cpuset.c 56.2 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.
7
 *  Copyright (C) 2004-2006 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 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>
Linus Torvalds's avatar
Linus Torvalds committed
43 44
#include <linux/sched.h>
#include <linux/seq_file.h>
45
#include <linux/security.h>
Linus Torvalds's avatar
Linus Torvalds committed
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>
Linus Torvalds's avatar
Linus Torvalds committed
57

58 59 60 61 62
/*
 * Tracks how many cpusets are currently defined in system.
 * When there is only one cpuset (the root cpuset) we can
 * short circuit some hooks.
 */
63
int number_of_cpusets __read_mostly;
64

65 66 67 68
/* Retrieve the cpuset from a cgroup */
struct cgroup_subsys cpuset_subsys;
struct cpuset;

69 70 71 72 73 74 75 76 77
/* 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
78
struct cpuset {
79 80
	struct cgroup_subsys_state css;

Linus Torvalds's avatar
Linus Torvalds committed
81 82 83 84 85 86 87 88 89 90
	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.
	 */
91 92 93
	int mems_generation;

	struct fmeter fmeter;		/* memory_pressure filter */
Linus Torvalds's avatar
Linus Torvalds committed
94 95
};

96 97 98 99 100 101 102 103 104 105 106 107 108 109 110
/* 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);
}


Linus Torvalds's avatar
Linus Torvalds committed
111 112 113 114
/* bits in struct cpuset flags field */
typedef enum {
	CS_CPU_EXCLUSIVE,
	CS_MEM_EXCLUSIVE,
115
	CS_MEMORY_MIGRATE,
116 117
	CS_SPREAD_PAGE,
	CS_SPREAD_SLAB,
Linus Torvalds's avatar
Linus Torvalds committed
118 119 120 121 122
} cpuset_flagbits_t;

/* convenient tests for these bits */
static inline int is_cpu_exclusive(const struct cpuset *cs)
{
123
	return test_bit(CS_CPU_EXCLUSIVE, &cs->flags);
Linus Torvalds's avatar
Linus Torvalds committed
124 125 126 127
}

static inline int is_mem_exclusive(const struct cpuset *cs)
{
128
	return test_bit(CS_MEM_EXCLUSIVE, &cs->flags);
Linus Torvalds's avatar
Linus Torvalds committed
129 130
}

131 132
static inline int is_memory_migrate(const struct cpuset *cs)
{
133
	return test_bit(CS_MEMORY_MIGRATE, &cs->flags);
134 135
}

136 137 138 139 140 141 142 143 144 145
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
146
/*
147
 * Increment this integer everytime any cpuset changes its
Linus Torvalds's avatar
Linus Torvalds committed
148 149 150 151 152 153 154 155 156 157 158 159 160
 * 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.
161 162 163
 *
 * Since cpuset_mems_generation is guarded by manage_mutex,
 * there is no need to mark it atomic.
Linus Torvalds's avatar
Linus Torvalds committed
164
 */
165
static int cpuset_mems_generation;
Linus Torvalds's avatar
Linus Torvalds committed
166 167 168 169 170 171 172 173

