cpuset.c 64.1 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4 5 6
/*
 *  kernel/cpuset.c
 *
 *  Processor and Memory placement constraints for sets of tasks.
 *
 *  Copyright (C) 2003 BULL SA.
P
Paul Jackson 已提交
7
 *  Copyright (C) 2004-2007 Silicon Graphics, Inc.
8
 *  Copyright (C) 2006 Google, Inc
L
Linus Torvalds 已提交
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.
L
Linus Torvalds 已提交
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
L
Linus Torvalds 已提交
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>
L
Linus Torvalds 已提交
36 37 38 39 40 41
#include <linux/mm.h>
#include <linux/module.h>
#include <linux/mount.h>
#include <linux/namei.h>
#include <linux/pagemap.h>
#include <linux/proc_fs.h>
42
#include <linux/rcupdate.h>
L
Linus Torvalds 已提交
43 44
#include <linux/sched.h>
#include <linux/seq_file.h>
45
#include <linux/security.h>
L
Linus Torvalds 已提交
46 47 48 49 50 51 52 53 54 55
#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>
56
#include <linux/mutex.h>
P
Paul Jackson 已提交
57
#include <linux/kfifo.h>
L
Linus Torvalds 已提交
58

59 60 61 62 63
/*
 * Tracks how many cpusets are currently defined in system.
 * When there is only one cpuset (the root cpuset) we can
 * short circuit some hooks.
 */
64
int number_of_cpusets __read_mostly;
65

66 67 68 69
/* Retrieve the cpuset from a cgroup */
struct cgroup_subsys cpuset_subsys;
struct cpuset;

70 71 72 73 74 75 76 77 78
/* 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 */
};

L
Linus Torvalds 已提交
79
struct cpuset {
80 81
	struct cgroup_subsys_state css;

L
Linus Torvalds 已提交
82 83 84 85 86 87 88 89 90 91
	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.
	 */
92 93 94
	int mems_generation;

	struct fmeter fmeter;		/* memory_pressure filter */
P
Paul Jackson 已提交
95 96 97

	/* partition number for rebuild_sched_domains() */
	int pn;
L
Linus Torvalds 已提交
98 99
};

100 101 102 103 104 105 106 107 108 109 110 111 112 113 114
/* 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);
}


L
Linus Torvalds 已提交
115 116 117 118
/* bits in struct cpuset flags field */
typedef enum {
	CS_CPU_EXCLUSIVE,
	CS_MEM_EXCLUSIVE,
119
	CS_MEMORY_MIGRATE,
P
Paul Jackson 已提交
120
	CS_SCHED_LOAD_BALANCE,
121 122
	CS_SPREAD_PAGE,
	CS_SPREAD_SLAB,
L
Linus Torvalds 已提交
123 124 125 126 127
} cpuset_flagbits_t;

/* convenient tests for these bits */
static inline int is_cpu_exclusive(const struct cpuset *cs)
{
128
	return test_bit(CS_CPU_EXCLUSIVE, &cs->flags);
L
Linus Torvalds 已提交
129 130 131 132
}

static inline int is_mem_exclusive(const struct cpuset *cs)
{
133
	return test_bit(CS_MEM_EXCLUSIVE, &cs->flags);
L
Linus Torvalds 已提交
134 135
}

P
Paul Jackson 已提交
136 137 138 139 140
static inline int is_sched_load_balance(const struct cpuset *cs)
{
	return test_bit(CS_SCHED_LOAD_BALANCE, &cs->flags);
}

141 142
static inline int is_memory_migrate(const struct cpuset *cs)
{
143
	return test_bit(CS_MEMORY_MIGRATE, &cs->flags);
144 145
}

146 147 148 149 150 151 152 153 154 155
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);
}

L
Linus Torvalds 已提交
156
/*
157
 * Increment this integer everytime any cpuset changes its
L
Linus Torvalds 已提交
158 159 160 161 162 163 164 165 166 167 168 169 170
 * 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.
171 172 173
 *
 * Since cpuset_mems_generation is guarded by manage_mutex,
 * there is no need to mark it atomic.
L
Linus Torvalds 已提交
174
 */
175
static int cpuset_mems_generation;
L
Linus Torvalds 已提交
176 177 178 179 180 181 182 183

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

/*
184 185
 * We have two global cpuset mutexes below.  They can nest.
 * It is ok to first take manage_mutex, then nest callback_mutex.  We also
186 187 188
 * require taking task_lock() when dereferencing a tasks cpuset pointer.
 * See "The task_lock() exception", at the end of this comment.
 *
189 190 191
 * 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
192 193
 * and be able to modify cpusets.  It can perform various checks on
 * the cpuset structure first, knowing nothing will change.  It can
194
 * also allocate memory while just holding manage_mutex.  While it is
195
 * performing these checks, various callback routines can briefly
196 197
 * acquire callback_mutex to query cpusets.  Once it is ready to make
 * the changes, it takes callback_mutex, blocking everyone else.
198 199
 *
 * Calls to the kernel memory allocator can not be made while holding
200
 * callback_mutex, as that would risk double tripping on callback_mutex
201 202 203
 * from one of the callbacks into the cpuset code from within
 * __alloc_pages().
 *
204
 * If a task is only holding callback_mutex, then it has read-only
205 206 207 208 209 210
 * 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.
211
 * So in general, code holding manage_mutex or callback_mutex can't rely
212
 * on the count field not changing.  However, if the count goes to
213
 * zero, then only attach_task(), which holds both mutexes, can
214 215 216
 * 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).
217
 * So code holding manage_mutex or callback_mutex can safely assume that
218
 * if the count is zero, it will stay zero.  Similarly, if a task
219
 * holds manage_mutex or callback_mutex on a cpuset with zero count, it
220
 * knows that the cpuset won't be removed, as cpuset_rmdir() needs
221
 * both of those mutexes.
222 223
 *
 * The cpuset_common_file_write handler for operations that modify
224
 * the cpuset hierarchy holds manage_mutex across the entire operation,
225 226
 * single threading all such cpuset modifications across the system.
 *
227
 * The cpuset_common_file_read() handlers only hold callback_mutex across
228 229 230 231
 * 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
232
 * (usually) take either mutex.  These are the two most performance
233
 * critical pieces of code here.  The exception occurs on cpuset_exit(),
234
 * when a task in a notify_on_release cpuset exits.  Then manage_mutex
235
 * is taken, and if the cpuset count is zero, a usermode call made
L
Linus Torvalds 已提交
236 237 238
 * to /sbin/cpuset_release_agent with the name of the cpuset (path
 * relative to the root of cpuset file system) as the argument.
 *
239 240 241
 * 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
242
 * least one task in the system (init), therefore, top_cpuset
243 244 245 246 247 248 249 250 251
 * 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
252
 * so using both mutexes, however there are several performance
253
 * critical places that need to reference task->cpuset without the
254
 * expense of grabbing a system global mutex.  Therefore except as
255 256 257 258
 * 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.
259 260 261 262 263
 *
 * 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().
L
Linus Torvalds 已提交
264 265
 */

266
static DEFINE_MUTEX(callback_mutex);
267

268 269 270
/* 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 */
271 272 273
static int cpuset_get_sb(struct file_system_type *fs_type,
			 int flags, const char *unused_dev_name,
			 void *data, struct vfsmount *mnt)
L
Linus Torvalds 已提交
274
{
275 276 277 278 279 280 281 282 283 284 285
	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;
L
Linus Torvalds 已提交
286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303
}

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.
 *
304
 * Call with callback_mutex held.
L
Linus Torvalds 已提交
305 306 307 308 309 310 311 312 313 314 315 316 317 318 319
 */

