cpuset.c 70.1 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
1 2 3 4 5 6
/*
 *  kernel/cpuset.c
 *
 *  Processor and Memory placement constraints for sets of tasks.
 *
 *  Copyright (C) 2003 BULL SA.
Paul Jackson's avatar
Paul Jackson committed
7
 *  Copyright (C) 2004-2007 Silicon Graphics, Inc.
8
 *  Copyright (C) 2006 Google, Inc
Linus Torvalds's avatar
Linus Torvalds committed
9 10 11 12
 *
 *  Portions derived from Patrick Mochel's sysfs code.
 *  sysfs is Copyright (c) 2001-3 Patrick Mochel
 *
13
 *  2003-10-10 Written by Simon Derr.
Linus Torvalds's avatar
Linus Torvalds committed
14
 *  2003-10-22 Updates by Stephen Hemminger.
15
 *  2004 May-July Rework by Paul Jackson.
16
 *  2006 Rework by Paul Menage to use generic cgroups
Linus Torvalds's avatar
Linus Torvalds committed
17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
 *
 *  This file is subject to the terms and conditions of the GNU General Public
 *  License.  See the file COPYING in the main directory of the Linux
 *  distribution for more details.
 */

#include <linux/cpu.h>
#include <linux/cpumask.h>
#include <linux/cpuset.h>
#include <linux/err.h>
#include <linux/errno.h>
#include <linux/file.h>
#include <linux/fs.h>
#include <linux/init.h>
#include <linux/interrupt.h>
#include <linux/kernel.h>
#include <linux/kmod.h>
#include <linux/list.h>
35
#include <linux/mempolicy.h>
Linus Torvalds's avatar
Linus Torvalds committed
36 37 38 39 40
#include <linux/mm.h>
#include <linux/module.h>
#include <linux/mount.h>
#include <linux/namei.h>
#include <linux/pagemap.h>
Paul Menage's avatar
Paul Menage committed
41
#include <linux/prio_heap.h>
Linus Torvalds's avatar
Linus Torvalds committed
42
#include <linux/proc_fs.h>
43
#include <linux/rcupdate.h>
Linus Torvalds's avatar
Linus Torvalds committed
44 45
#include <linux/sched.h>
#include <linux/seq_file.h>
46
#include <linux/security.h>
Linus Torvalds's avatar
Linus Torvalds committed
47 48 49 50 51 52 53 54 55 56
#include <linux/slab.h>
#include <linux/spinlock.h>
#include <linux/stat.h>
#include <linux/string.h>
#include <linux/time.h>
#include <linux/backing-dev.h>
#include <linux/sort.h>

#include <asm/uaccess.h>
#include <asm/atomic.h>
57
#include <linux/mutex.h>
Paul Jackson's avatar
Paul Jackson committed
58
#include <linux/kfifo.h>
59 60
#include <linux/workqueue.h>
#include <linux/cgroup.h>
Linus Torvalds's avatar
Linus Torvalds committed
61

62 63 64 65 66
/*
 * Tracks how many cpusets are currently defined in system.
 * When there is only one cpuset (the root cpuset) we can
 * short circuit some hooks.
 */
67
int number_of_cpusets __read_mostly;
68

69 70 71 72
/* Retrieve the cpuset from a cgroup */
struct cgroup_subsys cpuset_subsys;
struct cpuset;

73 74 75 76 77 78 79 80 81
/* See "Frequency meter" comments, below. */

struct fmeter {
	int cnt;		/* unprocessed events count */
	int val;		/* most recent output value */
	time_t time;		/* clock (secs) when val computed */
	spinlock_t lock;	/* guards read or write of above */
};

Linus Torvalds's avatar
Linus Torvalds committed
82
struct cpuset {
83 84
	struct cgroup_subsys_state css;

Linus Torvalds's avatar
Linus Torvalds committed
85 86 87 88 89 90 91 92 93 94
	unsigned long flags;		/* "unsigned long" so bitops work */
	cpumask_t cpus_allowed;		/* CPUs allowed to tasks in cpuset */
	nodemask_t mems_allowed;	/* Memory Nodes allowed to tasks */

	struct cpuset *parent;		/* my parent */

	/*
	 * Copy of global cpuset_mems_generation as of the most
	 * recent time this cpuset changed its mems_allowed.
	 */
95 96 97
	int mems_generation;

	struct fmeter fmeter;		/* memory_pressure filter */
Paul Jackson's avatar
Paul Jackson committed
98 99 100

	/* partition number for rebuild_sched_domains() */
	int pn;
101 102 103

	/* used for walking a cpuset heirarchy */
	struct list_head stack_list;
Linus Torvalds's avatar
Linus Torvalds committed
104 105
};

106 107 108 109 110 111 112 113 114 115 116 117 118
/* Retrieve the cpuset for a cgroup */
static inline struct cpuset *cgroup_cs(struct cgroup *cont)
{
	return container_of(cgroup_subsys_state(cont, cpuset_subsys_id),
			    struct cpuset, css);
}

/* Retrieve the cpuset for a task */
static inline struct cpuset *task_cs(struct task_struct *task)
{
	return container_of(task_subsys_state(task, cpuset_subsys_id),
			    struct cpuset, css);
}
119 120 121 122
struct cpuset_hotplug_scanner {
	struct cgroup_scanner scan;
	struct cgroup *to;
};
123

Linus Torvalds's avatar
Linus Torvalds committed
124 125 126 127
/* bits in struct cpuset flags field */
typedef enum {
	CS_CPU_EXCLUSIVE,
	CS_MEM_EXCLUSIVE,
128
	CS_MEMORY_MIGRATE,
Paul Jackson's avatar
Paul Jackson committed
129
	CS_SCHED_LOAD_BALANCE,
130 131
	CS_SPREAD_PAGE,
	CS_SPREAD_SLAB,
Linus Torvalds's avatar
Linus Torvalds committed
132 133 134 135 136
} cpuset_flagbits_t;

/* convenient tests for these bits */
static inline int is_cpu_exclusive(const struct cpuset *cs)
{
137
	return test_bit(CS_CPU_EXCLUSIVE, &cs->flags);
Linus Torvalds's avatar
Linus Torvalds committed
138 139 140 141
}

static inline int is_mem_exclusive(const struct cpuset *cs)
{
142
	return test_bit(CS_MEM_EXCLUSIVE, &cs->flags);
Linus Torvalds's avatar
Linus Torvalds committed
143 144
}

Paul Jackson's avatar
Paul Jackson committed
145 146 147 148 149
static inline int is_sched_load_balance(const struct cpuset *cs)
{
	return test_bit(CS_SCHED_LOAD_BALANCE, &cs->flags);
}

150 151
static inline int is_memory_migrate(const struct cpuset *cs)
{
152
	return test_bit(CS_MEMORY_MIGRATE, &cs->flags);
153 154
}

155 156 157 158 159 160 161 162 163 164
static inline int is_spread_page(const struct cpuset *cs)
{
	return test_bit(CS_SPREAD_PAGE, &cs->flags);
}

static inline int is_spread_slab(const struct cpuset *cs)
{
	return test_bit(CS_SPREAD_SLAB, &cs->flags);
}