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

/*
174 175
 * We have two global cpuset mutexes below.  They can nest.
 * It is ok to first take manage_mutex, then nest callback_mutex.  We also
176 177 178
 * require taking task_lock() when dereferencing a tasks cpuset pointer.
 * See "The task_lock() exception", at the end of this comment.
 *
179 180 181
 * 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
182 183
 * and be able to modify cpusets.  It can perform various checks on
 * the cpuset structure first, knowing nothing will change.  It can
184
 * also allocate memory while just holding manage_mutex.  While it is
185
 * performing these checks, various callback routines can briefly
186 187
 * acquire callback_mutex to query cpusets.  Once it is ready to make
 * the changes, it takes callback_mutex, blocking everyone else.
188 189
 *
 * Calls to the kernel memory allocator can not be made while holding
190
 * callback_mutex, as that would risk double tripping on callback_mutex
191 192 193
 * from one of the callbacks into the cpuset code from within
 * __alloc_pages().
 *
194
 * If a task is only holding callback_mutex, then it has read-only
195 196 197 198 199 200
 * 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.
201
 * So in general, code holding manage_mutex or callback_mutex can't rely
202
 * on the count field not changing.  However, if the count goes to
203
 * zero, then only attach_task(), which holds both mutexes, can
204 205 206
 * 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).
207
 * So code holding manage_mutex or callback_mutex can safely assume that
208
 * if the count is zero, it will stay zero.  Similarly, if a task
209
 * holds manage_mutex or callback_mutex on a cpuset with zero count, it
210
 * knows that the cpuset won't be removed, as cpuset_rmdir() needs
211
 * both of those mutexes.
212 213
 *
 * The cpuset_common_file_write handler for operations that modify
214
 * the cpuset hierarchy holds manage_mutex across the entire operation,
215 216
 * single threading all such cpuset modifications across the system.
 *
217
 * The cpuset_common_file_read() handlers only hold callback_mutex across
218 219 220 221
 * 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
222
 * (usually) take either mutex.  These are the two most performance
223
 * critical pieces of code here.  The exception occurs on cpuset_exit(),
224
 * when a task in a notify_on_release cpuset exits.  Then manage_mutex
225
 * is taken, and if the cpuset count is zero, a usermode call made
Linus Torvalds's avatar
Linus Torvalds committed
226 227 228
 * to /sbin/cpuset_release_agent with the name of the cpuset (path
 * relative to the root of cpuset file system) as the argument.
 *
229 230 231
 * 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
232
 * least one task in the system (init), therefore, top_cpuset
233 234 235 236 237 238 239 240 241
 * 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
242
 * so using both mutexes, however there are several performance
243
 * critical places that need to reference task->cpuset without the
244
 * expense of grabbing a system global mutex.  Therefore except as
245 246 247 248
 * 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.
249 250 251 252 253
 *
 * 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
254 255
 */

256
static DEFINE_MUTEX(callback_mutex);
257

258 259 260
/* 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 */
261 262 263
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
264
{
265 266 267 268 269 270 271 272 273 274 275
	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
276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293
}

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.
 *
294
 * Call with callback_mutex held.
Linus Torvalds's avatar
Linus Torvalds committed
295 296 297 298 299 300 301 302 303 304 305 306 307 308 309
 */

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
310 311 312 313
 * 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
314 315
 *
 * One way or another, we guarantee to return some non-empty subset
316
 * of node_states[N_HIGH_MEMORY].
Linus Torvalds's avatar
Linus Torvalds committed
317
 *
318
 * Call with callback_mutex held.
Linus Torvalds's avatar
Linus Torvalds committed
319 320 321 322
 */

static void guarantee_online_mems(const struct cpuset *cs, nodemask_t *pmask)
{
323 324
	while (cs && !nodes_intersects(cs->mems_allowed,
					node_states[N_HIGH_MEMORY]))
Linus Torvalds's avatar
Linus Torvalds committed
325 326
		cs = cs->parent;
	if (cs)
327 328
		nodes_and(*pmask, cs->mems_allowed,
					node_states[N_HIGH_MEMORY]);
Linus Torvalds's avatar
Linus Torvalds committed
329
	else
330 331
		*pmask = node_states[N_HIGH_MEMORY];
	BUG_ON(!nodes_intersects(*pmask, node_states[N_HIGH_MEMORY]));
Linus Torvalds's avatar
Linus Torvalds committed
332 333
}

334 335 336 337 338 339
/**
 * 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.
340
 *
341 342 343 344
 * 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().
 *
345 346 347 348
 * 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
349
 * current->mm->mmap_sem during call.
350
 *
351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368
 * 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.
369 370 371 372 373
 *
 * 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
374 375
 */

376
void cpuset_update_task_memory_state(void)
Linus Torvalds's avatar
Linus Torvalds committed
377
{
378
	int my_cpusets_mem_gen;
379
	struct task_struct *tsk = current;
380
	struct cpuset *cs;
381