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
320 321 322 323
 * 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].
L
Linus Torvalds 已提交
324 325
 *
 * One way or another, we guarantee to return some non-empty subset
326
 * of node_states[N_HIGH_MEMORY].
L
Linus Torvalds 已提交
327
 *
328
 * Call with callback_mutex held.
L
Linus Torvalds 已提交
329 330 331 332
 */

static void guarantee_online_mems(const struct cpuset *cs, nodemask_t *pmask)
{
333 334
	while (cs && !nodes_intersects(cs->mems_allowed,
					node_states[N_HIGH_MEMORY]))
L
Linus Torvalds 已提交
335 336
		cs = cs->parent;
	if (cs)
337 338
		nodes_and(*pmask, cs->mems_allowed,
					node_states[N_HIGH_MEMORY]);
L
Linus Torvalds 已提交
339
	else
340 341
		*pmask = node_states[N_HIGH_MEMORY];
	BUG_ON(!nodes_intersects(*pmask, node_states[N_HIGH_MEMORY]));
L
Linus Torvalds 已提交
342 343
}

344 345 346 347 348 349
/**
 * 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.
350
 *
351 352 353 354
 * 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().
 *
355 356 357 358
 * 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
359
 * current->mm->mmap_sem during call.
360
 *
361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378
 * 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.
379 380 381 382 383
 *
 * 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.
L
Linus Torvalds 已提交
384 385
 */

386
void cpuset_update_task_memory_state(void)
L
Linus Torvalds 已提交
387
{
388
	int my_cpusets_mem_gen;
389
	struct task_struct *tsk = current;
390
	struct cpuset *cs;
391

392
	if (task_cs(tsk) == &top_cpuset) {
393 394 395 396
		/* Don't need rcu for top_cpuset.  It's never freed. */
		my_cpusets_mem_gen = top_cpuset.mems_generation;
	} else {
		rcu_read_lock();
397
		my_cpusets_mem_gen = task_cs(current)->mems_generation;
398 399
		rcu_read_unlock();
	}
L
Linus Torvalds 已提交
400

401
	if (my_cpusets_mem_gen != tsk->cpuset_mems_generation) {
402
		mutex_lock(&callback_mutex);
403
		task_lock(tsk);
404
		cs = task_cs(tsk); /* Maybe changed when task not locked */
405 406
		guarantee_online_mems(cs, &tsk->mems_allowed);
		tsk->cpuset_mems_generation = cs->mems_generation;
407 408 409 410 411 412 413 414
		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;
415
		task_unlock(tsk);
416
		mutex_unlock(&callback_mutex);
417
		mpol_rebind_task(tsk, &tsk->mems_allowed);
L
Linus Torvalds 已提交
418 419 420 421 422 423 424 425
	}
}

/*
 * 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
426
 * are only set if the other's are set.  Call holding manage_mutex.
L
Linus Torvalds 已提交
427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443
 */

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
444
 * manage_mutex held.
L
Linus Torvalds 已提交
445 446 447 448 449 450 451 452 453 454 455 456 457 458
 *
 * '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)
{
459
	struct cgroup *cont;
L
Linus Torvalds 已提交
460 461 462
	struct cpuset *c, *par;

	/* Each of our child cpusets must be a subset of us */
463 464
	list_for_each_entry(cont, &cur->css.cgroup->children, sibling) {
		if (!is_cpuset_subset(cgroup_cs(cont), trial))
L
Linus Torvalds 已提交
465 466 467 468
			return -EBUSY;
	}

	/* Remaining checks don't apply to root cpuset */
469
	if (cur == &top_cpuset)
L
Linus Torvalds 已提交
470 471
		return 0;

472 473
	par = cur->parent;

L
Linus Torvalds 已提交
474 475 476 477 478
	/* 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 */
479 480
	list_for_each_entry(cont, &par->css.cgroup->children, sibling) {
		c = cgroup_cs(cont);
L
Linus Torvalds 已提交
481 482 483 484 485 486 487 488 489 490
		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;
	}

491 492 493 494 495 496 497 498
	/* 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;
		}
	}

L
Linus Torvalds 已提交
499 500 501
	return 0;
}

P
Paul Jackson 已提交
502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703
/*
 * 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
 * a call to the lock_cpu_hotplug()/unlock_cpu_hotplug() pair.
 * Must not be called holding callback_mutex, because we must not
 * call lock_cpu_hotplug() while holding callback_mutex.  Elsewhere
 * the kernel nests callback_mutex inside lock_cpu_hotplug() calls.
 * 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 */
	lock_cpu_hotplug();
	partition_sched_domains(ndoms, doms);
	unlock_cpu_hotplug();

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

704
/*
705
 * Call with manage_mutex held.  May take callback_mutex during call.
706 707
 */

L
Linus Torvalds 已提交
708 709 710
static int update_cpumask(struct cpuset *cs, char *buf)
{
	struct cpuset trialcs;
711
	int retval;
P
Paul Jackson 已提交
712
	int cpus_changed, is_load_balanced;
L
Linus Torvalds 已提交
713

714 715 716 717
	/* top_cpuset.cpus_allowed tracks cpu_online_map; it's read-only */
	if (cs == &top_cpuset)
		return -EACCES;

L
Linus Torvalds 已提交
718
	trialcs = *cs;
719 720

	/*
721 722 723 724
	 * 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.
725
	 */
726 727
	buf = strstrip(buf);
	if (!*buf) {
728 729 730 731 732 733
		cpus_clear(trialcs.cpus_allowed);
	} else {
		retval = cpulist_parse(buf, trialcs.cpus_allowed);
		if (retval < 0)
			return retval;
	}
L
Linus Torvalds 已提交
734 735
	cpus_and(trialcs.cpus_allowed, trialcs.cpus_allowed, cpu_online_map);
	retval = validate_change(cs, &trialcs);
736 737
	if (retval < 0)
		return retval;
P
Paul Jackson 已提交
738 739 740 741

	cpus_changed = !cpus_equal(cs->cpus_allowed, trialcs.cpus_allowed);
	is_load_balanced = is_sched_load_balance(&trialcs);

742
	mutex_lock(&callback_mutex);
743
	cs->cpus_allowed = trialcs.cpus_allowed;
744
	mutex_unlock(&callback_mutex);
P
Paul Jackson 已提交
745 746 747 748

	if (cpus_changed && is_load_balanced)
		rebuild_sched_domains();

749
	return 0;
L
Linus Torvalds 已提交
750 751
}

752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786 787 788 789 790 791 792 793 794 795 796
/*
 * 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);
797
	guarantee_online_mems(task_cs(tsk),&tsk->mems_allowed);
798 799 800
	mutex_unlock(&callback_mutex);
}

801
/*
802 803 804
 * 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
805 806 807
 * task in the cpuset, rebind any vma mempolicies and if
 * the cpuset is marked 'memory_migrate', migrate the tasks
 * pages to the new memory.
808
 *
809
 * Call with manage_mutex held.  May take callback_mutex during call.
810 811 812
 * 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.
813 814
 */

815 816
static void *cpuset_being_rebound;

L
Linus Torvalds 已提交
817 818 819
static int update_nodemask(struct cpuset *cs, char *buf)
{
	struct cpuset trialcs;
820
	nodemask_t oldmem;
821
	struct task_struct *p;
822 823
	struct mm_struct **mmarray;
	int i, n, ntasks;
824
	int migrate;
825
	int fudge;
L
Linus Torvalds 已提交
826
	int retval;
827
	struct cgroup_iter it;
L
Linus Torvalds 已提交
828

829 830 831 832
	/*
	 * top_cpuset.mems_allowed tracks node_stats[N_HIGH_MEMORY];
	 * it's read-only
	 */
833 834 835
	if (cs == &top_cpuset)
		return -EACCES;

L
Linus Torvalds 已提交
836
	trialcs = *cs;
837 838

	/*
839 840 841 842
	 * 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.
843
	 */
844 845
	buf = strstrip(buf);
	if (!*buf) {
846 847 848 849 850 851
		nodes_clear(trialcs.mems_allowed);
	} else {
		retval = nodelist_parse(buf, trialcs.mems_allowed);
		if (retval < 0)
			goto done;
	}
852 853
	nodes_and(trialcs.mems_allowed, trialcs.mems_allowed,
						node_states[N_HIGH_MEMORY]);
854 855 856 857 858
	oldmem = cs->mems_allowed;
	if (nodes_equal(oldmem, trialcs.mems_allowed)) {
		retval = 0;		/* Too easy - nothing to do */
		goto done;
	}
859 860 861 862
	retval = validate_change(cs, &trialcs);
	if (retval < 0)
		goto done;