Linus Torvalds's avatar
Linus Torvalds committed
165
/*
166
 * Increment this integer everytime any cpuset changes its
Linus Torvalds's avatar
Linus Torvalds committed
167 168 169 170 171 172 173 174 175 176 177 178 179
 * mems_allowed value.  Users of cpusets can track this generation
 * number, and avoid having to lock and reload mems_allowed unless
 * the cpuset they're using changes generation.
 *
 * A single, global generation is needed because attach_task() could
 * reattach a task to a different cpuset, which must not have its
 * generation numbers aliased with those of that tasks previous cpuset.
 *
 * Generations are needed for mems_allowed because one task cannot
 * modify anothers memory placement.  So we must enable every task,
 * on every visit to __alloc_pages(), to efficiently check whether
 * its current->cpuset->mems_allowed has changed, requiring an update
 * of its current->mems_allowed.
180 181 182
 *
 * Since cpuset_mems_generation is guarded by manage_mutex,
 * there is no need to mark it atomic.
Linus Torvalds's avatar
Linus Torvalds committed
183
 */
184
static int cpuset_mems_generation;
Linus Torvalds's avatar
Linus Torvalds committed
185 186 187 188 189 190 191 192

static struct cpuset top_cpuset = {
	.flags = ((1 << CS_CPU_EXCLUSIVE) | (1 << CS_MEM_EXCLUSIVE)),
	.cpus_allowed = CPU_MASK_ALL,
	.mems_allowed = NODE_MASK_ALL,
};

/*
193 194
 * We have two global cpuset mutexes below.  They can nest.
 * It is ok to first take manage_mutex, then nest callback_mutex.  We also
195 196 197
 * require taking task_lock() when dereferencing a tasks cpuset pointer.
 * See "The task_lock() exception", at the end of this comment.
 *
198 199 200
 * A task must hold both mutexes to modify cpusets.  If a task
 * holds manage_mutex, then it blocks others wanting that mutex,
 * ensuring that it is the only task able to also acquire callback_mutex
201 202
 * and be able to modify cpusets.  It can perform various checks on
 * the cpuset structure first, knowing nothing will change.  It can
203
 * also allocate memory while just holding manage_mutex.  While it is
204
 * performing these checks, various callback routines can briefly
205 206
 * acquire callback_mutex to query cpusets.  Once it is ready to make
 * the changes, it takes callback_mutex, blocking everyone else.
207 208
 *
 * Calls to the kernel memory allocator can not be made while holding
209
 * callback_mutex, as that would risk double tripping on callback_mutex
210 211 212
 * from one of the callbacks into the cpuset code from within
 * __alloc_pages().
 *
213
 * If a task is only holding callback_mutex, then it has read-only
214 215 216 217 218 219
 * access to cpusets.
 *
 * The task_struct fields mems_allowed and mems_generation may only
 * be accessed in the context of that task, so require no locks.
 *
 * Any task can increment and decrement the count field without lock.
220
 * So in general, code holding manage_mutex or callback_mutex can't rely
221
 * on the count field not changing.  However, if the count goes to
222
 * zero, then only attach_task(), which holds both mutexes, can
223 224 225
 * increment it again.  Because a count of zero means that no tasks
 * are currently attached, therefore there is no way a task attached
 * to that cpuset can fork (the other way to increment the count).
226
 * So code holding manage_mutex or callback_mutex can safely assume that
227
 * if the count is zero, it will stay zero.  Similarly, if a task
228
 * holds manage_mutex or callback_mutex on a cpuset with zero count, it
229
 * knows that the cpuset won't be removed, as cpuset_rmdir() needs
230
 * both of those mutexes.
231 232
 *
 * The cpuset_common_file_write handler for operations that modify
233
 * the cpuset hierarchy holds manage_mutex across the entire operation,
234 235
 * single threading all such cpuset modifications across the system.
 *
236
 * The cpuset_common_file_read() handlers only hold callback_mutex across
237 238 239 240
 * small pieces of code, such as when reading out possibly multi-word
 * cpumasks and nodemasks.
 *
 * The fork and exit callbacks cpuset_fork() and cpuset_exit(), don't
241
 * (usually) take either mutex.  These are the two most performance
242
 * critical pieces of code here.  The exception occurs on cpuset_exit(),
243
 * when a task in a notify_on_release cpuset exits.  Then manage_mutex
244
 * is taken, and if the cpuset count is zero, a usermode call made
Linus Torvalds's avatar
Linus Torvalds committed
245 246 247
 * to /sbin/cpuset_release_agent with the name of the cpuset (path
 * relative to the root of cpuset file system) as the argument.
 *
248 249 250
 * A cpuset can only be deleted if both its 'count' of using tasks
 * is zero, and its list of 'children' cpusets is empty.  Since all
 * tasks in the system use _some_ cpuset, and since there is always at
251
 * least one task in the system (init), therefore, top_cpuset
252 253 254 255 256 257 258 259 260
 * always has either children cpusets and/or using tasks.  So we don't
 * need a special hack to ensure that top_cpuset cannot be deleted.
 *
 * The above "Tale of Two Semaphores" would be complete, but for:
 *
 *	The task_lock() exception
 *
 * The need for this exception arises from the action of attach_task(),
 * which overwrites one tasks cpuset pointer with another.  It does
261
 * so using both mutexes, however there are several performance
262
 * critical places that need to reference task->cpuset without the
263
 * expense of grabbing a system global mutex.  Therefore except as
264 265 266 267
 * noted below, when dereferencing or, as in attach_task(), modifying
 * a tasks cpuset pointer we use task_lock(), which acts on a spinlock
 * (task->alloc_lock) already in the task_struct routinely used for
 * such matters.
268 269 270 271 272
 *
 * P.S.  One more locking exception.  RCU is used to guard the
 * update of a tasks cpuset pointer by attach_task() and the
 * access of task->cpuset->mems_generation via that pointer in
 * the routine cpuset_update_task_memory_state().
Linus Torvalds's avatar
Linus Torvalds committed
273 274
 */

275
static DEFINE_MUTEX(callback_mutex);
276

277 278 279
/* This is ugly, but preserves the userspace API for existing cpuset
 * users. If someone tries to mount the "cpuset" filesystem, we
 * silently switch it to mount "cgroup" instead */
280 281 282
static int cpuset_get_sb(struct file_system_type *fs_type,
			 int flags, const char *unused_dev_name,
			 void *data, struct vfsmount *mnt)
Linus Torvalds's avatar
Linus Torvalds committed
283
{
284 285 286 287 288 289 290 291 292 293 294
	struct file_system_type *cgroup_fs = get_fs_type("cgroup");
	int ret = -ENODEV;
	if (cgroup_fs) {
		char mountopts[] =
			"cpuset,noprefix,"
			"release_agent=/sbin/cpuset_release_agent";
		ret = cgroup_fs->get_sb(cgroup_fs, flags,
					   unused_dev_name, mountopts, mnt);
		put_filesystem(cgroup_fs);
	}
	return ret;
Linus Torvalds's avatar
Linus Torvalds committed
295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312
}

static struct file_system_type cpuset_fs_type = {
	.name = "cpuset",
	.get_sb = cpuset_get_sb,
};

/*
 * Return in *pmask the portion of a cpusets's cpus_allowed that
 * are online.  If none are online, walk up the cpuset hierarchy
 * until we find one that does have some online cpus.  If we get
 * all the way to the top and still haven't found any online cpus,
 * return cpu_online_map.  Or if passed a NULL cs from an exit'ing
 * task, return cpu_online_map.
 *
 * One way or another, we guarantee to return some non-empty subset
 * of cpu_online_map.
 *
313
 * Call with callback_mutex held.
Linus Torvalds's avatar
Linus Torvalds committed
314 315 316 317 318 319 320 321 322 323 324 325 326 327 328
 */

static void guarantee_online_cpus(const struct cpuset *cs, cpumask_t *pmask)
{
	while (cs && !cpus_intersects(cs->cpus_allowed, cpu_online_map))
		cs = cs->parent;
	if (cs)
		cpus_and(*pmask, cs->cpus_allowed, cpu_online_map);
	else
		*pmask = cpu_online_map;
	BUG_ON(!cpus_intersects(*pmask, cpu_online_map));
}