382
	if (task_cs(tsk) == &top_cpuset) {
383 384 385 386
		/* Don't need rcu for top_cpuset.  It's never freed. */
		my_cpusets_mem_gen = top_cpuset.mems_generation;
	} else {
		rcu_read_lock();
387
		my_cpusets_mem_gen = task_cs(current)->mems_generation;
388 389
		rcu_read_unlock();
	}
Linus Torvalds's avatar
Linus Torvalds committed
390

391
	if (my_cpusets_mem_gen != tsk->cpuset_mems_generation) {
392
		mutex_lock(&callback_mutex);
393
		task_lock(tsk);
394
		cs = task_cs(tsk); /* Maybe changed when task not locked */
395 396
		guarantee_online_mems(cs, &tsk->mems_allowed);
		tsk->cpuset_mems_generation = cs->mems_generation;
397 398 399 400 401 402 403 404
		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;
405
		task_unlock(tsk);
406
		mutex_unlock(&callback_mutex);
407
		mpol_rebind_task(tsk, &tsk->mems_allowed);
Linus Torvalds's avatar
Linus Torvalds committed
408 409 410 411 412 413 414 415
	}
}

/*
 * 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
416
 * are only set if the other's are set.  Call holding manage_mutex.
Linus Torvalds's avatar
Linus Torvalds committed
417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433
 */

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
434
 * manage_mutex held.
Linus Torvalds's avatar
Linus Torvalds committed
435 436 437 438 439 440 441 442 443 444 445 446 447 448
 *
 * '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)
{
449
	struct cgroup *cont;
Linus Torvalds's avatar
Linus Torvalds committed
450 451 452
	struct cpuset *c, *par;

	/* Each of our child cpusets must be a subset of us */
453 454
	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
455 456 457 458
			return -EBUSY;
	}

	/* Remaining checks don't apply to root cpuset */
459
	if (cur == &top_cpuset)
Linus Torvalds's avatar
Linus Torvalds committed
460 461
		return 0;

462 463
	par = cur->parent;

Linus Torvalds's avatar
Linus Torvalds committed
464 465 466 467 468
	/* 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 */
469 470
	list_for_each_entry(cont, &par->css.cgroup->children, sibling) {
		c = cgroup_cs(cont);
Linus Torvalds's avatar
Linus Torvalds committed
471 472 473 474 475 476 477 478 479 480 481 482 483
		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;
	}

	return 0;
}

484
/*
485
 * Call with manage_mutex held.  May take callback_mutex during call.
486 487
 */

Linus Torvalds's avatar
Linus Torvalds committed
488 489 490
static int update_cpumask(struct cpuset *cs, char *buf)
{
	struct cpuset trialcs;
491
	int retval;
Linus Torvalds's avatar
Linus Torvalds committed
492

493 494 495 496
	/* 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
497
	trialcs = *cs;
498 499 500 501 502 503 504 505 506 507 508 509 510

	/*
	 * We allow a cpuset's cpus_allowed to be empty; if it has attached
	 * tasks, we'll catch it later when we validate the change and return
	 * -ENOSPC.
	 */
	if (!buf[0] || (buf[0] == '\n' && !buf[1])) {
		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
511
	cpus_and(trialcs.cpus_allowed, trialcs.cpus_allowed, cpu_online_map);
512
	/* cpus_allowed cannot be empty for a cpuset with attached tasks. */
513 514
	if (cgroup_task_count(cs->css.cgroup) &&
	    cpus_empty(trialcs.cpus_allowed))
Linus Torvalds's avatar
Linus Torvalds committed
515 516
		return -ENOSPC;
	retval = validate_change(cs, &trialcs);
517 518
	if (retval < 0)
		return retval;
519
	mutex_lock(&callback_mutex);
520
	cs->cpus_allowed = trialcs.cpus_allowed;
521
	mutex_unlock(&callback_mutex);
522
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
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
/*
 * 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);
570
	guarantee_online_mems(task_cs(tsk),&tsk->mems_allowed);
571 572 573
	mutex_unlock(&callback_mutex);
}

