed\n");
636
		mutex_unlock(&policy->lock);
L
Linus Torvalds 已提交
637 638 639
		goto err_out;
	}

640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665
#ifdef CONFIG_SMP
	for_each_cpu_mask(j, policy->cpus) {
		if (cpu == j)
			continue;

		/* check for existing affected CPUs.  They may not be aware
		 * of it due to CPU Hotplug.
		 */
		managed_policy = cpufreq_cpu_get(j);
		if (unlikely(managed_policy)) {
			spin_lock_irqsave(&cpufreq_driver_lock, flags);
			managed_policy->cpus = policy->cpus;
			cpufreq_cpu_data[cpu] = managed_policy;
			spin_unlock_irqrestore(&cpufreq_driver_lock, flags);

			dprintk("CPU already managed, adding link\n");
			sysfs_create_link(&sys_dev->kobj,
					  &managed_policy->kobj, "cpufreq");

			cpufreq_debug_enable_ratelimit();
			mutex_unlock(&policy->lock);
			ret = 0;
			goto err_out_driver_exit; /* call driver->exit() */
		}
	}
#endif
L
Linus Torvalds 已提交
666 667 668 669 670 671 672 673
	memcpy(&new_policy, policy, sizeof(struct cpufreq_policy));

	/* prepare interface data */
	policy->kobj.parent = &sys_dev->kobj;
	policy->kobj.ktype = &ktype_cpufreq;
	strlcpy(policy->kobj.name, "cpufreq", KOBJ_NAME_LEN);

	ret = kobject_register(&policy->kobj);
674 675
	if (ret) {
		mutex_unlock(&policy->lock);
676
		goto err_out_driver_exit;
677
	}
L
Linus Torvalds 已提交
678 679 680 681 682 683 684 685 686 687 688 689 690 691 692
	/* set up files for this cpu device */
	drv_attr = cpufreq_driver->attr;
	while ((drv_attr) && (*drv_attr)) {
		sysfs_create_file(&policy->kobj, &((*drv_attr)->attr));
		drv_attr++;
	}
	if (cpufreq_driver->get)
		sysfs_create_file(&policy->kobj, &cpuinfo_cur_freq.attr);
	if (cpufreq_driver->target)
		sysfs_create_file(&policy->kobj, &scaling_cur_freq.attr);

	spin_lock_irqsave(&cpufreq_driver_lock, flags);
	for_each_cpu_mask(j, policy->cpus)
		cpufreq_cpu_data[j] = policy;
	spin_unlock_irqrestore(&cpufreq_driver_lock, flags);
693 694 695 696 697 698 699 700

	/* symlink affected CPUs */
	for_each_cpu_mask(j, policy->cpus) {
		if (j == cpu)
			continue;
		if (!cpu_online(j))
			continue;

701
		dprintk("CPU %u already managed, adding link\n", j);
702 703 704 705 706 707
		cpufreq_cpu_get(cpu);
		cpu_sys_dev = get_cpu_sysdev(j);
		sysfs_create_link(&cpu_sys_dev->kobj, &policy->kobj,
				  "cpufreq");
	}

L
Linus Torvalds 已提交
708 709
	policy->governor = NULL; /* to assure that the starting sequence is
				  * run in cpufreq_set_policy */
710
	mutex_unlock(&policy->lock);
711

L
Linus Torvalds 已提交
712 713 714 715 716 717 718 719 720 721
	/* set default policy */
	ret = cpufreq_set_policy(&new_policy);
	if (ret) {
		dprintk("setting policy failed\n");
		goto err_out_unregister;
	}

	module_put(cpufreq_driver->owner);
	dprintk("initialization complete\n");
	cpufreq_debug_enable_ratelimit();
722

L
Linus Torvalds 已提交
723 724 725 726 727 728 729 730 731 732 733 734
	return 0;


err_out_unregister:
	spin_lock_irqsave(&cpufreq_driver_lock, flags);
	for_each_cpu_mask(j, policy->cpus)
		cpufreq_cpu_data[j] = NULL;
	spin_unlock_irqrestore(&cpufreq_driver_lock, flags);

	kobject_unregister(&policy->kobj);
	wait_for_completion(&policy->kobj_unregister);

735 736 737 738
err_out_driver_exit:
	if (cpufreq_driver->exit)
		cpufreq_driver->exit(policy);

L
Linus Torvalds 已提交
739 740 741 742 743
err_out:
	kfree(policy);

nomem_out:
	module_put(cpufreq_driver->owner);
744
module_out:
L
Linus Torvalds 已提交
745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760
	cpufreq_debug_enable_ratelimit();
	return ret;
}