863
	mutex_lock(&callback_mutex);
864
	cs->mems_allowed = trialcs.mems_allowed;
865
	cs->mems_generation = cpuset_mems_generation++;
866
	mutex_unlock(&callback_mutex);
867

868
	cpuset_being_rebound = cs;		/* causes mpol_copy() rebind */
869 870 871 872 873 874 875 876 877 878 879 880 881

	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) {
882
		ntasks = cgroup_task_count(cs->css.cgroup);  /* guess */
883 884 885 886
		ntasks += fudge;
		mmarray = kmalloc(ntasks * sizeof(*mmarray), GFP_KERNEL);
		if (!mmarray)
			goto done;
887
		read_lock(&tasklist_lock);		/* block fork */
888
		if (cgroup_task_count(cs->css.cgroup) <= ntasks)
889
			break;				/* got enough */
890
		read_unlock(&tasklist_lock);		/* try again */
891 892 893 894 895 896
		kfree(mmarray);
	}

	n = 0;

	/* Load up mmarray[] with mm reference for each task in cpuset. */
897 898
	cgroup_iter_start(cs->css.cgroup, &it);
	while ((p = cgroup_iter_next(cs->css.cgroup, &it))) {
899 900 901 902 903
		struct mm_struct *mm;

		if (n >= ntasks) {
			printk(KERN_WARNING
				"Cpuset mempolicy rebind incomplete.\n");
904
			break;
905 906 907 908 909
		}
		mm = get_task_mm(p);
		if (!mm)
			continue;
		mmarray[n++] = mm;
910 911
	}
	cgroup_iter_end(cs->css.cgroup, &it);
912
	read_unlock(&tasklist_lock);
913 914 915 916 917 918 919 920 921

	/*
	 * 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
922
	 * cpuset manage_mutex, we know that no other rebind effort will
923 924
	 * be contending for the global variable cpuset_being_rebound.
	 * It's ok if we rebind the same mm twice; mpol_rebind_mm()
925
	 * is idempotent.  Also migrate pages in each mm to new nodes.
926
	 */
927
	migrate = is_memory_migrate(cs);
928 929 930 931
	for (i = 0; i < n; i++) {
		struct mm_struct *mm = mmarray[i];

		mpol_rebind_mm(mm, &cs->mems_allowed);
932 933
		if (migrate)
			cpuset_migrate_mm(mm, &oldmem, &cs->mems_allowed);
934 935 936 937 938
		mmput(mm);
	}

	/* We're done rebinding vma's to this cpusets new mems_allowed. */
	kfree(mmarray);
939
	cpuset_being_rebound = NULL;
940
	retval = 0;
941
done:
L
Linus Torvalds 已提交
942 943 944
	return retval;
}

945 946 947 948 949
int current_cpuset_is_being_rebound(void)
{
	return task_cs(current) == cpuset_being_rebound;
}

950
/*
951
 * Call with manage_mutex held.
952 953 954 955 956 957 958 959 960 961 962
 */

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;
}

L
Linus Torvalds 已提交
963 964 965
/*
 * 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,
P
Paul Jackson 已提交
966
 *				CS_SCHED_LOAD_BALANCE,
967 968
 *				CS_NOTIFY_ON_RELEASE, CS_MEMORY_MIGRATE,
 *				CS_SPREAD_PAGE, CS_SPREAD_SLAB)
L
Linus Torvalds 已提交
969 970
 * cs:	the cpuset to update
 * buf:	the buffer where we read the 0 or 1
971
 *
972
 * Call with manage_mutex held.
L
Linus Torvalds 已提交
973 974 975 976 977 978
 */

static int update_flag(cpuset_flagbits_t bit, struct cpuset *cs, char *buf)
{
	int turning_on;
	struct cpuset trialcs;
979
	int err;
P
Paul Jackson 已提交
980
	int cpus_nonempty, balance_flag_changed;
L
Linus Torvalds 已提交
981 982 983 984 985 986 987 988 989 990

	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);
991 992
	if (err < 0)
		return err;
P
Paul Jackson 已提交
993 994 995 996 997

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

998
	mutex_lock(&callback_mutex);
999
	cs->flags = trialcs.flags;
1000
	mutex_unlock(&callback_mutex);
1001

P
Paul Jackson 已提交
1002 1003 1004
	if (cpus_nonempty && balance_flag_changed)
		rebuild_sched_domains();

1005
	return 0;
L
Linus Torvalds 已提交
1006 1007
}

1008
/*
A
Adrian Bunk 已提交
1009
 * Frequency meter - How fast is some event occurring?
1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 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 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105
 *
 * 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;
}

1106 1107
static int cpuset_can_attach(struct cgroup_subsys *ss,
			     struct cgroup *cont, struct task_struct *tsk)
L
Linus Torvalds 已提交
1108
{
1109
	struct cpuset *cs = cgroup_cs(cont);
L
Linus Torvalds 已提交
1110 1111 1112 1113

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

1114 1115
	return security_task_setscheduler(tsk, 0, NULL);
}
L
Linus Torvalds 已提交
1116

1117 1118 1119 1120 1121 1122 1123 1124 1125
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);
1126

1127
	mutex_lock(&callback_mutex);
L
Linus Torvalds 已提交
1128 1129
	guarantee_online_cpus(cs, &cpus);
	set_cpus_allowed(tsk, cpus);
1130
	mutex_unlock(&callback_mutex);
L
Linus Torvalds 已提交
1131

1132 1133
	from = oldcs->mems_allowed;
	to = cs->mems_allowed;
1134 1135 1136
	mm = get_task_mm(tsk);
	if (mm) {
		mpol_rebind_mm(mm, &to);
1137
		if (is_memory_migrate(cs))
1138
			cpuset_migrate_mm(mm, &from, &to);
1139 1140 1141
		mmput(mm);
	}

L
Linus Torvalds 已提交
1142 1143 1144 1145 1146
}

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

typedef enum {
1147
	FILE_MEMORY_MIGRATE,
L
Linus Torvalds 已提交
1148 1149 1150 1151
	FILE_CPULIST,
	FILE_MEMLIST,
	FILE_CPU_EXCLUSIVE,
	FILE_MEM_EXCLUSIVE,
P
Paul Jackson 已提交
1152
	FILE_SCHED_LOAD_BALANCE,
1153 1154
	FILE_MEMORY_PRESSURE_ENABLED,
	FILE_MEMORY_PRESSURE,
1155 1156
	FILE_SPREAD_PAGE,
	FILE_SPREAD_SLAB,
L
Linus Torvalds 已提交
1157 1158
} cpuset_filetype_t;

1159 1160 1161
static ssize_t cpuset_common_file_write(struct cgroup *cont,
					struct cftype *cft,
					struct file *file,
1162
					const char __user *userbuf,
L
Linus Torvalds 已提交
1163 1164
					size_t nbytes, loff_t *unused_ppos)
{
1165
	struct cpuset *cs = cgroup_cs(cont);
L
Linus Torvalds 已提交
1166 1167 1168 1169 1170
	cpuset_filetype_t type = cft->private;
	char *buffer;
	int retval = 0;

	/* Crude upper limit on largest legitimate cpulist user might write. */
P
Paul Jackson 已提交
1171
	if (nbytes > 100U + 6 * max(NR_CPUS, MAX_NUMNODES))
L
Linus Torvalds 已提交
1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183
		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 */