574
/*
575 576 577
 * 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
578 579 580
 * task in the cpuset, rebind any vma mempolicies and if
 * the cpuset is marked 'memory_migrate', migrate the tasks
 * pages to the new memory.
581
 *
582
 * Call with manage_mutex held.  May take callback_mutex during call.
583 584 585
 * 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.
586 587
 */

588 589
static void *cpuset_being_rebound;

Linus Torvalds's avatar
Linus Torvalds committed
590 591 592
static int update_nodemask(struct cpuset *cs, char *buf)
{
	struct cpuset trialcs;
593
	nodemask_t oldmem;
594
	struct task_struct *p;
595 596
	struct mm_struct **mmarray;
	int i, n, ntasks;
597
	int migrate;
598
	int fudge;
Linus Torvalds's avatar
Linus Torvalds committed
599
	int retval;
600
	struct cgroup_iter it;
Linus Torvalds's avatar
Linus Torvalds committed
601

602 603 604 605
	/*
	 * top_cpuset.mems_allowed tracks node_stats[N_HIGH_MEMORY];
	 * it's read-only
	 */
606 607 608
	if (cs == &top_cpuset)
		return -EACCES;

Linus Torvalds's avatar
Linus Torvalds committed
609
	trialcs = *cs;
610 611 612 613 614 615 616 617 618 619 620 621

	/*
	 * We allow a cpuset's mems_allowed to be empty; if it has attached
	 * tasks, we'll catch it later when we validate the change and return
	 * -ENOSPC.
	 */
	if (!buf[0] || (buf[0] == '\n' && !buf[1])) {
		nodes_clear(trialcs.mems_allowed);
	} else {
		retval = nodelist_parse(buf, trialcs.mems_allowed);
		if (retval < 0)
			goto done;
622 623 624 625 626 627 628 629
		if (!nodes_intersects(trialcs.mems_allowed,
						node_states[N_HIGH_MEMORY])) {
			/*
			 * error if only memoryless nodes specified.
			 */
			retval = -ENOSPC;
			goto done;
		}
630
	}
631 632 633 634 635 636
	/*
	 * Exclude memoryless nodes.  We know that trialcs.mems_allowed
	 * contains at least one node with memory.
	 */
	nodes_and(trialcs.mems_allowed, trialcs.mems_allowed,
						node_states[N_HIGH_MEMORY]);
637 638 639 640 641
	oldmem = cs->mems_allowed;
	if (nodes_equal(oldmem, trialcs.mems_allowed)) {
		retval = 0;		/* Too easy - nothing to do */
		goto done;
	}
642
	/* mems_allowed cannot be empty for a cpuset with attached tasks. */
643 644
	if (cgroup_task_count(cs->css.cgroup) &&
	    nodes_empty(trialcs.mems_allowed)) {
645 646
		retval = -ENOSPC;
		goto done;
Linus Torvalds's avatar
Linus Torvalds committed
647
	}
648 649 650 651
	retval = validate_change(cs, &trialcs);
	if (retval < 0)
		goto done;

652
	mutex_lock(&callback_mutex);
653
	cs->mems_allowed = trialcs.mems_allowed;
654
	cs->mems_generation = cpuset_mems_generation++;
655
	mutex_unlock(&callback_mutex);
656

657
	cpuset_being_rebound = cs;		/* causes mpol_copy() rebind */
658 659 660 661 662 663 664 665 666 667 668 669 670

	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) {
671
		ntasks = cgroup_task_count(cs->css.cgroup);  /* guess */
672 673 674 675
		ntasks += fudge;
		mmarray = kmalloc(ntasks * sizeof(*mmarray), GFP_KERNEL);
		if (!mmarray)
			goto done;
676
		read_lock(&tasklist_lock);		/* block fork */
677
		if (cgroup_task_count(cs->css.cgroup) <= ntasks)
678
			break;				/* got enough */
679
		read_unlock(&tasklist_lock);		/* try again */
680 681 682 683 684 685
		kfree(mmarray);
	}

	n = 0;