/**
 * cpufreq_remove_dev - remove a CPU device
 *
 * Removes the cpufreq interface for a CPU device.
 */
static int cpufreq_remove_dev (struct sys_device * sys_dev)
{
	unsigned int cpu = sys_dev->id;
	unsigned long flags;
	struct cpufreq_policy *data;
#ifdef CONFIG_SMP
761
	struct sys_device *cpu_sys_dev;
L
Linus Torvalds 已提交
762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780
	unsigned int j;
#endif

	cpufreq_debug_disable_ratelimit();
	dprintk("unregistering CPU %u\n", cpu);

	spin_lock_irqsave(&cpufreq_driver_lock, flags);
	data = cpufreq_cpu_data[cpu];

	if (!data) {
		spin_unlock_irqrestore(&cpufreq_driver_lock, flags);
		cpufreq_debug_enable_ratelimit();
		return -EINVAL;
	}
	cpufreq_cpu_data[cpu] = NULL;


#ifdef CONFIG_SMP
	/* if this isn't the CPU which is the parent of the kobj, we
781
	 * only need to unlink, put and exit
L
Linus Torvalds 已提交
782 783 784
	 */
	if (unlikely(cpu != data->cpu)) {
		dprintk("removing link\n");
785
		cpu_clear(cpu, data->cpus);
L
Linus Torvalds 已提交
786 787 788 789 790 791 792 793 794 795 796 797
		spin_unlock_irqrestore(&cpufreq_driver_lock, flags);
		sysfs_remove_link(&sys_dev->kobj, "cpufreq");
		cpufreq_cpu_put(data);
		cpufreq_debug_enable_ratelimit();
		return 0;
	}
#endif


	if (!kobject_get(&data->kobj)) {
		spin_unlock_irqrestore(&cpufreq_driver_lock, flags);
		cpufreq_debug_enable_ratelimit();
798
		return -EFAULT;
L
Linus Torvalds 已提交
799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821
	}

#ifdef CONFIG_SMP
	/* if we have other CPUs still registered, we need to unlink them,
	 * or else wait_for_completion below will lock up. Clean the
	 * cpufreq_cpu_data[] while holding the lock, and remove the sysfs
	 * links afterwards.
	 */
	if (unlikely(cpus_weight(data->cpus) > 1)) {
		for_each_cpu_mask(j, data->cpus) {
			if (j == cpu)
				continue;
			cpufreq_cpu_data[j] = NULL;
		}
	}

	spin_unlock_irqrestore(&cpufreq_driver_lock, flags);

	if (unlikely(cpus_weight(data->cpus) > 1)) {
		for_each_cpu_mask(j, data->cpus) {
			if (j == cpu)
				continue;
			dprintk("removing link for cpu %u\n", j);
822 823
			cpu_sys_dev = get_cpu_sysdev(j);
			sysfs_remove_link(&cpu_sys_dev->kobj, "cpufreq");
L
Linus Torvalds 已提交
824 825 826 827 828 829 830
			cpufreq_cpu_put(data);
		}
	}
#else
	spin_unlock_irqrestore(&cpufreq_driver_lock, flags);
#endif

831
	mutex_lock(&data->lock);
L
Linus Torvalds 已提交
832 833
	if (cpufreq_driver->target)
		__cpufreq_governor(data, CPUFREQ_GOV_STOP);
834
	mutex_unlock(&data->lock);
L
Linus Torvalds 已提交
835 836 837 838 839 840

	kobject_unregister(&data->kobj);

	kobject_put(&data->kobj);

	/* we need to make sure that the underlying kobj is actually
841
	 * not referenced anymore by anybody before we proceed with
L
Linus Torvalds 已提交
842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877
	 * unloading.
	 */
	dprintk("waiting for dropping of refcount\n");
	wait_for_completion(&data->kobj_unregister);
	dprintk("wait complete\n");

	if (cpufreq_driver->exit)
		cpufreq_driver->exit(data);

	kfree(data);

	cpufreq_debug_enable_ratelimit();
	return 0;
}


static void handle_update(void *data)
{
	unsigned int cpu = (unsigned int)(long)data;
	dprintk("handle_update for cpu %u called\n", cpu);
	cpufreq_update_policy(cpu);
}

/**
 *	cpufreq_out_of_sync - If actual and saved CPU frequency differs, we're in deep trouble.
 *	@cpu: cpu number
 *	@old_freq: CPU frequency the kernel thinks the CPU runs at
 *	@new_freq: CPU frequency the CPU actually runs at
 *
 *	We adjust to current frequency first, and need to clean up later. So either call
 *	to cpufreq_update_policy() or schedule handle_update()).
 */
static void cpufreq_out_of_sync(unsigned int cpu, unsigned int old_freq, unsigned int new_freq)
{
	struct cpufreq_freqs freqs;