1184
	cgroup_lock();
L
Linus Torvalds 已提交
1185

1186
	if (cgroup_is_removed(cont)) {
L
Linus Torvalds 已提交
1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203
		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;
P
Paul Jackson 已提交
1204 1205 1206
	case FILE_SCHED_LOAD_BALANCE:
		retval = update_flag(CS_SCHED_LOAD_BALANCE, cs, buffer);
		break;
1207 1208 1209
	case FILE_MEMORY_MIGRATE:
		retval = update_flag(CS_MEMORY_MIGRATE, cs, buffer);
		break;
1210 1211 1212 1213 1214 1215
	case FILE_MEMORY_PRESSURE_ENABLED:
		retval = update_memory_pressure_enabled(cs, buffer);
		break;
	case FILE_MEMORY_PRESSURE:
		retval = -EACCES;
		break;
1216 1217
	case FILE_SPREAD_PAGE:
		retval = update_flag(CS_SPREAD_PAGE, cs, buffer);
1218
		cs->mems_generation = cpuset_mems_generation++;
1219 1220 1221
		break;
	case FILE_SPREAD_SLAB:
		retval = update_flag(CS_SPREAD_SLAB, cs, buffer);
1222
		cs->mems_generation = cpuset_mems_generation++;
1223
		break;
L
Linus Torvalds 已提交
1224 1225 1226 1227 1228 1229 1230 1231
	default:
		retval = -EINVAL;
		goto out2;
	}

	if (retval == 0)
		retval = nbytes;
out2:
1232
	cgroup_unlock();
L
Linus Torvalds 已提交
1233 1234 1235 1236 1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253
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;

1254
	mutex_lock(&callback_mutex);
L
Linus Torvalds 已提交
1255
	mask = cs->cpus_allowed;
1256
	mutex_unlock(&callback_mutex);
L
Linus Torvalds 已提交
1257 1258 1259 1260 1261 1262 1263 1264

	return cpulist_scnprintf(page, PAGE_SIZE, mask);
}

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

1265
	mutex_lock(&callback_mutex);
L
Linus Torvalds 已提交
1266
	mask = cs->mems_allowed;
1267
	mutex_unlock(&callback_mutex);
L
Linus Torvalds 已提交
1268 1269 1270 1271

	return nodelist_scnprintf(page, PAGE_SIZE, mask);
}

1272 1273 1274 1275 1276
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)
L
Linus Torvalds 已提交
1277
{
1278
	struct cpuset *cs = cgroup_cs(cont);
L
Linus Torvalds 已提交
1279 1280 1281 1282 1283
	cpuset_filetype_t type = cft->private;
	char *page;
	ssize_t retval = 0;
	char *s;

1284
	if (!(page = (char *)__get_free_page(GFP_TEMPORARY)))
L
Linus Torvalds 已提交
1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301
		return -ENOMEM;

	s = page;

	switch (type) {
	case FILE_CPULIST:
		s += cpuset_sprintf_cpulist(s, cs);
		break;
	case FILE_MEMLIST:
		s += cpuset_sprintf_memlist(s, cs);
		break;
	case FILE_CPU_EXCLUSIVE:
		*s++ = is_cpu_exclusive(cs) ? '1' : '0';
		break;
	case FILE_MEM_EXCLUSIVE:
		*s++ = is_mem_exclusive(cs) ? '1' : '0';
		break;
P
Paul Jackson 已提交
1302 1303 1304
	case FILE_SCHED_LOAD_BALANCE:
		*s++ = is_sched_load_balance(cs) ? '1' : '0';
		break;
1305 1306 1307
	case FILE_MEMORY_MIGRATE:
		*s++ = is_memory_migrate(cs) ? '1' : '0';
		break;
1308 1309 1310 1311 1312 1313
	case FILE_MEMORY_PRESSURE_ENABLED:
		*s++ = cpuset_memory_pressure_enabled ? '1' : '0';
		break;
	case FILE_MEMORY_PRESSURE:
		s += sprintf(s, "%d", fmeter_getrate(&cs->fmeter));
		break;
1314 1315 1316 1317 1318 1319
	case FILE_SPREAD_PAGE:
		*s++ = is_spread_page(cs) ? '1' : '0';
		break;
	case FILE_SPREAD_SLAB:
		*s++ = is_spread_slab(cs) ? '1' : '0';
		break;
L
Linus Torvalds 已提交
1320 1321 1322 1323 1324 1325
	default:
		retval = -EINVAL;
		goto out;
	}
	*s++ = '\n';

A
Al Viro 已提交
1326
	retval = simple_read_from_buffer(buf, nbytes, ppos, page, s - page);
L
Linus Torvalds 已提交
1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341
out:
	free_page((unsigned long)page);
	return retval;
}





/*
 * for the common functions, 'private' gives the type of file
 */

static struct cftype cft_cpus = {
	.name = "cpus",
1342 1343
	.read = cpuset_common_file_read,
	.write = cpuset_common_file_write,
L
Linus Torvalds 已提交
1344 1345 1346 1347 1348
	.private = FILE_CPULIST,
};

static struct cftype cft_mems = {
	.name = "mems",
1349 1350
	.read = cpuset_common_file_read,
	.write = cpuset_common_file_write,
L
Linus Torvalds 已提交
1351 1352 1353 1354 1355
	.private = FILE_MEMLIST,
};

static struct cftype cft_cpu_exclusive = {
	.name = "cpu_exclusive",
1356 1357
	.read = cpuset_common_file_read,
	.write = cpuset_common_file_write,
L
Linus Torvalds 已提交
1358 1359 1360 1361 1362
	.private = FILE_CPU_EXCLUSIVE,
};

static struct cftype cft_mem_exclusive = {
	.name = "mem_exclusive",
1363 1364
	.read = cpuset_common_file_read,
	.write = cpuset_common_file_write,
L
Linus Torvalds 已提交
1365 1366 1367
	.private = FILE_MEM_EXCLUSIVE,
};

P
Paul Jackson 已提交
1368 1369 1370 1371 1372 1373 1374
static struct cftype cft_sched_load_balance = {
	.name = "sched_load_balance",
	.read = cpuset_common_file_read,
	.write = cpuset_common_file_write,
	.private = FILE_SCHED_LOAD_BALANCE,
};

1375 1376
static struct cftype cft_memory_migrate = {
	.name = "memory_migrate",
1377 1378
	.read = cpuset_common_file_read,
	.write = cpuset_common_file_write,
1379 1380 1381
	.private = FILE_MEMORY_MIGRATE,
};

1382 1383
static struct cftype cft_memory_pressure_enabled = {
	.name = "memory_pressure_enabled",
1384 1385
	.read = cpuset_common_file_read,
	.write = cpuset_common_file_write,
1386 1387 1388 1389 1390
	.private = FILE_MEMORY_PRESSURE_ENABLED,
};

static struct cftype cft_memory_pressure = {
	.name = "memory_pressure",
1391 1392
	.read = cpuset_common_file_read,
	.write = cpuset_common_file_write,
1393 1394 1395
	.private = FILE_MEMORY_PRESSURE,
};

1396 1397
static struct cftype cft_spread_page = {
	.name = "memory_spread_page",
1398 1399
	.read = cpuset_common_file_read,
	.write = cpuset_common_file_write,
1400 1401 1402 1403 1404
	.private = FILE_SPREAD_PAGE,
};

static struct cftype cft_spread_slab = {
	.name = "memory_spread_slab",
1405 1406
	.read = cpuset_common_file_read,
	.write = cpuset_common_file_write,
1407 1408 1409
	.private = FILE_SPREAD_SLAB,
};

1410
static int cpuset_populate(struct cgroup_subsys *ss, struct cgroup *cont)
L
Linus Torvalds 已提交
1411 1412 1413
{
	int err;