/*
 * Return in *pmask the portion of a cpusets's mems_allowed that
329 330 331 332
 * are online, with memory.  If none are online with memory, walk
 * up the cpuset hierarchy until we find one that does have some
 * online mems.  If we get all the way to the top and still haven't
 * found any online mems, return node_states[N_HIGH_MEMORY].
Linus Torvalds's avatar
Linus Torvalds committed
333 334
 *
 * One way or another, we guarantee to return some non-empty subset
335
 * of node_states[N_HIGH_MEMORY].
Linus Torvalds's avatar
Linus Torvalds committed
336
 *
337
 * Call with callback_mutex held.
Linus Torvalds's avatar
Linus Torvalds committed
338 339 340 341
 */

static void guarantee_online_mems(const struct cpuset *cs, nodemask_t *pmask)
{
342 343
	while (cs && !nodes_intersects(cs->mems_allowed,
					node_states[N_HIGH_MEMORY]))
Linus Torvalds's avatar
Linus Torvalds committed
344 345
		cs = cs->parent;
	if (cs)
346 347
		nodes_and(*pmask, cs->mems_allowed,
					node_states[N_HIGH_MEMORY]);
Linus Torvalds's avatar
Linus Torvalds committed
348
	else
349 350
		*pmask = node_states[N_HIGH_MEMORY];
	BUG_ON(!nodes_intersects(*pmask, node_states[N_HIGH_MEMORY]));
Linus Torvalds's avatar
Linus Torvalds committed
351 352
}

353 354 355 356 357 358
/**
 * cpuset_update_task_memory_state - update task memory placement
 *
 * If the current tasks cpusets mems_allowed changed behind our
 * backs, update current->mems_allowed, mems_generation and task NUMA
 * mempolicy to the new value.
359
 *
360 361 362 363
 * Task mempolicy is updated by rebinding it relative to the
 * current->cpuset if a task has its memory placement changed.
 * Do not call this routine if in_interrupt().
 *
364 365 366 367
 * Call without callback_mutex or task_lock() held.  May be
 * called with or without manage_mutex held.  Thanks in part to
 * 'the_top_cpuset_hack', the tasks cpuset pointer will never
 * be NULL.  This routine also might acquire callback_mutex and
368
 * current->mm->mmap_sem during call.
369
 *
370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387
 * Reading current->cpuset->mems_generation doesn't need task_lock
 * to guard the current->cpuset derefence, because it is guarded
 * from concurrent freeing of current->cpuset by attach_task(),
 * using RCU.
 *
 * The rcu_dereference() is technically probably not needed,
 * as I don't actually mind if I see a new cpuset pointer but
 * an old value of mems_generation.  However this really only
 * matters on alpha systems using cpusets heavily.  If I dropped
 * that rcu_dereference(), it would save them a memory barrier.
 * For all other arch's, rcu_dereference is a no-op anyway, and for
 * alpha systems not using cpusets, another planned optimization,
 * avoiding the rcu critical section for tasks in the root cpuset
 * which is statically allocated, so can't vanish, will make this
 * irrelevant.  Better to use RCU as intended, than to engage in
 * some cute trick to save a memory barrier that is impossible to
 * test, for alpha systems using cpusets heavily, which might not
 * even exist.
388 389 390 391 392
 *
 * This routine is needed to update the per-task mems_allowed data,
 * within the tasks context, when it is trying to allocate memory
 * (in various mm/mempolicy.c routines) and notices that some other
 * task has been modifying its cpuset.
Linus Torvalds's avatar
Linus Torvalds committed
393 394
 */

395
void cpuset_update_task_memory_state(void)
Linus Torvalds's avatar
Linus Torvalds committed
396
{
397
	int my_cpusets_mem_gen;
398
	struct task_struct *tsk = current;
399
	struct cpuset *cs;
400

401
	if (task_cs(tsk) == &top_cpuset) {
402 403 404 405
		/* Don't need rcu for top_cpuset.  It's never freed. */
		my_cpusets_mem_gen = top_cpuset.mems_generation;
	} else {
		rcu_read_lock();
406
		my_cpusets_mem_gen = task_cs(current)->mems_generation;
407 408
		rcu_read_unlock();
	}
Linus Torvalds's avatar
Linus Torvalds committed
409

410
	if (my_cpusets_mem_gen != tsk->cpuset_mems_generation) {
411
		mutex_lock(&callback_mutex);
412
		task_lock(tsk);
413
		cs = task_cs(tsk); /* Maybe changed when task not locked */
414 415
		guarantee_online_mems(cs, &tsk->mems_allowed);
		tsk->cpuset_mems_generation = cs->mems_generation;
416 417 418 419 420 421 422 423
		if (is_spread_page(cs))
			tsk->flags |= PF_SPREAD_PAGE;
		else
			tsk->flags &= ~PF_SPREAD_PAGE;
		if (is_spread_slab(cs))
			tsk->flags |= PF_SPREAD_SLAB;
		else
			tsk->flags &= ~PF_SPREAD_SLAB;
424
		task_unlock(tsk);
425
		mutex_unlock(&callback_mutex);
426
		mpol_rebind_task(tsk, &tsk->mems_allowed);
Linus Torvalds's avatar
Linus Torvalds committed
427 428 429 430 431 432 433 434
	}
}

/*
 * is_cpuset_subset(p, q) - Is cpuset p a subset of cpuset q?
 *
 * One cpuset is a subset of another if all its allowed CPUs and
 * Memory Nodes are a subset of the other, and its exclusive flags
435
 * are only set if the other's are set.  Call holding manage_mutex.
Linus Torvalds's avatar
Linus Torvalds committed
436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452
 */

static int is_cpuset_subset(const struct cpuset *p, const struct cpuset *q)
{
	return	cpus_subset(p->cpus_allowed, q->cpus_allowed) &&
		nodes_subset(p->mems_allowed, q->mems_allowed) &&
		is_cpu_exclusive(p) <= is_cpu_exclusive(q) &&
		is_mem_exclusive(p) <= is_mem_exclusive(q);
}

/*
 * validate_change() - Used to validate that any proposed cpuset change
 *		       follows the structural rules for cpusets.
 *
 * If we replaced the flag and mask values of the current cpuset
 * (cur) with those values in the trial cpuset (trial), would
 * our various subset and exclusive rules still be valid?  Presumes
453
 * manage_mutex held.
Linus Torvalds's avatar
Linus Torvalds committed
454 455 456 457 458 459 460 461 462 463 464 465 466 467
 *
 * 'cur' is the address of an actual, in-use cpuset.  Operations
 * such as list traversal that depend on the actual address of the
 * cpuset in the list must use cur below, not trial.
 *
 * 'trial' is the address of bulk structure copy of cur, with
 * perhaps one or more of the fields cpus_allowed, mems_allowed,
 * or flags changed to new, trial values.
 *
 * Return 0 if valid, -errno if not.
 */

static int validate_change(const struct cpuset *cur, const struct cpuset *trial)
{
468
	struct cgroup *cont;
Linus Torvalds's avatar
Linus Torvalds committed
469 470 471
	struct cpuset *c, *par;

	/* Each of our child cpusets must be a subset of us */
472 473
	list_for_each_entry(cont, &cur->css.cgroup->children, sibling) {
		if (!is_cpuset_subset(cgroup_cs(cont), trial))
Linus Torvalds's avatar
Linus Torvalds committed
474 475 476 477
			return -EBUSY;
	}