878
	dprintk(KERN_WARNING "Warning: CPU frequency out of sync: cpufreq and timing "
L
Linus Torvalds 已提交
879 880 881 882 883 884 885 886 887 888
	       "core thinks of %u, is %u kHz.\n", old_freq, new_freq);

	freqs.cpu = cpu;
	freqs.old = old_freq;
	freqs.new = new_freq;
	cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE);
	cpufreq_notify_transition(&freqs, CPUFREQ_POSTCHANGE);
}


889
/**
890 891 892 893 894 895 896 897 898 899 900 901
 * cpufreq_quick_get - get the CPU frequency (in kHz) frpm policy->cur
 * @cpu: CPU number
 *
 * This is the last known freq, without actually getting it from the driver.
 * Return value will be same as what is shown in scaling_cur_freq in sysfs.
 */
unsigned int cpufreq_quick_get(unsigned int cpu)
{
	struct cpufreq_policy *policy = cpufreq_cpu_get(cpu);
	unsigned int ret = 0;

	if (policy) {
902
		mutex_lock(&policy->lock);
903
		ret = policy->cur;
904
		mutex_unlock(&policy->lock);
905 906 907 908 909 910 911 912
		cpufreq_cpu_put(policy);
	}

	return (ret);
}
EXPORT_SYMBOL(cpufreq_quick_get);


913
/**
L
Linus Torvalds 已提交
914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929
 * cpufreq_get - get the current CPU frequency (in kHz)
 * @cpu: CPU number
 *
 * Get the CPU current (static) CPU frequency
 */
unsigned int cpufreq_get(unsigned int cpu)
{
	struct cpufreq_policy *policy = cpufreq_cpu_get(cpu);
	unsigned int ret = 0;

	if (!policy)
		return 0;

	if (!cpufreq_driver->get)
		goto out;

930
	mutex_lock(&policy->lock);
L
Linus Torvalds 已提交
931 932 933

	ret = cpufreq_driver->get(cpu);

934
	if (ret && policy->cur && !(cpufreq_driver->flags & CPUFREQ_CONST_LOOPS)) {
L
Linus Torvalds 已提交
935 936 937 938 939 940 941
		/* verify no discrepancy between actual and saved value exists */
		if (unlikely(ret != policy->cur)) {
			cpufreq_out_of_sync(cpu, policy->cur, ret);
			schedule_work(&policy->update);
		}
	}

942
	mutex_unlock(&policy->lock);
L
Linus Torvalds 已提交
943

944
out:
L
Linus Torvalds 已提交
945 946 947 948 949 950 951
	cpufreq_cpu_put(policy);

	return (ret);
}
EXPORT_SYMBOL(cpufreq_get);


952 953 954 955
/**
 *	cpufreq_suspend - let the low level driver prepare for suspend
 */

956
static int cpufreq_suspend(struct sys_device * sysdev, pm_message_t pmsg)
957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982 983
{
	int cpu = sysdev->id;
	unsigned int ret = 0;
	unsigned int cur_freq = 0;
	struct cpufreq_policy *cpu_policy;

	dprintk("resuming cpu %u\n", cpu);

	if (!cpu_online(cpu))
		return 0;

	/* we may be lax here as interrupts are off. Nonetheless
	 * we need to grab the correct cpu policy, as to check
	 * whether we really run on this CPU.
	 */

	cpu_policy = cpufreq_cpu_get(cpu);
	if (!cpu_policy)
		return -EINVAL;

	/* only handle each CPU group once */
	if (unlikely(cpu_policy->cpu != cpu)) {
		cpufreq_cpu_put(cpu_policy);
		return 0;
	}

	if (cpufreq_driver->suspend) {
984
		ret = cpufreq_driver->suspend(cpu_policy, pmsg);
985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009
		if (ret) {
			printk(KERN_ERR "cpufreq: suspend failed in ->suspend "
					"step on CPU %u\n", cpu_policy->cpu);
			cpufreq_cpu_put(cpu_policy);
			return ret;
		}
	}


	if (cpufreq_driver->flags & CPUFREQ_CONST_LOOPS)
		goto out;

	if (cpufreq_driver->get)
		cur_freq = cpufreq_driver->get(cpu_policy->cpu);

	if (!cur_freq || !cpu_policy->cur) {
		printk(KERN_ERR "cpufreq: suspend failed to assert current "
		       "frequency is what timing core thinks it is.\n");
		goto out;
	}

	if (unlikely(cur_freq != cpu_policy->cur)) {
		struct cpufreq_freqs freqs;

		if (!(cpufreq_driver->flags & CPUFREQ_PM_NO_WARN))
1010
			dprintk(KERN_DEBUG "Warning: CPU frequency is %u, "
1011 1012 1013 1014 1015 1016 1017
			       "cpufreq assumed %u kHz.\n",
			       cur_freq, cpu_policy->cur);

		freqs.cpu = cpu;
		freqs.old = cpu_policy->cur;
		freqs.new = cur_freq;