1414
	if ((err = cgroup_add_file(cont, ss, &cft_cpus)) < 0)
L
Linus Torvalds 已提交
1415
		return err;
1416
	if ((err = cgroup_add_file(cont, ss, &cft_mems)) < 0)
L
Linus Torvalds 已提交
1417
		return err;
1418
	if ((err = cgroup_add_file(cont, ss, &cft_cpu_exclusive)) < 0)
L
Linus Torvalds 已提交
1419
		return err;
1420
	if ((err = cgroup_add_file(cont, ss, &cft_mem_exclusive)) < 0)
L
Linus Torvalds 已提交
1421
		return err;
1422
	if ((err = cgroup_add_file(cont, ss, &cft_memory_migrate)) < 0)
L
Linus Torvalds 已提交
1423
		return err;
P
Paul Jackson 已提交
1424 1425
	if ((err = cgroup_add_file(cont, ss, &cft_sched_load_balance)) < 0)
		return err;
1426
	if ((err = cgroup_add_file(cont, ss, &cft_memory_pressure)) < 0)
1427
		return err;
1428
	if ((err = cgroup_add_file(cont, ss, &cft_spread_page)) < 0)
1429
		return err;
1430
	if ((err = cgroup_add_file(cont, ss, &cft_spread_slab)) < 0)
L
Linus Torvalds 已提交
1431
		return err;
1432 1433 1434 1435
	/* memory_pressure_enabled is in root cpuset only */
	if (err == 0 && !cont->parent)
		err = cgroup_add_file(cont, ss,
					 &cft_memory_pressure_enabled);
L
Linus Torvalds 已提交
1436 1437 1438
	return 0;
}

1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474
/*
 * post_clone() is called at the end of cgroup_clone().
 * 'cgroup' was just created automatically as a result of
 * a cgroup_clone(), and the current task is about to
 * be moved into 'cgroup'.
 *
 * Currently we refuse to set up the cgroup - thereby
 * refusing the task to be entered, and as a result refusing
 * the sys_unshare() or clone() which initiated it - if any
 * sibling cpusets have exclusive cpus or mem.
 *
 * If this becomes a problem for some users who wish to
 * allow that scenario, then cpuset_post_clone() could be
 * changed to grant parent->cpus_allowed-sibling_cpus_exclusive
 * (and likewise for mems) to the new cgroup.
 */
static void cpuset_post_clone(struct cgroup_subsys *ss,
			      struct cgroup *cgroup)
{
	struct cgroup *parent, *child;
	struct cpuset *cs, *parent_cs;

	parent = cgroup->parent;
	list_for_each_entry(child, &parent->children, sibling) {
		cs = cgroup_cs(child);
		if (is_mem_exclusive(cs) || is_cpu_exclusive(cs))
			return;
	}
	cs = cgroup_cs(cgroup);
	parent_cs = cgroup_cs(parent);

	cs->mems_allowed = parent_cs->mems_allowed;
	cs->cpus_allowed = parent_cs->cpus_allowed;
	return;
}

L
Linus Torvalds 已提交
1475 1476 1477 1478 1479 1480
/*
 *	cpuset_create - create a cpuset
 *	parent:	cpuset that will be parent of the new cpuset.
 *	name:		name of the new cpuset. Will be strcpy'ed.
 *	mode:		mode to set on new inode
 *
1481
 *	Must be called with the mutex on the parent inode held
L
Linus Torvalds 已提交
1482 1483
 */

1484 1485 1486
static struct cgroup_subsys_state *cpuset_create(
	struct cgroup_subsys *ss,
	struct cgroup *cont)
L
Linus Torvalds 已提交
1487 1488
{
	struct cpuset *cs;
1489
	struct cpuset *parent;
L
Linus Torvalds 已提交
1490

1491 1492 1493 1494 1495 1496
	if (!cont->parent) {
		/* This is early initialization for the top cgroup */
		top_cpuset.mems_generation = cpuset_mems_generation++;
		return &top_cpuset.css;
	}
	parent = cgroup_cs(cont->parent);
L
Linus Torvalds 已提交
1497 1498
	cs = kmalloc(sizeof(*cs), GFP_KERNEL);
	if (!cs)
1499
		return ERR_PTR(-ENOMEM);
L
Linus Torvalds 已提交
1500

1501
	cpuset_update_task_memory_state();
L
Linus Torvalds 已提交
1502
	cs->flags = 0;
1503 1504 1505 1506
	if (is_spread_page(parent))
		set_bit(CS_SPREAD_PAGE, &cs->flags);
	if (is_spread_slab(parent))
		set_bit(CS_SPREAD_SLAB, &cs->flags);
P
Paul Jackson 已提交
1507
	set_bit(CS_SCHED_LOAD_BALANCE, &cs->flags);
L
Linus Torvalds 已提交
1508 1509
	cs->cpus_allowed = CPU_MASK_NONE;
	cs->mems_allowed = NODE_MASK_NONE;
1510
	cs->mems_generation = cpuset_mems_generation++;
1511
	fmeter_init(&cs->fmeter);
L
Linus Torvalds 已提交
1512 1513

	cs->parent = parent;
1514
	number_of_cpusets++;
1515
	return &cs->css ;
L
Linus Torvalds 已提交
1516 1517
}

P
Paul Jackson 已提交
1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529
/*
 * Locking note on the strange update_flag() call below:
 *
 * If the cpuset being removed has its flag 'sched_load_balance'
 * enabled, then simulate turning sched_load_balance off, which
 * will call rebuild_sched_domains().  The lock_cpu_hotplug()
 * call in rebuild_sched_domains() must not be made while holding
 * callback_mutex.  Elsewhere the kernel nests callback_mutex inside
 * lock_cpu_hotplug() calls.  So the reverse nesting would risk an
 * ABBA deadlock.
 */

1530
static void cpuset_destroy(struct cgroup_subsys *ss, struct cgroup *cont)
L
Linus Torvalds 已提交
1531
{
1532
	struct cpuset *cs = cgroup_cs(cont);
L
Linus Torvalds 已提交
1533

1534
	cpuset_update_task_memory_state();
P
Paul Jackson 已提交
1535 1536 1537 1538

	if (is_sched_load_balance(cs))
		update_flag(CS_SCHED_LOAD_BALANCE, cs, "0");

1539
	number_of_cpusets--;
1540
	kfree(cs);
L
Linus Torvalds 已提交
1541 1542
}

1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554
struct cgroup_subsys cpuset_subsys = {
	.name = "cpuset",
	.create = cpuset_create,
	.destroy  = cpuset_destroy,
	.can_attach = cpuset_can_attach,
	.attach = cpuset_attach,
	.populate = cpuset_populate,
	.post_clone = cpuset_post_clone,
	.subsys_id = cpuset_subsys_id,
	.early_init = 1,
};

1555 1556 1557 1558 1559 1560 1561 1562
/*
 * cpuset_init_early - just enough so that the calls to
 * cpuset_update_task_memory_state() in early init code
 * are harmless.
 */

int __init cpuset_init_early(void)
{
1563
	top_cpuset.mems_generation = cpuset_mems_generation++;
1564 1565 1566
	return 0;
}

1567

L
Linus Torvalds 已提交
1568 1569 1570 1571 1572 1573 1574 1575
/**
 * cpuset_init - initialize cpusets at system boot
 *
 * Description: Initialize top_cpuset and the cpuset internal file system,
 **/

int __init cpuset_init(void)
{
1576
	int err = 0;
L
Linus Torvalds 已提交
1577 1578 1579 1580

	top_cpuset.cpus_allowed = CPU_MASK_ALL;
	top_cpuset.mems_allowed = NODE_MASK_ALL;

1581
	fmeter_init(&top_cpuset.fmeter);
1582
	top_cpuset.mems_generation = cpuset_mems_generation++;
P
Paul Jackson 已提交
1583
	set_bit(CS_SCHED_LOAD_BALANCE, &top_cpuset.flags);
L
Linus Torvalds 已提交
1584 1585 1586

	err = register_filesystem(&cpuset_fs_type);
	if (err < 0)
1587 1588
		return err;