	/* Load up mmarray[] with mm reference for each task in cpuset. */
686 687
	cgroup_iter_start(cs->css.cgroup, &it);
	while ((p = cgroup_iter_next(cs->css.cgroup, &it))) {
688 689 690 691 692
		struct mm_struct *mm;

		if (n >= ntasks) {
			printk(KERN_WARNING
				"Cpuset mempolicy rebind incomplete.\n");
693
			break;
694 695 696 697 698
		}
		mm = get_task_mm(p);
		if (!mm)
			continue;
		mmarray[n++] = mm;
699 700
	}
	cgroup_iter_end(cs->css.cgroup, &it);
701
	read_unlock(&tasklist_lock);
702 703 704 705 706 707 708 709 710

	/*
	 * 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
711
	 * cpuset manage_mutex, we know that no other rebind effort will
712 713
	 * be contending for the global variable cpuset_being_rebound.
	 * It's ok if we rebind the same mm twice; mpol_rebind_mm()
714
	 * is idempotent.  Also migrate pages in each mm to new nodes.
715
	 */
716
	migrate = is_memory_migrate(cs);
717 718 719 720
	for (i = 0; i < n; i++) {
		struct mm_struct *mm = mmarray[i];

		mpol_rebind_mm(mm, &cs->mems_allowed);
721 722
		if (migrate)
			cpuset_migrate_mm(mm, &oldmem, &cs->mems_allowed);
723 724 725 726 727
		mmput(mm);
	}

	/* We're done rebinding vma's to this cpusets new mems_allowed. */
	kfree(mmarray);
728
	cpuset_being_rebound = NULL;
729
	retval = 0;
730
done:
Linus Torvalds's avatar
Linus Torvalds committed
731 732 733
	return retval;
}

734 735 736 737 738
int current_cpuset_is_being_rebound(void)
{
	return task_cs(current) == cpuset_being_rebound;
}

739
/*
740
 * Call with manage_mutex held.
741 742 743 744 745 746 747 748 749 750 751
 */

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
752 753 754
/*
 * 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,
755 756
 *				CS_NOTIFY_ON_RELEASE, CS_MEMORY_MIGRATE,
 *				CS_SPREAD_PAGE, CS_SPREAD_SLAB)
Linus Torvalds's avatar
Linus Torvalds committed
757 758
 * cs:	the cpuset to update
 * buf:	the buffer where we read the 0 or 1
759
 *
760
 * Call with manage_mutex held.
Linus Torvalds's avatar
Linus Torvalds committed
761 762 763 764 765 766
 */

static int update_flag(cpuset_flagbits_t bit, struct cpuset *cs, char *buf)
{
	int turning_on;
	struct cpuset trialcs;
767
	int err;
Linus Torvalds's avatar
Linus Torvalds committed
768 769 770 771 772 773 774 775 776 777

	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);
778 779
	if (err < 0)
		return err;
780
	mutex_lock(&callback_mutex);
781
	cs->flags = trialcs.flags;
782
	mutex_unlock(&callback_mutex);
783 784

	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
785 786
}

787
/*
788
 * Frequency meter - How fast is some event occurring?
789 790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884
 *
 * 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;
}

885 886
static int cpuset_can_attach(struct cgroup_subsys *ss,
			     struct cgroup *cont, struct task_struct *tsk)
Linus Torvalds's avatar
Linus Torvalds committed
887
{
888
	struct cpuset *cs = cgroup_cs(cont);
Linus Torvalds's avatar
Linus Torvalds committed
889 890 891 892

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

893 894
	return security_task_setscheduler(tsk, 0, NULL);
}
Linus Torvalds's avatar
Linus Torvalds committed
895

896 897 898 899 900 901 902 903 904
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);
905