1018
		blocking_notifier_call_chain(&cpufreq_transition_notifier_list,
1019 1020 1021 1022 1023 1024
				    CPUFREQ_SUSPENDCHANGE, &freqs);
		adjust_jiffies(CPUFREQ_SUSPENDCHANGE, &freqs);

		cpu_policy->cur = cur_freq;
	}

1025
out:
1026 1027 1028 1029
	cpufreq_cpu_put(cpu_policy);
	return 0;
}

L
Linus Torvalds 已提交
1030 1031 1032 1033 1034
/**
 *	cpufreq_resume -  restore proper CPU frequency handling after resume
 *
 *	1.) resume CPUfreq hardware support (cpufreq_driver->resume())
 *	2.) if ->target and !CPUFREQ_CONST_LOOPS: verify we're in sync
1035 1036
 *	3.) schedule call cpufreq_update_policy() ASAP as interrupts are
 *	    restored.
L
Linus Torvalds 已提交
1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080
 */
static int cpufreq_resume(struct sys_device * sysdev)
{
	int cpu = sysdev->id;
	unsigned int ret = 0;
	struct cpufreq_policy *cpu_policy;

	dprintk("resuming cpu %u\n", cpu);

	if (!cpu_online(cpu))
		return 0;

	/* we may be lax here as interrupts are off. Nonetheless
	 * we need to grab the correct cpu policy, as to check
	 * whether we really run on this CPU.
	 */

	cpu_policy = cpufreq_cpu_get(cpu);
	if (!cpu_policy)
		return -EINVAL;

	/* only handle each CPU group once */
	if (unlikely(cpu_policy->cpu != cpu)) {
		cpufreq_cpu_put(cpu_policy);
		return 0;
	}

	if (cpufreq_driver->resume) {
		ret = cpufreq_driver->resume(cpu_policy);
		if (ret) {
			printk(KERN_ERR "cpufreq: resume failed in ->resume "
					"step on CPU %u\n", cpu_policy->cpu);
			cpufreq_cpu_put(cpu_policy);
			return ret;
		}
	}

	if (!(cpufreq_driver->flags & CPUFREQ_CONST_LOOPS)) {
		unsigned int cur_freq = 0;

		if (cpufreq_driver->get)
			cur_freq = cpufreq_driver->get(cpu_policy->cpu);

		if (!cur_freq || !cpu_policy->cur) {
1081 1082 1083
			printk(KERN_ERR "cpufreq: resume failed to assert "
					"current frequency is what timing core "
					"thinks it is.\n");
L
Linus Torvalds 已提交
1084 1085 1086 1087 1088 1089
			goto out;
		}

		if (unlikely(cur_freq != cpu_policy->cur)) {
			struct cpufreq_freqs freqs;

1090
			if (!(cpufreq_driver->flags & CPUFREQ_PM_NO_WARN))
1091
				dprintk(KERN_WARNING "Warning: CPU frequency"
1092 1093
				       "is %u, cpufreq assumed %u kHz.\n",
				       cur_freq, cpu_policy->cur);
L
Linus Torvalds 已提交
1094 1095 1096 1097 1098

			freqs.cpu = cpu;
			freqs.old = cpu_policy->cur;
			freqs.new = cur_freq;

1099 1100
			blocking_notifier_call_chain(
					&cpufreq_transition_notifier_list,
1101
					CPUFREQ_RESUMECHANGE, &freqs);
L
Linus Torvalds 已提交
1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116
			adjust_jiffies(CPUFREQ_RESUMECHANGE, &freqs);

			cpu_policy->cur = cur_freq;
		}
	}

out:
	schedule_work(&cpu_policy->update);
	cpufreq_cpu_put(cpu_policy);
	return ret;
}

static struct sysdev_driver cpufreq_sysdev_driver = {
	.add		= cpufreq_add_dev,
	.remove		= cpufreq_remove_dev,
1117
	.suspend	= cpufreq_suspend,
L
Linus Torvalds 已提交
1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130
	.resume		= cpufreq_resume,
};


/*********************************************************************
 *                     NOTIFIER LISTS INTERFACE                      *
 *********************************************************************/