1589
	number_of_cpusets = 1;
1590
	return 0;
L
Linus Torvalds 已提交
1591 1592
}

1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615
/*
 * If common_cpu_mem_hotplug_unplug(), below, unplugs any CPUs
 * or memory nodes, we need to walk over the cpuset hierarchy,
 * removing that CPU or node from all cpusets.  If this removes the
 * last CPU or node from a cpuset, then the guarantee_online_cpus()
 * or guarantee_online_mems() code will use that emptied cpusets
 * parent online CPUs or nodes.  Cpusets that were already empty of
 * CPUs or nodes are left empty.
 *
 * This routine is intentionally inefficient in a couple of regards.
 * It will check all cpusets in a subtree even if the top cpuset of
 * the subtree has no offline CPUs or nodes.  It checks both CPUs and
 * nodes, even though the caller could have been coded to know that
 * only one of CPUs or nodes needed to be checked on a given call.
 * This was done to minimize text size rather than cpu cycles.
 *
 * Call with both manage_mutex and callback_mutex held.
 *
 * Recursive, on depth of cpuset subtree.
 */

static void guarantee_online_cpus_mems_in_subtree(const struct cpuset *cur)
{
1616
	struct cgroup *cont;
1617 1618 1619
	struct cpuset *c;

	/* Each of our child cpusets mems must be online */
1620 1621
	list_for_each_entry(cont, &cur->css.cgroup->children, sibling) {
		c = cgroup_cs(cont);
1622 1623 1624 1625 1626 1627 1628 1629 1630 1631
		guarantee_online_cpus_mems_in_subtree(c);
		if (!cpus_empty(c->cpus_allowed))
			guarantee_online_cpus(c, &c->cpus_allowed);
		if (!nodes_empty(c->mems_allowed))
			guarantee_online_mems(c, &c->mems_allowed);
	}
}

/*
 * The cpus_allowed and mems_allowed nodemasks in the top_cpuset track
1632 1633 1634
 * cpu_online_map and node_states[N_HIGH_MEMORY].  Force the top cpuset to
 * track what's online after any CPU or memory node hotplug or unplug
 * event.
1635 1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648
 *
 * To ensure that we don't remove a CPU or node from the top cpuset
 * that is currently in use by a child cpuset (which would violate
 * the rule that cpusets must be subsets of their parent), we first
 * call the recursive routine guarantee_online_cpus_mems_in_subtree().
 *
 * Since there are two callers of this routine, one for CPU hotplug
 * events and one for memory node hotplug events, we could have coded
 * two separate routines here.  We code it as a single common routine
 * in order to minimize text size.
 */

static void common_cpu_mem_hotplug_unplug(void)
{
1649
	cgroup_lock();
1650 1651 1652 1653
	mutex_lock(&callback_mutex);

	guarantee_online_cpus_mems_in_subtree(&top_cpuset);
	top_cpuset.cpus_allowed = cpu_online_map;
1654
	top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY];
1655 1656

	mutex_unlock(&callback_mutex);
1657
	cgroup_unlock();
1658 1659
}

1660 1661 1662 1663 1664 1665
/*
 * The top_cpuset tracks what CPUs and Memory Nodes are online,
 * period.  This is necessary in order to make cpusets transparent
 * (of no affect) on systems that are actively using CPU hotplug
 * but making no active use of cpusets.
 *
1666 1667
 * This routine ensures that top_cpuset.cpus_allowed tracks
 * cpu_online_map on each CPU hotplug (cpuhp) event.
1668 1669
 */

P
Paul Jackson 已提交
1670 1671
static int cpuset_handle_cpuhp(struct notifier_block *unused_nb,
				unsigned long phase, void *unused_cpu)
1672
{
1673 1674 1675
	if (phase == CPU_DYING || phase == CPU_DYING_FROZEN)
		return NOTIFY_DONE;

1676
	common_cpu_mem_hotplug_unplug();
1677 1678 1679
	return 0;
}

1680
#ifdef CONFIG_MEMORY_HOTPLUG
1681
/*
1682 1683 1684
 * Keep top_cpuset.mems_allowed tracking node_states[N_HIGH_MEMORY].
 * Call this routine anytime after you change
 * node_states[N_HIGH_MEMORY].
1685 1686 1687
 * See also the previous routine cpuset_handle_cpuhp().
 */

A
Al Viro 已提交
1688
void cpuset_track_online_nodes(void)
1689
{
1690
	common_cpu_mem_hotplug_unplug();
1691 1692 1693
}
#endif

L
Linus Torvalds 已提交
1694 1695 1696 1697 1698 1699 1700 1701 1702
/**
 * cpuset_init_smp - initialize cpus_allowed
 *
 * Description: Finish top cpuset after cpu, node maps are initialized
 **/

void __init cpuset_init_smp(void)
{
	top_cpuset.cpus_allowed = cpu_online_map;
1703
	top_cpuset.mems_allowed = node_states[N_HIGH_MEMORY];
1704 1705

	hotcpu_notifier(cpuset_handle_cpuhp, 0);
L
Linus Torvalds 已提交
1706 1707 1708
}

/**
1709

L
Linus Torvalds 已提交
1710 1711 1712 1713 1714 1715 1716 1717 1718
 * cpuset_cpus_allowed - return cpus_allowed mask from a tasks cpuset.
 * @tsk: pointer to task_struct from which to obtain cpuset->cpus_allowed.
 *
 * Description: Returns the cpumask_t cpus_allowed of the cpuset
 * attached to the specified @tsk.  Guaranteed to return some non-empty
 * subset of cpu_online_map, even if this means going outside the
 * tasks cpuset.
 **/

1719
cpumask_t cpuset_cpus_allowed(struct task_struct *tsk)
L
Linus Torvalds 已提交
1720 1721 1722
{
	cpumask_t mask;

1723
	mutex_lock(&callback_mutex);
1724
	task_lock(tsk);
1725
	guarantee_online_cpus(task_cs(tsk), &mask);
1726
	task_unlock(tsk);
1727
	mutex_unlock(&callback_mutex);
L
Linus Torvalds 已提交
1728 1729 1730 1731 1732 1733 1734 1735 1736

	return mask;
}

void cpuset_init_current_mems_allowed(void)
{
	current->mems_allowed = NODE_MASK_ALL;
}

1737 1738 1739 1740 1741 1742
/**
 * cpuset_mems_allowed - return mems_allowed mask from a tasks cpuset.
 * @tsk: pointer to task_struct from which to obtain cpuset->mems_allowed.
 *
 * Description: Returns the nodemask_t mems_allowed of the cpuset
 * attached to the specified @tsk.  Guaranteed to return some non-empty
1743
 * subset of node_states[N_HIGH_MEMORY], even if this means going outside the
1744 1745 1746 1747 1748 1749 1750
 * tasks cpuset.
 **/

nodemask_t cpuset_mems_allowed(struct task_struct *tsk)
{
	nodemask_t mask;

1751
	mutex_lock(&callback_mutex);
1752
	task_lock(tsk);
1753
	guarantee_online_mems(task_cs(tsk), &mask);
1754
	task_unlock(tsk);
1755
	mutex_unlock(&callback_mutex);
1756 1757 1758 1759

	return mask;
}

1760 1761 1762 1763
/**
 * cpuset_zonelist_valid_mems_allowed - check zonelist vs. curremt mems_allowed
 * @zl: the zonelist to be checked
 *
L
Linus Torvalds 已提交
1764 1765 1766 1767 1768 1769 1770
 * Are any of the nodes on zonelist zl allowed in current->mems_allowed?
 */
int cpuset_zonelist_valid_mems_allowed(struct zonelist *zl)
{
	int i;

	for (i = 0; zl->zones[i]; i++) {
1771
		int nid = zone_to_nid(zl->zones[i]);
L
Linus Torvalds 已提交
1772 1773 1774 1775 1776 1777 1778

		if (node_isset(nid, current->mems_allowed))
			return 1;
	}
	return 0;
}

1779 1780
/*
 * nearest_exclusive_ancestor() - Returns the nearest mem_exclusive
1781
 * ancestor to the specified cpuset.  Call holding callback_mutex.
1782 1783 1784 1785 1786 1787 1788 1789 1790 1791
 * If no ancestor is mem_exclusive (an unusual configuration), then
 * returns the root cpuset.
 */
static const struct cpuset *nearest_exclusive_ancestor(const struct cpuset *cs)
{
	while (!is_mem_exclusive(cs) && cs->parent)
		cs = cs->parent;
	return cs;
}