	/* Remaining checks don't apply to root cpuset */
478
	if (cur == &top_cpuset)
Linus Torvalds's avatar
Linus Torvalds committed
479 480
		return 0;

481 482
	par = cur->parent;

Linus Torvalds's avatar
Linus Torvalds committed
483 484 485 486 487
	/* We must be a subset of our parent cpuset */
	if (!is_cpuset_subset(trial, par))
		return -EACCES;

	/* If either I or some sibling (!= me) is exclusive, we can't overlap */
488 489
	list_for_each_entry(cont, &par->css.cgroup->children, sibling) {
		c = cgroup_cs(cont);
Linus Torvalds's avatar
Linus Torvalds committed
490 491 492 493 494 495 496 497 498 499
		if ((is_cpu_exclusive(trial) || is_cpu_exclusive(c)) &&
		    c != cur &&
		    cpus_intersects(trial->cpus_allowed, c->cpus_allowed))
			return -EINVAL;
		if ((is_mem_exclusive(trial) || is_mem_exclusive(c)) &&
		    c != cur &&
		    nodes_intersects(trial->mems_allowed, c->mems_allowed))
			return -EINVAL;
	}

500 501 502 503 504 505 506 507
	/* Cpusets with tasks can't have empty cpus_allowed or mems_allowed */
	if (cgroup_task_count(cur->css.cgroup)) {
		if (cpus_empty(trial->cpus_allowed) ||
		    nodes_empty(trial->mems_allowed)) {
			return -ENOSPC;
		}
	}

Linus Torvalds's avatar
Linus Torvalds committed
508 509 510
	return 0;
}

Paul Jackson's avatar
Paul Jackson committed
511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547
/*
 * Helper routine for rebuild_sched_domains().
 * Do cpusets a, b have overlapping cpus_allowed masks?
 */

static int cpusets_overlap(struct cpuset *a, struct cpuset *b)
{
	return cpus_intersects(a->cpus_allowed, b->cpus_allowed);
}

/*
 * rebuild_sched_domains()
 *
 * If the flag 'sched_load_balance' of any cpuset with non-empty
 * 'cpus' changes, or if the 'cpus' allowed changes in any cpuset
 * which has that flag enabled, or if any cpuset with a non-empty
 * 'cpus' is removed, then call this routine to rebuild the
 * scheduler's dynamic sched domains.
 *
 * This routine builds a partial partition of the systems CPUs
 * (the set of non-overlappping cpumask_t's in the array 'part'
 * below), and passes that partial partition to the kernel/sched.c
 * partition_sched_domains() routine, which will rebuild the
 * schedulers load balancing domains (sched domains) as specified
 * by that partial partition.  A 'partial partition' is a set of
 * non-overlapping subsets whose union is a subset of that set.
 *
 * See "What is sched_load_balance" in Documentation/cpusets.txt
 * for a background explanation of this.
 *
 * Does not return errors, on the theory that the callers of this
 * routine would rather not worry about failures to rebuild sched
 * domains when operating in the severe memory shortage situations
 * that could cause allocation failures below.
 *
 * Call with cgroup_mutex held.  May take callback_mutex during
 * call due to the kfifo_alloc() and kmalloc() calls.  May nest
548
 * a call to the get_online_cpus()/put_online_cpus() pair.
Paul Jackson's avatar
Paul Jackson committed
549
 * Must not be called holding callback_mutex, because we must not
550 551
 * call get_online_cpus() while holding callback_mutex.  Elsewhere
 * the kernel nests callback_mutex inside get_online_cpus() calls.
Paul Jackson's avatar
Paul Jackson committed
552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701
 * So the reverse nesting would risk an ABBA deadlock.
 *
 * The three key local variables below are:
 *    q  - a kfifo queue of cpuset pointers, used to implement a
 *	   top-down scan of all cpusets.  This scan loads a pointer
 *	   to each cpuset marked is_sched_load_balance into the
 *	   array 'csa'.  For our purposes, rebuilding the schedulers
 *	   sched domains, we can ignore !is_sched_load_balance cpusets.
 *  csa  - (for CpuSet Array) Array of pointers to all the cpusets
 *	   that need to be load balanced, for convenient iterative
 *	   access by the subsequent code that finds the best partition,
 *	   i.e the set of domains (subsets) of CPUs such that the
 *	   cpus_allowed of every cpuset marked is_sched_load_balance
 *	   is a subset of one of these domains, while there are as
 *	   many such domains as possible, each as small as possible.
 * doms  - Conversion of 'csa' to an array of cpumasks, for passing to
 *	   the kernel/sched.c routine partition_sched_domains() in a
 *	   convenient format, that can be easily compared to the prior
 *	   value to determine what partition elements (sched domains)
 *	   were changed (added or removed.)
 *
 * Finding the best partition (set of domains):
 *	The triple nested loops below over i, j, k scan over the
 *	load balanced cpusets (using the array of cpuset pointers in
 *	csa[]) looking for pairs of cpusets that have overlapping
 *	cpus_allowed, but which don't have the same 'pn' partition
 *	number and gives them in the same partition number.  It keeps
 *	looping on the 'restart' label until it can no longer find
 *	any such pairs.
 *
 *	The union of the cpus_allowed masks from the set of
 *	all cpusets having the same 'pn' value then form the one
 *	element of the partition (one sched domain) to be passed to
 *	partition_sched_domains().
 */

static void rebuild_sched_domains(void)
{
	struct kfifo *q;	/* queue of cpusets to be scanned */
	struct cpuset *cp;	/* scans q */
	struct cpuset **csa;	/* array of all cpuset ptrs */
	int csn;		/* how many cpuset ptrs in csa so far */
	int i, j, k;		/* indices for partition finding loops */
	cpumask_t *doms;	/* resulting partition; i.e. sched domains */
	int ndoms;		/* number of sched domains in result */
	int nslot;		/* next empty doms[] cpumask_t slot */

	q = NULL;
	csa = NULL;
	doms = NULL;

	/* Special case for the 99% of systems with one, full, sched domain */
	if (is_sched_load_balance(&top_cpuset)) {
		ndoms = 1;
		doms = kmalloc(sizeof(cpumask_t), GFP_KERNEL);
		if (!doms)
			goto rebuild;
		*doms = top_cpuset.cpus_allowed;
		goto rebuild;
	}

	q = kfifo_alloc(number_of_cpusets * sizeof(cp), GFP_KERNEL, NULL);
	if (IS_ERR(q))
		goto done;
	csa = kmalloc(number_of_cpusets * sizeof(cp), GFP_KERNEL);
	if (!csa)
		goto done;
	csn = 0;

	cp = &top_cpuset;
	__kfifo_put(q, (void *)&cp, sizeof(cp));
	while (__kfifo_get(q, (void *)&cp, sizeof(cp))) {
		struct cgroup *cont;
		struct cpuset *child;   /* scans child cpusets of cp */
		if (is_sched_load_balance(cp))
			csa[csn++] = cp;
		list_for_each_entry(cont, &cp->css.cgroup->children, sibling) {
			child = cgroup_cs(cont);
			__kfifo_put(q, (void *)&child, sizeof(cp));
		}
  	}

	for (i = 0; i < csn; i++)
		csa[i]->pn = i;
	ndoms = csn;

restart:
	/* Find the best partition (set of sched domains) */
	for (i = 0; i < csn; i++) {
		struct cpuset *a = csa[i];
		int apn = a->pn;

		for (j = 0; j < csn; j++) {
			struct cpuset *b = csa[j];
			int bpn = b->pn;

			if (apn != bpn && cpusets_overlap(a, b)) {
				for (k = 0; k < csn; k++) {
					struct cpuset *c = csa[k];

					if (c->pn == bpn)
						c->pn = apn;
				}
				ndoms--;	/* one less element */
				goto restart;
			}
		}
	}