/**
 *	cpufreq_register_notifier - register a driver with cpufreq
 *	@nb: notifier function to register
 *      @list: CPUFREQ_TRANSITION_NOTIFIER or CPUFREQ_POLICY_NOTIFIER
 *
1131
 *	Add a driver to one of two lists: either a list of drivers that
L
Linus Torvalds 已提交
1132 1133 1134 1135 1136
 *      are notified about clock rate changes (once before and once after
 *      the transition), or a list of drivers that are notified about
 *      changes in cpufreq policy.
 *
 *	This function may sleep, and has the same return conditions as
1137
 *	blocking_notifier_chain_register.
L
Linus Torvalds 已提交
1138 1139 1140 1141 1142 1143 1144
 */
int cpufreq_register_notifier(struct notifier_block *nb, unsigned int list)
{
	int ret;

	switch (list) {
	case CPUFREQ_TRANSITION_NOTIFIER:
1145 1146
		ret = blocking_notifier_chain_register(
				&cpufreq_transition_notifier_list, nb);
L
Linus Torvalds 已提交
1147 1148
		break;
	case CPUFREQ_POLICY_NOTIFIER:
1149 1150
		ret = blocking_notifier_chain_register(
				&cpufreq_policy_notifier_list, nb);
L
Linus Torvalds 已提交
1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168
		break;
	default:
		ret = -EINVAL;
	}

	return ret;
}
EXPORT_SYMBOL(cpufreq_register_notifier);


/**
 *	cpufreq_unregister_notifier - unregister a driver with cpufreq
 *	@nb: notifier block to be unregistered
 *      @list: CPUFREQ_TRANSITION_NOTIFIER or CPUFREQ_POLICY_NOTIFIER
 *
 *	Remove a driver from the CPU frequency notifier list.
 *
 *	This function may sleep, and has the same return conditions as
1169
 *	blocking_notifier_chain_unregister.
L
Linus Torvalds 已提交
1170 1171 1172 1173 1174 1175 1176
 */
int cpufreq_unregister_notifier(struct notifier_block *nb, unsigned int list)
{
	int ret;

	switch (list) {
	case CPUFREQ_TRANSITION_NOTIFIER:
1177 1178
		ret = blocking_notifier_chain_unregister(
				&cpufreq_transition_notifier_list, nb);
L
Linus Torvalds 已提交
1179 1180
		break;
	case CPUFREQ_POLICY_NOTIFIER:
1181 1182
		ret = blocking_notifier_chain_unregister(
				&cpufreq_policy_notifier_list, nb);
L
Linus Torvalds 已提交
1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202
		break;
	default:
		ret = -EINVAL;
	}

	return ret;
}
EXPORT_SYMBOL(cpufreq_unregister_notifier);


/*********************************************************************
 *                              GOVERNORS                            *
 *********************************************************************/


int __cpufreq_driver_target(struct cpufreq_policy *policy,
			    unsigned int target_freq,
			    unsigned int relation)
{
	int retval = -EINVAL;
1203

1204
	lock_cpu_hotplug();
L
Linus Torvalds 已提交
1205 1206 1207 1208
	dprintk("target for CPU %u: %u kHz, relation %u\n", policy->cpu,
		target_freq, relation);
	if (cpu_online(policy->cpu) && cpufreq_driver->target)
		retval = cpufreq_driver->target(policy, target_freq, relation);
1209

1210
	unlock_cpu_hotplug();
1211

L
Linus Torvalds 已提交
1212 1213 1214 1215 1216 1217 1218 1219
	return retval;
}
EXPORT_SYMBOL_GPL(__cpufreq_driver_target);

int cpufreq_driver_target(struct cpufreq_policy *policy,
			  unsigned int target_freq,
			  unsigned int relation)
{
1220
	int ret;
L
Linus Torvalds 已提交
1221 1222 1223 1224 1225

	policy = cpufreq_cpu_get(policy->cpu);
	if (!policy)
		return -EINVAL;

1226
	mutex_lock(&policy->lock);
L
Linus Torvalds 已提交
1227 1228 1229

	ret = __cpufreq_driver_target(policy, target_freq, relation);

1230
	mutex_unlock(&policy->lock);
L
Linus Torvalds 已提交
1231 1232 1233 1234 1235 1236 1237 1238 1239

	cpufreq_cpu_put(policy);
	return ret;
}
EXPORT_SYMBOL_GPL(cpufreq_driver_target);


static int __cpufreq_governor(struct cpufreq_policy *policy, unsigned int event)
{
1240
	int ret;
L
Linus Torvalds 已提交
1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265

	if (!try_module_get(policy->governor->owner))
		return -EINVAL;

	dprintk("__cpufreq_governor for CPU %u, event %u\n", policy->cpu, event);
	ret = policy->governor->governor(policy, event);

	/* we keep one module reference alive for each CPU governed by this CPU */
	if ((event != CPUFREQ_GOV_START) || ret)
		module_put(policy->governor->owner);
	if ((event == CPUFREQ_GOV_STOP) && !ret)
		module_put(policy->governor->owner);

	return ret;
}


int cpufreq_governor(unsigned int cpu, unsigned int event)
{
	int ret = 0;
	struct cpufreq_policy *policy = cpufreq_cpu_get(cpu);

	if (!policy)
		return -EINVAL;