1792
/**
1793
 * cpuset_zone_allowed_softwall - Can we allocate on zone z's memory node?
1794
 * @z: is this zone on an allowed node?
1795
 * @gfp_mask: memory allocation flags
1796
 *
1797 1798
 * If we're in interrupt, yes, we can always allocate.  If
 * __GFP_THISNODE is set, yes, we can always allocate.  If zone
1799 1800 1801
 * z's node is in our tasks mems_allowed, yes.  If it's not a
 * __GFP_HARDWALL request and this zone's nodes is in the nearest
 * mem_exclusive cpuset ancestor to this tasks cpuset, yes.
1802 1803
 * If the task has been OOM killed and has access to memory reserves
 * as specified by the TIF_MEMDIE flag, yes.
1804 1805
 * Otherwise, no.
 *
1806 1807 1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819
 * If __GFP_HARDWALL is set, cpuset_zone_allowed_softwall()
 * reduces to cpuset_zone_allowed_hardwall().  Otherwise,
 * cpuset_zone_allowed_softwall() might sleep, and might allow a zone
 * from an enclosing cpuset.
 *
 * cpuset_zone_allowed_hardwall() only handles the simpler case of
 * hardwall cpusets, and never sleeps.
 *
 * The __GFP_THISNODE placement logic is really handled elsewhere,
 * by forcibly using a zonelist starting at a specified node, and by
 * (in get_page_from_freelist()) refusing to consider the zones for
 * any node on the zonelist except the first.  By the time any such
 * calls get to this routine, we should just shut up and say 'yes'.
 *
1820
 * GFP_USER allocations are marked with the __GFP_HARDWALL bit,
1821 1822
 * and do not allow allocations outside the current tasks cpuset
 * unless the task has been OOM killed as is marked TIF_MEMDIE.
1823
 * GFP_KERNEL allocations are not so marked, so can escape to the
1824
 * nearest enclosing mem_exclusive ancestor cpuset.
1825
 *
1826 1827 1828 1829 1830 1831 1832
 * Scanning up parent cpusets requires callback_mutex.  The
 * __alloc_pages() routine only calls here with __GFP_HARDWALL bit
 * _not_ set if it's a GFP_KERNEL allocation, and all nodes in the
 * current tasks mems_allowed came up empty on the first pass over
 * the zonelist.  So only GFP_KERNEL allocations, if all nodes in the
 * cpuset are short of memory, might require taking the callback_mutex
 * mutex.
1833
 *
1834
 * The first call here from mm/page_alloc:get_page_from_freelist()
1835 1836 1837
 * has __GFP_HARDWALL set in gfp_mask, enforcing hardwall cpusets,
 * so no allocation on a node outside the cpuset is allowed (unless
 * in interrupt, of course).
1838 1839 1840 1841 1842 1843
 *
 * The second pass through get_page_from_freelist() doesn't even call
 * here for GFP_ATOMIC calls.  For those calls, the __alloc_pages()
 * variable 'wait' is not set, and the bit ALLOC_CPUSET is not set
 * in alloc_flags.  That logic and the checks below have the combined
 * affect that:
1844 1845
 *	in_interrupt - any node ok (current task context irrelevant)
 *	GFP_ATOMIC   - any node ok
1846
 *	TIF_MEMDIE   - any node ok
1847 1848
 *	GFP_KERNEL   - any node in enclosing mem_exclusive cpuset ok
 *	GFP_USER     - only nodes in current tasks mems allowed ok.
1849 1850
 *
 * Rule:
1851
 *    Don't call cpuset_zone_allowed_softwall if you can't sleep, unless you
1852 1853
 *    pass in the __GFP_HARDWALL flag set in gfp_flag, which disables
 *    the code that might scan up ancestor cpusets and sleep.
1854
 */
1855

1856
int __cpuset_zone_allowed_softwall(struct zone *z, gfp_t gfp_mask)
L
Linus Torvalds 已提交
1857
{
1858 1859
	int node;			/* node that zone z is on */
	const struct cpuset *cs;	/* current cpuset ancestors */
1860
	int allowed;			/* is allocation in zone z allowed? */
1861

1862
	if (in_interrupt() || (gfp_mask & __GFP_THISNODE))
1863
		return 1;
1864
	node = zone_to_nid(z);
1865
	might_sleep_if(!(gfp_mask & __GFP_HARDWALL));
1866 1867
	if (node_isset(node, current->mems_allowed))
		return 1;
1868 1869 1870 1871 1872 1873
	/*
	 * Allow tasks that have access to memory reserves because they have
	 * been OOM killed to get memory anywhere.
	 */
	if (unlikely(test_thread_flag(TIF_MEMDIE)))
		return 1;
1874 1875 1876
	if (gfp_mask & __GFP_HARDWALL)	/* If hardwall request, stop here */
		return 0;

1877 1878 1879
	if (current->flags & PF_EXITING) /* Let dying task have memory */
		return 1;

1880
	/* Not hardwall and node outside mems_allowed: scan up cpusets */
1881
	mutex_lock(&callback_mutex);
1882 1883

	task_lock(current);
1884
	cs = nearest_exclusive_ancestor(task_cs(current));
1885 1886
	task_unlock(current);

1887
	allowed = node_isset(node, cs->mems_allowed);
1888
	mutex_unlock(&callback_mutex);
1889
	return allowed;
L
Linus Torvalds 已提交
1890 1891
}

1892 1893 1894 1895 1896 1897 1898
/*
 * cpuset_zone_allowed_hardwall - Can we allocate on zone z's memory node?
 * @z: is this zone on an allowed node?
 * @gfp_mask: memory allocation flags
 *
 * If we're in interrupt, yes, we can always allocate.
 * If __GFP_THISNODE is set, yes, we can always allocate.  If zone
1899 1900 1901
 * z's node is in our tasks mems_allowed, yes.   If the task has been
 * OOM killed and has access to memory reserves as specified by the
 * TIF_MEMDIE flag, yes.  Otherwise, no.
1902 1903 1904 1905 1906 1907 1908 1909 1910 1911 1912 1913 1914 1915 1916 1917 1918 1919 1920 1921 1922 1923 1924
 *
 * The __GFP_THISNODE placement logic is really handled elsewhere,
 * by forcibly using a zonelist starting at a specified node, and by
 * (in get_page_from_freelist()) refusing to consider the zones for
 * any node on the zonelist except the first.  By the time any such
 * calls get to this routine, we should just shut up and say 'yes'.
 *
 * Unlike the cpuset_zone_allowed_softwall() variant, above,
 * this variant requires that the zone be in the current tasks
 * mems_allowed or that we're in interrupt.  It does not scan up the
 * cpuset hierarchy for the nearest enclosing mem_exclusive cpuset.
 * It never sleeps.
 */

int __cpuset_zone_allowed_hardwall(struct zone *z, gfp_t gfp_mask)
{
	int node;			/* node that zone z is on */

	if (in_interrupt() || (gfp_mask & __GFP_THISNODE))
		return 1;
	node = zone_to_nid(z);
	if (node_isset(node, current->mems_allowed))
		return 1;
D
Daniel Walker 已提交
1925 1926 1927 1928 1929 1930
	/*
	 * Allow tasks that have access to memory reserves because they have
	 * been OOM killed to get memory anywhere.
	 */
	if (unlikely(test_thread_flag(TIF_MEMDIE)))
		return 1;
1931 1932 1933
	return 0;
}

P
Paul Jackson 已提交
1934 1935 1936
/**
 * cpuset_lock - lock out any changes to cpuset structures
 *
1937
 * The out of memory (oom) code needs to mutex_lock cpusets
P
Paul Jackson 已提交
1938
 * from being changed while it scans the tasklist looking for a
1939
 * task in an overlapping cpuset.  Expose callback_mutex via this
P
Paul Jackson 已提交
1940 1941
 * cpuset_lock() routine, so the oom code can lock it, before
 * locking the task list.  The tasklist_lock is a spinlock, so
1942
 * must be taken inside callback_mutex.
P
Paul Jackson 已提交
1943 1944 1945 1946
 */

void cpuset_lock(void)
{
1947
	mutex_lock(&callback_mutex);
P
Paul Jackson 已提交
1948 1949 1950 1951 1952 1953 1954 1955 1956 1957
}