	/* Convert <csn, csa> to <ndoms, doms> */
	doms = kmalloc(ndoms * sizeof(cpumask_t), GFP_KERNEL);
	if (!doms)
		goto rebuild;

	for (nslot = 0, i = 0; i < csn; i++) {
		struct cpuset *a = csa[i];
		int apn = a->pn;

		if (apn >= 0) {
			cpumask_t *dp = doms + nslot;

			if (nslot == ndoms) {
				static int warnings = 10;
				if (warnings) {
					printk(KERN_WARNING
					 "rebuild_sched_domains confused:"
					  " nslot %d, ndoms %d, csn %d, i %d,"
					  " apn %d\n",
					  nslot, ndoms, csn, i, apn);
					warnings--;
				}
				continue;
			}

			cpus_clear(*dp);
			for (j = i; j < csn; j++) {
				struct cpuset *b = csa[j];

				if (apn == b->pn) {
					cpus_or(*dp, *dp, b->cpus_allowed);
					b->pn = -1;
				}
			}
			nslot++;
		}
	}
	BUG_ON(nslot != ndoms);

rebuild:
	/* Have scheduler rebuild sched domains */
702
	get_online_cpus();
Paul Jackson's avatar
Paul Jackson committed
703
	partition_sched_domains(ndoms, doms);
704
	put_online_cpus();
Paul Jackson's avatar
Paul Jackson committed
705 706 707 708 709 710 711 712

done:
	if (q && !IS_ERR(q))
		kfifo_free(q);
	kfree(csa);
	/* Don't kfree(doms) -- partition_sched_domains() does that. */
}

Paul Menage's avatar
Paul Menage committed
713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742
static inline int started_after_time(struct task_struct *t1,
				     struct timespec *time,
				     struct task_struct *t2)
{
	int start_diff = timespec_compare(&t1->start_time, time);
	if (start_diff > 0) {
		return 1;
	} else if (start_diff < 0) {
		return 0;
	} else {
		/*
		 * Arbitrarily, if two processes started at the same
		 * time, we'll say that the lower pointer value
		 * started first. Note that t2 may have exited by now
		 * so this may not be a valid pointer any longer, but
		 * that's fine - it still serves to distinguish
		 * between two tasks started (effectively)
		 * simultaneously.
		 */
		return t1 > t2;
	}
}

static inline int started_after(void *p1, void *p2)
{
	struct task_struct *t1 = p1;
	struct task_struct *t2 = p2;
	return started_after_time(t1, &t2->start_time, t2);
}

743
/*
744
 * Call with manage_mutex held.  May take callback_mutex during call.
745 746
 */

Linus Torvalds's avatar
Linus Torvalds committed
747 748 749
static int update_cpumask(struct cpuset *cs, char *buf)
{
	struct cpuset trialcs;
Paul Menage's avatar
Paul Menage committed
750 751 752 753 754 755 756 757 758
	int retval, i;
	int is_load_balanced;
	struct cgroup_iter it;
	struct cgroup *cgrp = cs->css.cgroup;
	struct task_struct *p, *dropped;
	/* Never dereference latest_task, since it's not refcounted */
	struct task_struct *latest_task = NULL;
	struct ptr_heap heap;
	struct timespec latest_time = { 0, 0 };
Linus Torvalds's avatar
Linus Torvalds committed
759

760 761 762 763
	/* top_cpuset.cpus_allowed tracks cpu_online_map; it's read-only */
	if (cs == &top_cpuset)
		return -EACCES;

Linus Torvalds's avatar
Linus Torvalds committed
764
	trialcs = *cs;
765 766

	/*
767 768 769 770
	 * An empty cpus_allowed is ok iff there are no tasks in the cpuset.
	 * Since cpulist_parse() fails on an empty mask, we special case
	 * that parsing.  The validate_change() call ensures that cpusets
	 * with tasks have cpus.
771
	 */
772 773
	buf = strstrip(buf);
	if (!*buf) {
774 775 776 777 778 779
		cpus_clear(trialcs.cpus_allowed);
	} else {
		retval = cpulist_parse(buf, trialcs.cpus_allowed);
		if (retval < 0)
			return retval;
	}
Linus Torvalds's avatar
Linus Torvalds committed
780 781
	cpus_and(trialcs.cpus_allowed, trialcs.cpus_allowed, cpu_online_map);
	retval = validate_change(cs, &trialcs);
782 783
	if (retval < 0)
		return retval;
Paul Jackson's avatar
Paul Jackson committed
784

Paul Menage's avatar
Paul Menage committed
785 786 787 788 789 790 791
	/* Nothing to do if the cpus didn't change */
	if (cpus_equal(cs->cpus_allowed, trialcs.cpus_allowed))
		return 0;
	retval = heap_init(&heap, PAGE_SIZE, GFP_KERNEL, &started_after);
	if (retval)
		return retval;

Paul Jackson's avatar
Paul Jackson committed
792 793
	is_load_balanced = is_sched_load_balance(&trialcs);

794
	mutex_lock(&callback_mutex);
795
	cs->cpus_allowed = trialcs.cpus_allowed;
796
	mutex_unlock(&callback_mutex);
Paul Jackson's avatar
Paul Jackson committed
797

Paul Menage's avatar
Paul Menage committed
798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851
 again:
	/*
	 * Scan tasks in the cpuset, and update the cpumasks of any
	 * that need an update. Since we can't call set_cpus_allowed()
	 * while holding tasklist_lock, gather tasks to be processed
	 * in a heap structure. If the statically-sized heap fills up,
	 * overflow tasks that started later, and in future iterations
	 * only consider tasks that started after the latest task in
	 * the previous pass. This guarantees forward progress and
	 * that we don't miss any tasks
	 */
	heap.size = 0;
	cgroup_iter_start(cgrp, &it);
	while ((p = cgroup_iter_next(cgrp, &it))) {
		/* Only affect tasks that don't have the right cpus_allowed */
		if (cpus_equal(p->cpus_allowed, cs->cpus_allowed))
			continue;
		/*
		 * Only process tasks that started after the last task
		 * we processed
		 */
		if (!started_after_time(p, &latest_time, latest_task))
			continue;
		dropped = heap_insert(&heap, p);
		if (dropped == NULL) {
			get_task_struct(p);
		} else if (dropped != p) {
			get_task_struct(p);
			put_task_struct(dropped);
		}
	}
	cgroup_iter_end(cgrp, &it);
	if (heap.size) {
		for (i = 0; i < heap.size; i++) {
			struct task_struct *p = heap.ptrs[i];
			if (i == 0) {
				latest_time = p->start_time;
				latest_task = p;
			}
			set_cpus_allowed(p, cs->cpus_allowed);
			put_task_struct(p);
		}
		/*
		 * If we had to process any tasks at all, scan again
		 * in case some of them were in the middle of forking
		 * children that didn't notice the new cpumask
		 * restriction.  Not the most efficient way to do it,
		 * but it avoids having to take callback_mutex in the
		 * fork path
		 */
		goto again;
	}
	heap_free(&heap);
	if (is_load_balanced)
Paul Jackson's avatar
Paul Jackson committed
852 853
		rebuild_sched_domains();