1266
	mutex_lock(&policy->lock);
L
Linus Torvalds 已提交
1267
	ret = __cpufreq_governor(policy, event);
1268
	mutex_unlock(&policy->lock);
L
Linus Torvalds 已提交
1269 1270 1271 1272 1273 1274 1275 1276 1277 1278 1279 1280 1281 1282

	cpufreq_cpu_put(policy);
	return ret;
}
EXPORT_SYMBOL_GPL(cpufreq_governor);


int cpufreq_register_governor(struct cpufreq_governor *governor)
{
	struct cpufreq_governor *t;

	if (!governor)
		return -EINVAL;

1283
	mutex_lock(&cpufreq_governor_mutex);
1284

L
Linus Torvalds 已提交
1285 1286
	list_for_each_entry(t, &cpufreq_governor_list, governor_list) {
		if (!strnicmp(governor->name,t->name,CPUFREQ_NAME_LEN)) {
1287
			mutex_unlock(&cpufreq_governor_mutex);
L
Linus Torvalds 已提交
1288 1289 1290 1291 1292
			return -EBUSY;
		}
	}
	list_add(&governor->governor_list, &cpufreq_governor_list);

1293
	mutex_unlock(&cpufreq_governor_mutex);
L
Linus Torvalds 已提交
1294 1295 1296 1297 1298 1299 1300 1301 1302 1303
	return 0;
}
EXPORT_SYMBOL_GPL(cpufreq_register_governor);


void cpufreq_unregister_governor(struct cpufreq_governor *governor)
{
	if (!governor)
		return;

1304
	mutex_lock(&cpufreq_governor_mutex);
L
Linus Torvalds 已提交
1305
	list_del(&governor->governor_list);
1306
	mutex_unlock(&cpufreq_governor_mutex);
L
Linus Torvalds 已提交
1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332
	return;
}
EXPORT_SYMBOL_GPL(cpufreq_unregister_governor);



/*********************************************************************
 *                          POLICY INTERFACE                         *
 *********************************************************************/

/**
 * cpufreq_get_policy - get the current cpufreq_policy
 * @policy: struct cpufreq_policy into which the current cpufreq_policy is written
 *
 * Reads the current cpufreq policy.
 */
int cpufreq_get_policy(struct cpufreq_policy *policy, unsigned int cpu)
{
	struct cpufreq_policy *cpu_policy;
	if (!policy)
		return -EINVAL;

	cpu_policy = cpufreq_cpu_get(cpu);
	if (!cpu_policy)
		return -EINVAL;

1333
	mutex_lock(&cpu_policy->lock);
L
Linus Torvalds 已提交
1334
	memcpy(policy, cpu_policy, sizeof(struct cpufreq_policy));
1335
	mutex_unlock(&cpu_policy->lock);
L
Linus Torvalds 已提交
1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350

	cpufreq_cpu_put(cpu_policy);
	return 0;
}
EXPORT_SYMBOL(cpufreq_get_policy);


static int __cpufreq_set_policy(struct cpufreq_policy *data, struct cpufreq_policy *policy)
{
	int ret = 0;

	cpufreq_debug_disable_ratelimit();
	dprintk("setting new policy for CPU %u: %u - %u kHz\n", policy->cpu,
		policy->min, policy->max);

1351
	memcpy(&policy->cpuinfo, &data->cpuinfo, sizeof(struct cpufreq_cpuinfo));
L
Linus Torvalds 已提交
1352 1353 1354 1355 1356 1357 1358

	/* verify the cpu speed can be set within this limit */
	ret = cpufreq_driver->verify(policy);
	if (ret)
		goto error_out;

	/* adjust if necessary - all reasons */
1359 1360
	blocking_notifier_call_chain(&cpufreq_policy_notifier_list,
			CPUFREQ_ADJUST, policy);
L
Linus Torvalds 已提交
1361 1362

	/* adjust if necessary - hardware incompatibility*/
1363 1364
	blocking_notifier_call_chain(&cpufreq_policy_notifier_list,
			CPUFREQ_INCOMPATIBLE, policy);
L
Linus Torvalds 已提交
1365 1366 1367 1368

	/* verify the cpu speed can be set within this limit,
	   which might be different to the first one */
	ret = cpufreq_driver->verify(policy);
1369
	if (ret)
L
Linus Torvalds 已提交
1370 1371 1372
		goto error_out;

	/* notification of the new policy */
1373 1374
	blocking_notifier_call_chain(&cpufreq_policy_notifier_list,
			CPUFREQ_NOTIFY, policy);
L
Linus Torvalds 已提交
1375