/**
 * cpuset_unlock - release lock on cpuset changes
 *
 * Undo the lock taken in a previous cpuset_lock() call.
 */

void cpuset_unlock(void)
{
1958
	mutex_unlock(&callback_mutex);
P
Paul Jackson 已提交
1959 1960
}

1961 1962 1963 1964 1965 1966 1967 1968 1969 1970 1971 1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998
/**
 * cpuset_mem_spread_node() - On which node to begin search for a page
 *
 * If a task is marked PF_SPREAD_PAGE or PF_SPREAD_SLAB (as for
 * tasks in a cpuset with is_spread_page or is_spread_slab set),
 * and if the memory allocation used cpuset_mem_spread_node()
 * to determine on which node to start looking, as it will for
 * certain page cache or slab cache pages such as used for file
 * system buffers and inode caches, then instead of starting on the
 * local node to look for a free page, rather spread the starting
 * node around the tasks mems_allowed nodes.
 *
 * We don't have to worry about the returned node being offline
 * because "it can't happen", and even if it did, it would be ok.
 *
 * The routines calling guarantee_online_mems() are careful to
 * only set nodes in task->mems_allowed that are online.  So it
 * should not be possible for the following code to return an
 * offline node.  But if it did, that would be ok, as this routine
 * is not returning the node where the allocation must be, only
 * the node where the search should start.  The zonelist passed to
 * __alloc_pages() will include all nodes.  If the slab allocator
 * is passed an offline node, it will fall back to the local node.
 * See kmem_cache_alloc_node().
 */

int cpuset_mem_spread_node(void)
{
	int node;

	node = next_node(current->cpuset_mem_spread_rotor, current->mems_allowed);
	if (node == MAX_NUMNODES)
		node = first_node(current->mems_allowed);
	current->cpuset_mem_spread_rotor = node;
	return node;
}
EXPORT_SYMBOL_GPL(cpuset_mem_spread_node);

1999
/**
2000 2001 2002 2003 2004 2005 2006 2007
 * cpuset_mems_allowed_intersects - Does @tsk1's mems_allowed intersect @tsk2's?
 * @tsk1: pointer to task_struct of some task.
 * @tsk2: pointer to task_struct of some other task.
 *
 * Description: Return true if @tsk1's mems_allowed intersects the
 * mems_allowed of @tsk2.  Used by the OOM killer to determine if
 * one of the task's memory usage might impact the memory available
 * to the other.
2008 2009
 **/

2010 2011
int cpuset_mems_allowed_intersects(const struct task_struct *tsk1,
				   const struct task_struct *tsk2)
2012
{
2013
	return nodes_intersects(tsk1->mems_allowed, tsk2->mems_allowed);
2014 2015
}

2016 2017 2018 2019 2020 2021
/*
 * Collection of memory_pressure is suppressed unless
 * this flag is enabled by writing "1" to the special
 * cpuset file 'memory_pressure_enabled' in the root cpuset.
 */

2022
int cpuset_memory_pressure_enabled __read_mostly;
2023 2024 2025 2026 2027 2028 2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044

/**
 * cpuset_memory_pressure_bump - keep stats of per-cpuset reclaims.
 *
 * Keep a running average of the rate of synchronous (direct)
 * page reclaim efforts initiated by tasks in each cpuset.
 *
 * This represents the rate at which some task in the cpuset
 * ran low on memory on all nodes it was allowed to use, and
 * had to enter the kernels page reclaim code in an effort to
 * create more free memory by tossing clean pages or swapping
 * or writing dirty pages.
 *
 * Display to user space in the per-cpuset read-only file
 * "memory_pressure".  Value displayed is an integer
 * representing the recent rate of entry into the synchronous
 * (direct) page reclaim by any task attached to the cpuset.
 **/

void __cpuset_memory_pressure_bump(void)
{
	task_lock(current);
2045
	fmeter_markevent(&task_cs(current)->fmeter);
2046 2047 2048
	task_unlock(current);
}

2049
#ifdef CONFIG_PROC_PID_CPUSET
L
Linus Torvalds 已提交
2050 2051 2052 2053
/*
 * proc_cpuset_show()
 *  - Print tasks cpuset path into seq_file.
 *  - Used for /proc/<pid>/cpuset.
2054 2055
 *  - No need to task_lock(tsk) on this tsk->cpuset reference, as it
 *    doesn't really matter if tsk->cpuset changes after we read it,
2056
 *    and we take manage_mutex, keeping attach_task() from changing it
2057 2058 2059
 *    anyway.  No need to check that tsk->cpuset != NULL, thanks to
 *    the_top_cpuset_hack in cpuset_exit(), which sets an exiting tasks
 *    cpuset to top_cpuset.
L
Linus Torvalds 已提交
2060
 */
P
Paul Jackson 已提交
2061
static int proc_cpuset_show(struct seq_file *m, void *unused_v)
L
Linus Torvalds 已提交
2062
{
2063
	struct pid *pid;
L
Linus Torvalds 已提交
2064 2065
	struct task_struct *tsk;
	char *buf;
2066
	struct cgroup_subsys_state *css;
2067
	int retval;
L
Linus Torvalds 已提交
2068

2069
	retval = -ENOMEM;
L
Linus Torvalds 已提交
2070 2071
	buf = kmalloc(PAGE_SIZE, GFP_KERNEL);
	if (!buf)
2072 2073 2074
		goto out;

	retval = -ESRCH;
2075 2076
	pid = m->private;
	tsk = get_pid_task(pid, PIDTYPE_PID);
2077 2078
	if (!tsk)
		goto out_free;
L
Linus Torvalds 已提交
2079

2080
	retval = -EINVAL;
2081 2082 2083
	cgroup_lock();
	css = task_subsys_state(tsk, cpuset_subsys_id);
	retval = cgroup_path(css->cgroup, buf, PAGE_SIZE);
L
Linus Torvalds 已提交
2084
	if (retval < 0)
2085
		goto out_unlock;
L
Linus Torvalds 已提交
2086 2087
	seq_puts(m, buf);
	seq_putc(m, '\n');
2088
out_unlock:
2089
	cgroup_unlock();
2090 2091
	put_task_struct(tsk);
out_free:
L
Linus Torvalds 已提交
2092
	kfree(buf);
2093
out:
L
Linus Torvalds 已提交
2094 2095 2096 2097 2098
	return retval;
}

static int cpuset_open(struct inode *inode, struct file *file)
{
2099 2100
	struct pid *pid = PROC_I(inode)->pid;
	return single_open(file, proc_cpuset_show, pid);
L
Linus Torvalds 已提交
2101 2102
}

2103
const struct file_operations proc_cpuset_operations = {
L
Linus Torvalds 已提交
2104 2105 2106 2107 2108
	.open		= cpuset_open,
	.read		= seq_read,
	.llseek		= seq_lseek,
	.release	= single_release,
};
2109
#endif /* CONFIG_PROC_PID_CPUSET */
L
Linus Torvalds 已提交
2110 2111 2112 2113 2114 2115 2116 2117 2118 2119 2120 2121

/* Display task cpus_allowed, mems_allowed in /proc/<pid>/status file. */
char *cpuset_task_status_allowed(struct task_struct *task, char *buffer)
{
	buffer += sprintf(buffer, "Cpus_allowed:\t");
	buffer += cpumask_scnprintf(buffer, PAGE_SIZE, task->cpus_allowed);
	buffer += sprintf(buffer, "\n");
	buffer += sprintf(buffer, "Mems_allowed:\t");
	buffer += nodemask_scnprintf(buffer, PAGE_SIZE, task->mems_allowed);
	buffer += sprintf(buffer, "\n");
	return buffer;
}