906
	mutex_lock(&callback_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
907 908
	guarantee_online_cpus(cs, &cpus);
	set_cpus_allowed(tsk, cpus);
909
	mutex_unlock(&callback_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
910

911 912
	from = oldcs->mems_allowed;
	to = cs->mems_allowed;
913 914 915
	mm = get_task_mm(tsk);
	if (mm) {
		mpol_rebind_mm(mm, &to);
916
		if (is_memory_migrate(cs))
917
			cpuset_migrate_mm(mm, &from, &to);
918 919 920
		mmput(mm);
	}

Linus Torvalds's avatar
Linus Torvalds committed
921 922 923 924 925
}

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

typedef enum {
926
	FILE_MEMORY_MIGRATE,
Linus Torvalds's avatar
Linus Torvalds committed
927 928 929 930
	FILE_CPULIST,
	FILE_MEMLIST,
	FILE_CPU_EXCLUSIVE,
	FILE_MEM_EXCLUSIVE,
931 932
	FILE_MEMORY_PRESSURE_ENABLED,
	FILE_MEMORY_PRESSURE,
933 934
	FILE_SPREAD_PAGE,
	FILE_SPREAD_SLAB,
Linus Torvalds's avatar
Linus Torvalds committed
935 936
} cpuset_filetype_t;

937 938 939
static ssize_t cpuset_common_file_write(struct cgroup *cont,
					struct cftype *cft,
					struct file *file,
940
					const char __user *userbuf,
Linus Torvalds's avatar
Linus Torvalds committed
941 942
					size_t nbytes, loff_t *unused_ppos)
{
943
	struct cpuset *cs = cgroup_cs(cont);
Linus Torvalds's avatar
Linus Torvalds committed
944 945 946 947 948
	cpuset_filetype_t type = cft->private;
	char *buffer;
	int retval = 0;

	/* Crude upper limit on largest legitimate cpulist user might write. */
949
	if (nbytes > 100 + 6 * max(NR_CPUS, MAX_NUMNODES))
Linus Torvalds's avatar
Linus Torvalds committed
950 951 952 953 954 955 956 957 958 959 960 961
		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 */

962
	cgroup_lock();
Linus Torvalds's avatar
Linus Torvalds committed
963

964
	if (cgroup_is_removed(cont)) {
Linus Torvalds's avatar
Linus Torvalds committed
965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981
		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;
982 983 984
	case FILE_MEMORY_MIGRATE:
		retval = update_flag(CS_MEMORY_MIGRATE, cs, buffer);
		break;
985 986 987 988 989 990
	case FILE_MEMORY_PRESSURE_ENABLED:
		retval = update_memory_pressure_enabled(cs, buffer);
		break;
	case FILE_MEMORY_PRESSURE:
		retval = -EACCES;
		break;
991 992
	case FILE_SPREAD_PAGE:
		retval = update_flag(CS_SPREAD_PAGE, cs, buffer);
993
		cs->mems_generation = cpuset_mems_generation++;
994 995 996
		break;
	case FILE_SPREAD_SLAB:
		retval = update_flag(CS_SPREAD_SLAB, cs, buffer);
997
		cs->mems_generation = cpuset_mems_generation++;
998
		break;
Linus Torvalds's avatar
Linus Torvalds committed
999 1000 1001 1002 1003 1004 1005 1006
	default:
		retval = -EINVAL;
		goto out2;
	}

	if (retval == 0)
		retval = nbytes;
out2:
1007
	cgroup_unlock();
Linus Torvalds's avatar
Linus Torvalds committed
1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028
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;