854
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
855 856
}

857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901
/*
 * cpuset_migrate_mm
 *
 *    Migrate memory region from one set of nodes to another.
 *
 *    Temporarilly set tasks mems_allowed to target nodes of migration,
 *    so that the migration code can allocate pages on these nodes.
 *
 *    Call holding manage_mutex, so our current->cpuset won't change
 *    during this call, as manage_mutex holds off any attach_task()
 *    calls.  Therefore we don't need to take task_lock around the
 *    call to guarantee_online_mems(), as we know no one is changing
 *    our tasks cpuset.
 *
 *    Hold callback_mutex around the two modifications of our tasks
 *    mems_allowed to synchronize with cpuset_mems_allowed().
 *
 *    While the mm_struct we are migrating is typically from some
 *    other task, the task_struct mems_allowed that we are hacking
 *    is for our current task, which must allocate new pages for that
 *    migrating memory region.
 *
 *    We call cpuset_update_task_memory_state() before hacking
 *    our tasks mems_allowed, so that we are assured of being in
 *    sync with our tasks cpuset, and in particular, callbacks to
 *    cpuset_update_task_memory_state() from nested page allocations
 *    won't see any mismatch of our cpuset and task mems_generation
 *    values, so won't overwrite our hacked tasks mems_allowed
 *    nodemask.
 */

static void cpuset_migrate_mm(struct mm_struct *mm, const nodemask_t *from,
							const nodemask_t *to)
{
	struct task_struct *tsk = current;

	cpuset_update_task_memory_state();

	mutex_lock(&callback_mutex);
	tsk->mems_allowed = *to;
	mutex_unlock(&callback_mutex);

	do_migrate_pages(mm, from, to, MPOL_MF_MOVE_ALL);

	mutex_lock(&callback_mutex);
902
	guarantee_online_mems(task_cs(tsk),&tsk->mems_allowed);
903 904 905
	mutex_unlock(&callback_mutex);
}

906
/*
907 908 909
 * Handle user request to change the 'mems' memory placement
 * of a cpuset.  Needs to validate the request, update the
 * cpusets mems_allowed and mems_generation, and for each
910 911 912
 * task in the cpuset, rebind any vma mempolicies and if
 * the cpuset is marked 'memory_migrate', migrate the tasks
 * pages to the new memory.
913
 *
914
 * Call with manage_mutex held.  May take callback_mutex during call.
915 916 917
 * Will take tasklist_lock, scan tasklist for tasks in cpuset cs,
 * lock each such tasks mm->mmap_sem, scan its vma's and rebind
 * their mempolicies to the cpusets new mems_allowed.
918 919
 */

920 921
static void *cpuset_being_rebound;

Linus Torvalds's avatar
Linus Torvalds committed
922 923 924
static int update_nodemask(struct cpuset *cs, char *buf)
{
	struct cpuset trialcs;
925
	nodemask_t oldmem;
926
	struct task_struct *p;
927 928
	struct mm_struct **mmarray;
	int i, n, ntasks;
929
	int migrate;
930
	int fudge;
Linus Torvalds's avatar
Linus Torvalds committed
931
	int retval;
932
	struct cgroup_iter it;
Linus Torvalds's avatar
Linus Torvalds committed
933

934 935 936 937
	/*
	 * top_cpuset.mems_allowed tracks node_stats[N_HIGH_MEMORY];
	 * it's read-only
	 */
938 939 940
	if (cs == &top_cpuset)
		return -EACCES;

Linus Torvalds's avatar
Linus Torvalds committed
941
	trialcs = *cs;
942 943

	/*
944 945 946 947
	 * An empty mems_allowed is ok iff there are no tasks in the cpuset.
	 * Since nodelist_parse() fails on an empty mask, we special case
	 * that parsing.  The validate_change() call ensures that cpusets
	 * with tasks have memory.
948
	 */
949 950
	buf = strstrip(buf);
	if (!*buf) {
951 952 953 954 955 956
		nodes_clear(trialcs.mems_allowed);
	} else {
		retval = nodelist_parse(buf, trialcs.mems_allowed);
		if (retval < 0)
			goto done;
	}
957 958
	nodes_and(trialcs.mems_allowed, trialcs.mems_allowed,
						node_states[N_HIGH_MEMORY]);
959 960 961 962 963
	oldmem = cs->mems_allowed;
	if (nodes_equal(oldmem, trialcs.mems_allowed)) {
		retval = 0;		/* Too easy - nothing to do */
		goto done;
	}
964 965 966 967
	retval = validate_change(cs, &trialcs);
	if (retval < 0)
		goto done;

968
	mutex_lock(&callback_mutex);
969
	cs->mems_allowed = trialcs.mems_allowed;
970
	cs->mems_generation = cpuset_mems_generation++;
971
	mutex_unlock(&callback_mutex);
972

973
	cpuset_being_rebound = cs;		/* causes mpol_copy() rebind */
974 975 976 977 978 979 980 981 982 983 984 985 986

	fudge = 10;				/* spare mmarray[] slots */
	fudge += cpus_weight(cs->cpus_allowed);	/* imagine one fork-bomb/cpu */
	retval = -ENOMEM;

	/*
	 * Allocate mmarray[] to hold mm reference for each task
	 * in cpuset cs.  Can't kmalloc GFP_KERNEL while holding
	 * tasklist_lock.  We could use GFP_ATOMIC, but with a
	 * few more lines of code, we can retry until we get a big
	 * enough mmarray[] w/o using GFP_ATOMIC.
	 */
	while (1) {
987
		ntasks = cgroup_task_count(cs->css.cgroup);  /* guess */
988 989 990 991
		ntasks += fudge;
		mmarray = kmalloc(ntasks * sizeof(*mmarray), GFP_KERNEL);
		if (!mmarray)
			goto done;
992
		read_lock(&tasklist_lock);		/* block fork */
993
		if (cgroup_task_count(cs->css.cgroup) <= ntasks)
994
			break;				/* got enough */
995
		read_unlock(&tasklist_lock);		/* try again */
996 997 998 999 1000 1001
		kfree(mmarray);
	}

	n = 0;

	/* Load up mmarray[] with mm reference for each task in cpuset. */
1002 1003
	cgroup_iter_start(cs->css.cgroup, &it);
	while ((p = cgroup_iter_next(cs->css.cgroup, &it))) {
1004 1005 1006 1007 1008
		struct mm_struct *mm;

		if (n >= ntasks) {
			printk(KERN_WARNING
				"Cpuset mempolicy rebind incomplete.\n");
1009
			break;
1010 1011 1012 1013 1014
		}
		mm = get_task_mm(p);
		if (!mm)
			continue;
		mmarray[n++] = mm;
1015 1016
	}
	cgroup_iter_end(cs->css.cgroup, &it);
1017
	read_unlock(&tasklist_lock);
1018 1019 1020 1021 1022 1023 1024 1025 1026

	/*
	 * Now that we've dropped the tasklist spinlock, we can
	 * rebind the vma mempolicies of each mm in mmarray[] to their
	 * new cpuset, and release that mm.  The mpol_rebind_mm()
	 * call takes mmap_sem, which we couldn't take while holding
	 * tasklist_lock.  Forks can happen again now - the mpol_copy()
	 * cpuset_being_rebound check will catch such forks, and rebind
	 * their vma mempolicies too.  Because we still hold the global
1027
	 * cpuset manage_mutex, we know that no other rebind effort will
1028 1029
	 * be contending for the global variable cpuset_being_rebound.
	 * It's ok if we rebind the same mm twice; mpol_rebind_mm()
1030
	 * is idempotent.  Also migrate pages in each mm to new nodes.
1031
	 */
1032
	migrate = is_memory_migrate(cs);
1033 1034 1035 1036
	for (i = 0; i < n; i++) {
		struct mm_struct *mm = mmarray[i];

		mpol_rebind_mm(mm, &cs->mems_allowed);
1037 1038
		if (migrate)
			cpuset_migrate_mm(mm, &oldmem, &cs->mems_allowed);
1039 1040 1041 1042 1043
		mmput(mm);
	}