1376 1377
	data->min = policy->min;
	data->max = policy->max;
L
Linus Torvalds 已提交
1378 1379 1380 1381 1382 1383 1384 1385 1386 1387 1388 1389 1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408 1409 1410 1411 1412 1413

	dprintk("new min and max freqs are %u - %u kHz\n", data->min, data->max);

	if (cpufreq_driver->setpolicy) {
		data->policy = policy->policy;
		dprintk("setting range\n");
		ret = cpufreq_driver->setpolicy(policy);
	} else {
		if (policy->governor != data->governor) {
			/* save old, working values */
			struct cpufreq_governor *old_gov = data->governor;

			dprintk("governor switch\n");

			/* end old governor */
			if (data->governor)
				__cpufreq_governor(data, CPUFREQ_GOV_STOP);

			/* start new governor */
			data->governor = policy->governor;
			if (__cpufreq_governor(data, CPUFREQ_GOV_START)) {
				/* new governor failed, so re-start old one */
				dprintk("starting governor %s failed\n", data->governor->name);
				if (old_gov) {
					data->governor = old_gov;
					__cpufreq_governor(data, CPUFREQ_GOV_START);
				}
				ret = -EINVAL;
				goto error_out;
			}
			/* might be a policy change, too, so fall through */
		}
		dprintk("governor: change or update limits\n");
		__cpufreq_governor(data, CPUFREQ_GOV_LIMITS);
	}

1414
error_out:
L
Linus Torvalds 已提交
1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429 1430 1431 1432 1433 1434 1435 1436 1437
	cpufreq_debug_enable_ratelimit();
	return ret;
}

/**
 *	cpufreq_set_policy - set a new CPUFreq policy
 *	@policy: policy to be set.
 *
 *	Sets a new CPU frequency and voltage scaling policy.
 */
int cpufreq_set_policy(struct cpufreq_policy *policy)
{
	int ret = 0;
	struct cpufreq_policy *data;

	if (!policy)
		return -EINVAL;

	data = cpufreq_cpu_get(policy->cpu);
	if (!data)
		return -EINVAL;

	/* lock this CPU */
1438
	mutex_lock(&data->lock);
L
Linus Torvalds 已提交
1439 1440 1441 1442 1443 1444 1445

	ret = __cpufreq_set_policy(data, policy);
	data->user_policy.min = data->min;
	data->user_policy.max = data->max;
	data->user_policy.policy = data->policy;
	data->user_policy.governor = data->governor;

1446
	mutex_unlock(&data->lock);
L
Linus Torvalds 已提交
1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469
	cpufreq_cpu_put(data);

	return ret;
}
EXPORT_SYMBOL(cpufreq_set_policy);


/**
 *	cpufreq_update_policy - re-evaluate an existing cpufreq policy
 *	@cpu: CPU which shall be re-evaluated
 *
 *	Usefull for policy notifiers which have different necessities
 *	at different times.
 */
int cpufreq_update_policy(unsigned int cpu)
{
	struct cpufreq_policy *data = cpufreq_cpu_get(cpu);
	struct cpufreq_policy policy;
	int ret = 0;

	if (!data)
		return -ENODEV;

1470
	mutex_lock(&data->lock);
L
Linus Torvalds 已提交
1471 1472

	dprintk("updating policy for CPU %u\n", cpu);
1473
	memcpy(&policy, data, sizeof(struct cpufreq_policy));
L
Linus Torvalds 已提交
1474 1475 1476 1477 1478
	policy.min = data->user_policy.min;
	policy.max = data->user_policy.max;
	policy.policy = data->user_policy.policy;
	policy.governor = data->user_policy.governor;

1479 1480 1481 1482
	/* BIOS might change freq behind our back
	  -> ask driver for current freq and notify governors about a change */
	if (cpufreq_driver->get) {
		policy.cur = cpufreq_driver->get(cpu);
1483 1484 1485 1486 1487 1488 1489
		if (!data->cur) {
			dprintk("Driver did not initialize current freq");
			data->cur = policy.cur;
		} else {
			if (data->cur != policy.cur)
				cpufreq_out_of_sync(cpu, data->cur, policy.cur);
		}
1490 1491
	}

L
Linus Torvalds 已提交
1492 1493
	ret = __cpufreq_set_policy(data, &policy);

1494
	mutex_unlock(&data->lock);
L
Linus Torvalds 已提交
1495 1496 1497 1498 1499 1500

	cpufreq_cpu_put(data);
	return ret;
}
EXPORT_SYMBOL(cpufreq_update_policy);