1029
	mutex_lock(&callback_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
1030
	mask = cs->cpus_allowed;
1031
	mutex_unlock(&callback_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
1032 1033 1034 1035 1036 1037 1038 1039

	return cpulist_scnprintf(page, PAGE_SIZE, mask);
}

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

1040
	mutex_lock(&callback_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
1041
	mask = cs->mems_allowed;
1042
	mutex_unlock(&callback_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
1043 1044 1045 1046

	return nodelist_scnprintf(page, PAGE_SIZE, mask);
}

1047 1048 1049 1050 1051
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
1052
{
1053
	struct cpuset *cs = cgroup_cs(cont);
Linus Torvalds's avatar
Linus Torvalds committed
1054 1055 1056 1057 1058
	cpuset_filetype_t type = cft->private;
	char *page;
	ssize_t retval = 0;
	char *s;

1059
	if (!(page = (char *)__get_free_page(GFP_TEMPORARY)))
Linus Torvalds's avatar
Linus Torvalds committed
1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070 1071 1072 1073 1074 1075 1076
		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;
1077 1078 1079
	case FILE_MEMORY_MIGRATE:
		*s++ = is_memory_migrate(cs) ? '1' : '0';
		break;
1080 1081 1082 1083 1084 1085
	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;
1086 1087 1088 1089 1090 1091
	case FILE_SPREAD_PAGE:
		*s++ = is_spread_page(cs) ? '1' : '0';
		break;
	case FILE_SPREAD_SLAB:
		*s++ = is_spread_slab(cs) ? '1' : '0';
		break;
Linus Torvalds's avatar
Linus Torvalds committed
1092 1093 1094 1095 1096 1097
	default:
		retval = -EINVAL;
		goto out;
	}
	*s++ = '\n';

Al Viro's avatar
Al Viro committed
1098
	retval = simple_read_from_buffer(buf, nbytes, ppos, page, s - page);
Linus Torvalds's avatar
Linus Torvalds committed
1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113
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",
1114 1115
	.read = cpuset_common_file_read,
	.write = cpuset_common_file_write,
Linus Torvalds's avatar
Linus Torvalds committed
1116 1117 1118 1119 1120
	.private = FILE_CPULIST,
};

static struct cftype cft_mems = {
	.name = "mems",
1121 1122
	.read = cpuset_common_file_read,
	.write = cpuset_common_file_write,
Linus Torvalds's avatar
Linus Torvalds committed
1123 1124 1125 1126 1127
	.private = FILE_MEMLIST,
};

static struct cftype cft_cpu_exclusive = {
	.name = "cpu_exclusive",
1128 1129
	.read = cpuset_common_file_read,
	.write = cpuset_common_file_write,
Linus Torvalds's avatar
Linus Torvalds committed
1130 1131 1132 1133 1134
	.private = FILE_CPU_EXCLUSIVE,
};

static struct cftype cft_mem_exclusive = {
	.name = "mem_exclusive",
1135 1136
	.read = cpuset_common_file_read,
	.write = cpuset_common_file_write,
Linus Torvalds's avatar
Linus Torvalds committed
1137 1138 1139
	.private = FILE_MEM_EXCLUSIVE,
};

1140 1141
static struct cftype cft_memory_migrate = {
	.name = "memory_migrate",
1142 1143
	.read = cpuset_common_file_read,
	.write = cpuset_common_file_write,
1144 1145 1146
	.private = FILE_MEMORY_MIGRATE,
};

1147 1148
static struct cftype cft_memory_pressure_enabled = {
	.name = "memory_pressure_enabled",
1149 1150
	.read = cpuset_common_file_read,
	.write = cpuset_common_file_write,
1151 1152 1153 1154 1155
	.private = FILE_MEMORY_PRESSURE_ENABLED,
};

static struct cftype cft_memory_pressure = {
	.name = "memory_pressure",
1156 1157
	.read = cpuset_common_file_read,
	.write = cpuset_common_file_write,
1158 1159 1160
	.private = FILE_MEMORY_PRESSURE,
};

1161 1162
static struct cftype cft_spread_page = {
	.name = "memory_spread_page",
1163 1164
	.read = cpuset_common_file_read,
	.write = cpuset_common_file_write,
1165 1166 1167 1168 1169
	.private = FILE_SPREAD_PAGE,
};

static struct cftype cft_spread_slab = {
	.name = "memory_spread_slab",
1170 1171
	.read = cpuset_common_file_read,
	.write = cpuset_common_file_write,
1172 1173 1174
	.private = FILE_SPREAD_SLAB,
};