	/* We're done rebinding vma's to this cpusets new mems_allowed. */
	kfree(mmarray);
1044
	cpuset_being_rebound = NULL;
1045
	retval = 0;
1046
done:
Linus Torvalds's avatar
Linus Torvalds committed
1047 1048 1049
	return retval;
}

1050 1051 1052 1053 1054
int current_cpuset_is_being_rebound(void)
{
	return task_cs(current) == cpuset_being_rebound;
}

1055
/*
1056
 * Call with manage_mutex held.
1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067
 */

static int update_memory_pressure_enabled(struct cpuset *cs, char *buf)
{
	if (simple_strtoul(buf, NULL, 10) != 0)
		cpuset_memory_pressure_enabled = 1;
	else
		cpuset_memory_pressure_enabled = 0;
	return 0;
}

Linus Torvalds's avatar
Linus Torvalds committed
1068 1069 1070
/*
 * update_flag - read a 0 or a 1 in a file and update associated flag
 * bit:	the bit to update (CS_CPU_EXCLUSIVE, CS_MEM_EXCLUSIVE,
Paul Jackson's avatar
Paul Jackson committed
1071
 *				CS_SCHED_LOAD_BALANCE,
1072 1073
 *				CS_NOTIFY_ON_RELEASE, CS_MEMORY_MIGRATE,
 *				CS_SPREAD_PAGE, CS_SPREAD_SLAB)
Linus Torvalds's avatar
Linus Torvalds committed
1074 1075
 * cs:	the cpuset to update
 * buf:	the buffer where we read the 0 or 1
1076
 *
1077
 * Call with manage_mutex held.
Linus Torvalds's avatar
Linus Torvalds committed
1078 1079 1080 1081 1082 1083
 */

static int update_flag(cpuset_flagbits_t bit, struct cpuset *cs, char *buf)
{
	int turning_on;
	struct cpuset trialcs;
1084
	int err;
Paul Jackson's avatar
Paul Jackson committed
1085
	int cpus_nonempty, balance_flag_changed;
Linus Torvalds's avatar
Linus Torvalds committed
1086 1087 1088 1089 1090 1091 1092 1093 1094 1095

	turning_on = (simple_strtoul(buf, NULL, 10) != 0);

	trialcs = *cs;
	if (turning_on)
		set_bit(bit, &trialcs.flags);
	else
		clear_bit(bit, &trialcs.flags);

	err = validate_change(cs, &trialcs);
1096 1097
	if (err < 0)
		return err;
Paul Jackson's avatar
Paul Jackson committed
1098 1099 1100 1101 1102

	cpus_nonempty = !cpus_empty(trialcs.cpus_allowed);
	balance_flag_changed = (is_sched_load_balance(cs) !=
		 			is_sched_load_balance(&trialcs));

1103
	mutex_lock(&callback_mutex);
1104
	cs->flags = trialcs.flags;
1105
	mutex_unlock(&callback_mutex);
1106

Paul Jackson's avatar
Paul Jackson committed
1107 1108 1109
	if (cpus_nonempty && balance_flag_changed)
		rebuild_sched_domains();

1110
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
1111 1112
}

1113
/*
1114
 * Frequency meter - How fast is some event occurring?
1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210
 *
 * These routines manage a digitally filtered, constant time based,
 * event frequency meter.  There are four routines:
 *   fmeter_init() - initialize a frequency meter.
 *   fmeter_markevent() - called each time the event happens.
 *   fmeter_getrate() - returns the recent rate of such events.
 *   fmeter_update() - internal routine used to update fmeter.
 *
 * A common data structure is passed to each of these routines,
 * which is used to keep track of the state required to manage the
 * frequency meter and its digital filter.
 *
 * The filter works on the number of events marked per unit time.
 * The filter is single-pole low-pass recursive (IIR).  The time unit
 * is 1 second.  Arithmetic is done using 32-bit integers scaled to
 * simulate 3 decimal digits of precision (multiplied by 1000).
 *
 * With an FM_COEF of 933, and a time base of 1 second, the filter
 * has a half-life of 10 seconds, meaning that if the events quit
 * happening, then the rate returned from the fmeter_getrate()
 * will be cut in half each 10 seconds, until it converges to zero.
 *
 * It is not worth doing a real infinitely recursive filter.  If more
 * than FM_MAXTICKS ticks have elapsed since the last filter event,
 * just compute FM_MAXTICKS ticks worth, by which point the level
 * will be stable.
 *
 * Limit the count of unprocessed events to FM_MAXCNT, so as to avoid
 * arithmetic overflow in the fmeter_update() routine.
 *
 * Given the simple 32 bit integer arithmetic used, this meter works
 * best for reporting rates between one per millisecond (msec) and
 * one per 32 (approx) seconds.  At constant rates faster than one
 * per msec it maxes out at values just under 1,000,000.  At constant
 * rates between one per msec, and one per second it will stabilize
 * to a value N*1000, where N is the rate of events per second.
 * At constant rates between one per second and one per 32 seconds,
 * it will be choppy, moving up on the seconds that have an event,
 * and then decaying until the next event.  At rates slower than
 * about one in 32 seconds, it decays all the way back to zero between
 * each event.
 */

#define FM_COEF 933		/* coefficient for half-life of 10 secs */
#define FM_MAXTICKS ((time_t)99) /* useless computing more ticks than this */
#define FM_MAXCNT 1000000	/* limit cnt to avoid overflow */
#define FM_SCALE 1000		/* faux fixed point scale */

/* Initialize a frequency meter */
static void fmeter_init(struct fmeter *fmp)
{
	fmp->cnt = 0;
	fmp->val = 0;
	fmp->time = 0;
	spin_lock_init(&fmp->lock);
}

/* Internal meter update - process cnt events and update value */
static void fmeter_update(struct fmeter *fmp)
{
	time_t now = get_seconds();
	time_t ticks = now - fmp->time;

	if (ticks == 0)
		return;

	ticks = min(FM_MAXTICKS, ticks);
	while (ticks-- > 0)
		fmp->val = (FM_COEF * fmp->val) / FM_SCALE;
	fmp->time = now;

	fmp->val += ((FM_SCALE - FM_COEF) * fmp->cnt) / FM_SCALE;
	fmp->cnt = 0;
}

/* Process any previous ticks, then bump cnt by one (times scale). */
static void fmeter_markevent(struct fmeter *fmp)
{
	spin_lock(&fmp->lock);
	fmeter_update(fmp);
	fmp->cnt = min(FM_MAXCNT, fmp->cnt + FM_SCALE);
	spin_unlock(&fmp->lock);
}

/* Process any previous ticks, then return current value. */
static int fmeter_getrate(struct fmeter *fmp)
{
	int val;

	spin_lock(&fmp->lock);
	fmeter_update(fmp);
	val = fmp->val;
	spin_unlock(&fmp->lock);
	return val;
}