1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531 1532 1533 1534 1535 1536 1537 1538 1539
static int __cpuinit cpufreq_cpu_callback(struct notifier_block *nfb,
					unsigned long action, void *hcpu)
{
	unsigned int cpu = (unsigned long)hcpu;
	struct cpufreq_policy *policy;
	struct sys_device *sys_dev;

	sys_dev = get_cpu_sysdev(cpu);

	if (sys_dev) {
		switch (action) {
		case CPU_ONLINE:
			cpufreq_add_dev(sys_dev);
			break;
		case CPU_DOWN_PREPARE:
			/*
			 * We attempt to put this cpu in lowest frequency
			 * possible before going down. This will permit
			 * hardware-managed P-State to switch other related
			 * threads to min or higher speeds if possible.
			 */
			policy = cpufreq_cpu_data[cpu];
			if (policy) {
				cpufreq_driver_target(policy, policy->min,
						CPUFREQ_RELATION_H);
			}
			break;
		case CPU_DEAD:
			cpufreq_remove_dev(sys_dev);
			break;
		}
	}
	return NOTIFY_OK;
}

static struct notifier_block cpufreq_cpu_notifier =
{
    .notifier_call = cpufreq_cpu_callback,
};
L
Linus Torvalds 已提交
1540 1541 1542 1543 1544 1545 1546 1547 1548 1549

/*********************************************************************
 *               REGISTER / UNREGISTER CPUFREQ DRIVER                *
 *********************************************************************/

/**
 * cpufreq_register_driver - register a CPU Frequency driver
 * @driver_data: A struct cpufreq_driver containing the values#
 * submitted by the CPU Frequency driver.
 *
1550
 *   Registers a CPU Frequency driver to this core code. This code
L
Linus Torvalds 已提交
1551
 * returns zero on success, -EBUSY when another driver got here first
1552
 * (and isn't unregistered in the meantime).
L
Linus Torvalds 已提交
1553 1554 1555 1556 1557 1558 1559 1560 1561 1562 1563 1564 1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599
 *
 */
int cpufreq_register_driver(struct cpufreq_driver *driver_data)
{
	unsigned long flags;
	int ret;

	if (!driver_data || !driver_data->verify || !driver_data->init ||
	    ((!driver_data->setpolicy) && (!driver_data->target)))
		return -EINVAL;

	dprintk("trying to register driver %s\n", driver_data->name);

	if (driver_data->setpolicy)
		driver_data->flags |= CPUFREQ_CONST_LOOPS;

	spin_lock_irqsave(&cpufreq_driver_lock, flags);
	if (cpufreq_driver) {
		spin_unlock_irqrestore(&cpufreq_driver_lock, flags);
		return -EBUSY;
	}
	cpufreq_driver = driver_data;
	spin_unlock_irqrestore(&cpufreq_driver_lock, flags);

	ret = sysdev_driver_register(&cpu_sysdev_class,&cpufreq_sysdev_driver);

	if ((!ret) && !(cpufreq_driver->flags & CPUFREQ_STICKY)) {
		int i;
		ret = -ENODEV;

		/* check for at least one working CPU */
		for (i=0; i<NR_CPUS; i++)
			if (cpufreq_cpu_data[i])
				ret = 0;

		/* if all ->init() calls failed, unregister */
		if (ret) {
			dprintk("no CPU initialized for driver %s\n", driver_data->name);
			sysdev_driver_unregister(&cpu_sysdev_class, &cpufreq_sysdev_driver);

			spin_lock_irqsave(&cpufreq_driver_lock, flags);
			cpufreq_driver = NULL;
			spin_unlock_irqrestore(&cpufreq_driver_lock, flags);
		}
	}

	if (!ret) {
1600
		register_cpu_notifier(&cpufreq_cpu_notifier);
L
Linus Torvalds 已提交
1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612
		dprintk("driver %s up and running\n", driver_data->name);
		cpufreq_debug_enable_ratelimit();
	}

	return (ret);
}
EXPORT_SYMBOL_GPL(cpufreq_register_driver);


/**
 * cpufreq_unregister_driver - unregister the current CPUFreq driver
 *
1613
 *    Unregister the current CPUFreq driver. Only call this if you have
L
Linus Torvalds 已提交
1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631
 * the right to do so, i.e. if you have succeeded in initialising before!
 * Returns zero if successful, and -EINVAL if the cpufreq_driver is
 * currently not initialised.
 */
int cpufreq_unregister_driver(struct cpufreq_driver *driver)
{
	unsigned long flags;

	cpufreq_debug_disable_ratelimit();

	if (!cpufreq_driver || (driver != cpufreq_driver)) {
		cpufreq_debug_enable_ratelimit();
		return -EINVAL;
	}

	dprintk("unregistering driver %s\n", driver->name);

	sysdev_driver_unregister(&cpu_sysdev_class, &cpufreq_sysdev_driver);
1632
	unregister_cpu_notifier(&cpufreq_cpu_notifier);
L
Linus Torvalds 已提交
1633 1634 1635 1636 1637 1638 1639 1640

	spin_lock_irqsave(&cpufreq_driver_lock, flags);
	cpufreq_driver = NULL;
	spin_unlock_irqrestore(&cpufreq_driver_lock, flags);

	return 0;
}
EXPORT_SYMBOL_GPL(cpufreq_unregister_driver);