1175
static int cpuset_populate(struct cgroup_subsys *ss, struct cgroup *cont)
Linus Torvalds's avatar
Linus Torvalds committed
1176 1177 1178
{
	int err;

1179
	if ((err = cgroup_add_file(cont, ss, &cft_cpus)) < 0)
Linus Torvalds's avatar
Linus Torvalds committed
1180
		return err;
1181
	if ((err = cgroup_add_file(cont, ss, &cft_mems)) < 0)
Linus Torvalds's avatar
Linus Torvalds committed
1182
		return err;
1183
	if ((err = cgroup_add_file(cont, ss, &cft_cpu_exclusive)) < 0)
Linus Torvalds's avatar
Linus Torvalds committed
1184
		return err;
1185
	if ((err = cgroup_add_file(cont, ss, &cft_mem_exclusive)) < 0)
Linus Torvalds's avatar
Linus Torvalds committed
1186
		return err;
1187
	if ((err = cgroup_add_file(cont, ss, &cft_memory_migrate)) < 0)
Linus Torvalds's avatar
Linus Torvalds committed
1188
		return err;
1189
	if ((err = cgroup_add_file(cont, ss, &cft_memory_pressure)) < 0)
1190
		return err;
1191
	if ((err = cgroup_add_file(cont, ss, &cft_spread_page)) < 0)
1192
		return err;
1193
	if ((err = cgroup_add_file(cont, ss, &cft_spread_slab)) < 0)
Linus Torvalds's avatar
Linus Torvalds committed
1194
		return err;
1195 1196 1197 1198
	/* memory_pressure_enabled is in root cpuset only */
	if (err == 0 && !cont->parent)
		err = cgroup_add_file(cont, ss,
					 &cft_memory_pressure_enabled);
Linus Torvalds's avatar
Linus Torvalds committed
1199 1200 1201
	return 0;
}

1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231 1232 1233 1234 1235 1236 1237
/*
 * 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;
}

Linus Torvalds's avatar
Linus Torvalds committed
1238 1239 1240 1241 1242 1243
/*
 *	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
 *
1244
 *	Must be called with the mutex on the parent inode held
Linus Torvalds's avatar
Linus Torvalds committed
1245 1246
 */

1247 1248 1249
static struct cgroup_subsys_state *cpuset_create(
	struct cgroup_subsys *ss,
	struct cgroup *cont)
Linus Torvalds's avatar
Linus Torvalds committed
1250 1251
{
	struct cpuset *cs;
1252
	struct cpuset *parent;
Linus Torvalds's avatar
Linus Torvalds committed
1253

1254 1255 1256 1257 1258 1259
	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);
Linus Torvalds's avatar
Linus Torvalds committed
1260 1261
	cs = kmalloc(sizeof(*cs), GFP_KERNEL);
	if (!cs)
1262
		return ERR_PTR(-ENOMEM);
Linus Torvalds's avatar
Linus Torvalds committed
1263

1264
	cpuset_update_task_memory_state();
Linus Torvalds's avatar
Linus Torvalds committed
1265
	cs->flags = 0;
1266 1267 1268 1269
	if (is_spread_page(parent))
		set_bit(CS_SPREAD_PAGE, &cs->flags);
	if (is_spread_slab(parent))
		set_bit(CS_SPREAD_SLAB, &cs->flags);
Linus Torvalds's avatar
Linus Torvalds committed
1270 1271
	cs->cpus_allowed = CPU_MASK_NONE;
	cs->mems_allowed = NODE_MASK_NONE;
1272
	cs->mems_generation = cpuset_mems_generation++;
1273
	fmeter_init(&cs->fmeter);
Linus Torvalds's avatar
Linus Torvalds committed
1274 1275

	cs->parent = parent;
1276
	number_of_cpusets++;