1211 1212
static int cpuset_can_attach(struct cgroup_subsys *ss,
			     struct cgroup *cont, struct task_struct *tsk)
Linus Torvalds's avatar
Linus Torvalds committed
1213
{
1214
	struct cpuset *cs = cgroup_cs(cont);
Linus Torvalds's avatar
Linus Torvalds committed
1215 1216 1217 1218

	if (cpus_empty(cs->cpus_allowed) || nodes_empty(cs->mems_allowed))
		return -ENOSPC;

1219 1220
	return security_task_setscheduler(tsk, 0, NULL);
}
Linus Torvalds's avatar
Linus Torvalds committed
1221

1222 1223 1224 1225 1226 1227 1228 1229 1230
static void cpuset_attach(struct cgroup_subsys *ss,
			  struct cgroup *cont, struct cgroup *oldcont,
			  struct task_struct *tsk)
{
	cpumask_t cpus;
	nodemask_t from, to;
	struct mm_struct *mm;
	struct cpuset *cs = cgroup_cs(cont);
	struct cpuset *oldcs = cgroup_cs(oldcont);
1231

1232
	mutex_lock(&callback_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
1233 1234
	guarantee_online_cpus(cs, &cpus);
	set_cpus_allowed(tsk, cpus);
1235
	mutex_unlock(&callback_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
1236

1237 1238
	from = oldcs->mems_allowed;
	to = cs->mems_allowed;
1239 1240 1241
	mm = get_task_mm(tsk);
	if (mm) {
		mpol_rebind_mm(mm, &to);
1242
		if (is_memory_migrate(cs))
1243
			cpuset_migrate_mm(mm, &from, &to);
1244 1245 1246
		mmput(mm);
	}

Linus Torvalds's avatar
Linus Torvalds committed
1247 1248 1249 1250 1251
}

/* The various types of files and directories in a cpuset file system */

typedef enum {
1252
	FILE_MEMORY_MIGRATE,
Linus Torvalds's avatar
Linus Torvalds committed
1253 1254 1255 1256
	FILE_CPULIST,
	FILE_MEMLIST,
	FILE_CPU_EXCLUSIVE,
	FILE_MEM_EXCLUSIVE,
Paul Jackson's avatar
Paul Jackson committed
1257
	FILE_SCHED_LOAD_BALANCE,
1258 1259
	FILE_MEMORY_PRESSURE_ENABLED,
	FILE_MEMORY_PRESSURE,
1260 1261
	FILE_SPREAD_PAGE,
	FILE_SPREAD_SLAB,
Linus Torvalds's avatar
Linus Torvalds committed
1262 1263
} cpuset_filetype_t;

1264 1265 1266
static ssize_t cpuset_common_file_write(struct cgroup *cont,
					struct cftype *cft,
					struct file *file,
1267
					const char __user *userbuf,
Linus Torvalds's avatar
Linus Torvalds committed
1268 1269
					size_t nbytes, loff_t *unused_ppos)
{
1270
	struct cpuset *cs = cgroup_cs(cont);
Linus Torvalds's avatar
Linus Torvalds committed
1271 1272 1273 1274 1275
	cpuset_filetype_t type = cft->private;
	char *buffer;
	int retval = 0;

	/* Crude upper limit on largest legitimate cpulist user might write. */
Paul Jackson's avatar
Paul Jackson committed
1276
	if (nbytes > 100U + 6 * max(NR_CPUS, MAX_NUMNODES))
Linus Torvalds's avatar
Linus Torvalds committed
1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288
		return -E2BIG;

	/* +1 for nul-terminator */
	if ((buffer = kmalloc(nbytes + 1, GFP_KERNEL)) == 0)
		return -ENOMEM;

	if (copy_from_user(buffer, userbuf, nbytes)) {
		retval = -EFAULT;
		goto out1;
	}
	buffer[nbytes] = 0;	/* nul-terminate */

1289
	cgroup_lock();
Linus Torvalds's avatar
Linus Torvalds committed
1290

1291
	if (cgroup_is_removed(cont)) {
Linus Torvalds's avatar
Linus Torvalds committed
1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308
		retval = -ENODEV;
		goto out2;
	}

	switch (type) {
	case FILE_CPULIST:
		retval = update_cpumask(cs, buffer);
		break;
	case FILE_MEMLIST:
		retval = update_nodemask(cs, buffer);
		break;
	case FILE_CPU_EXCLUSIVE:
		retval = update_flag(CS_CPU_EXCLUSIVE, cs, buffer);
		break;
	case FILE_MEM_EXCLUSIVE:
		retval = update_flag(CS_MEM_EXCLUSIVE, cs, buffer);
		break;
Paul Jackson's avatar
Paul Jackson committed
1309 1310 1311
	case FILE_SCHED_LOAD_BALANCE:
		retval = update_flag(CS_SCHED_LOAD_BALANCE, cs, buffer);
		break;
1312 1313 1314
	case FILE_MEMORY_MIGRATE:
		retval = update_flag(CS_MEMORY_MIGRATE, cs, buffer);
		break;
1315 1316 1317 1318 1319 1320
	case FILE_MEMORY_PRESSURE_ENABLED:
		retval = update_memory_pressure_enabled(cs, buffer);
		break;
	case FILE_MEMORY_PRESSURE:
		retval = -EACCES;
		break;
1321 1322
	case FILE_SPREAD_PAGE:
		retval = update_flag(CS_SPREAD_PAGE, cs, buffer);
1323
		cs->mems_generation = cpuset_mems_generation++;
1324 1325 1326
		break;
	case FILE_SPREAD_SLAB:
		retval = update_flag(CS_SPREAD_SLAB, cs, buffer);
1327
		cs->mems_generation = cpuset_mems_generation++;
1328
		break;
Linus Torvalds's avatar
Linus Torvalds committed
1329 1330 1331 1332 1333 1334 1335 1336
	default:
		retval = -EINVAL;
		goto out2;
	}

	if (retval == 0)
		retval = nbytes;
out2:
1337
	cgroup_unlock();
Linus Torvalds's avatar
Linus Torvalds committed
1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358
out1:
	kfree(buffer);
	return retval;
}

/*
 * These ascii lists should be read in a single call, by using a user
 * buffer large enough to hold the entire map.  If read in smaller
 * chunks, there is no guarantee of atomicity.  Since the display format
 * used, list of ranges of sequential numbers, is variable length,
 * and since these maps can change value dynamically, one could read
 * gibberish by doing partial reads while a list was changing.
 * A single large read to a buffer that crosses a page boundary is
 * ok, because the result being copied to user land is not recomputed
 * across a page fault.
 */

static int cpuset_sprintf_cpulist(char *page, struct cpuset *cs)
{
	cpumask_t mask;

1359
	mutex_lock(&callback_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
1360
	mask = cs->cpus_allowed;
1361
	mutex_unlock(&callback_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
1362 1363 1364 1365 1366 1367 1368 1369

	return cpulist_scnprintf(page, PAGE_SIZE, mask);
}

static int cpuset_sprintf_memlist(char *page, struct cpuset *cs)
{
	nodemask_t mask;

1370
	mutex_lock(&callback_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
1371
	mask = cs->mems_allowed;
1372
	mutex_unlock(&callback_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
1373 1374 1375 1376

	return nodelist_scnprintf(page, PAGE_SIZE, mask);
}

1377 1378 1379 1380 1381
static ssize_t cpuset_common_file_read(struct cgroup *cont,
				       struct cftype *cft,
				       struct file *file,
				       char __user *buf,
				       size_t nbytes, loff_t *ppos)
Linus Torvalds's avatar
Linus Torvalds committed
1382
{
1383
	struct cpuset *cs = cgroup_cs(cont);
Linus Torvalds's avatar
Linus Torvalds committed
1384 1385 1386