cpuset.c 68.9 KB
Newer Older
Linus Torvalds's avatar
Linus Torvalds committed
1 2 3 4 5 6
/*
 *  kernel/cpuset.c
 *
 *  Processor and Memory placement constraints for sets of tasks.
 *
 *  Copyright (C) 2003 BULL SA.
Paul Jackson's avatar
Paul Jackson committed
7
 *  Copyright (C) 2004-2007 Silicon Graphics, Inc.
8
 *  Copyright (C) 2006 Google, Inc
Linus Torvalds's avatar
Linus Torvalds committed
9 10 11 12
 *
 *  Portions derived from Patrick Mochel's sysfs code.
 *  sysfs is Copyright (c) 2001-3 Patrick Mochel
 *
13
 *  2003-10-10 Written by Simon Derr.
Linus Torvalds's avatar
Linus Torvalds committed
14
 *  2003-10-22 Updates by Stephen Hemminger.
15
 *  2004 May-July Rework by Paul Jackson.
16
 *  2006 Rework by Paul Menage to use generic cgroups
Linus Torvalds's avatar
Linus Torvalds committed
17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34
 *
 *  This file is subject to the terms and conditions of the GNU General Public
 *  License.  See the file COPYING in the main directory of the Linux
 *  distribution for more details.
 */

#include <linux/cpu.h>
#include <linux/cpumask.h>
#include <linux/cpuset.h>
#include <linux/err.h>
#include <linux/errno.h>
#include <linux/file.h>
#include <linux/fs.h>
#include <linux/init.h>
#include <linux/interrupt.h>
#include <linux/kernel.h>
#include <linux/kmod.h>
#include <linux/list.h>
35
#include <linux/mempolicy.h>
Linus Torvalds's avatar
Linus Torvalds committed
36 37 38 39 40 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>
Paul Jackson's avatar
Paul Jackson committed
57
#include <linux/kfifo.h>
58 59
#include <linux/workqueue.h>
#include <linux/cgroup.h>
Linus Torvalds's avatar
Linus Torvalds committed
60

61 62 63 64 65
/*
 * Tracks how many cpusets are currently defined in system.
 * When there is only one cpuset (the root cpuset) we can
 * short circuit some hooks.
 */
66
int number_of_cpusets __read_mostly;
67

68
/* Forward declare cgroup structures */
69 70 71
struct cgroup_subsys cpuset_subsys;
struct cpuset;

72 73 74 75 76 77 78 79 80
/* 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
81
struct cpuset {
82 83
	struct cgroup_subsys_state css;

Linus Torvalds's avatar
Linus Torvalds committed
84 85 86 87 88 89 90 91 92 93
	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.
	 */
94 95 96
	int mems_generation;

	struct fmeter fmeter;		/* memory_pressure filter */
Paul Jackson's avatar
Paul Jackson committed
97 98 99

	/* partition number for rebuild_sched_domains() */
	int pn;
100

101 102 103
	/* for custom sched domain */
	int relax_domain_level;

104 105
	/* used for walking a cpuset heirarchy */
	struct list_head stack_list;
Linus Torvalds's avatar
Linus Torvalds committed
106 107
};

108 109 110 111 112 113 114 115 116 117 118 119 120
/* 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);
}
121 122 123 124
struct cpuset_hotplug_scanner {
	struct cgroup_scanner scan;
	struct cgroup *to;
};
125

Linus Torvalds's avatar
Linus Torvalds committed
126 127 128 129
/* bits in struct cpuset flags field */
typedef enum {
	CS_CPU_EXCLUSIVE,
	CS_MEM_EXCLUSIVE,
130
	CS_MEMORY_MIGRATE,
Paul Jackson's avatar
Paul Jackson committed
131
	CS_SCHED_LOAD_BALANCE,
132 133
	CS_SPREAD_PAGE,
	CS_SPREAD_SLAB,
Linus Torvalds's avatar
Linus Torvalds committed
134 135 136 137 138
} cpuset_flagbits_t;

/* convenient tests for these bits */
static inline int is_cpu_exclusive(const struct cpuset *cs)
{
139
	return test_bit(CS_CPU_EXCLUSIVE, &cs->flags);
Linus Torvalds's avatar
Linus Torvalds committed
140 141 142 143
}

static inline int is_mem_exclusive(const struct cpuset *cs)
{
144
	return test_bit(CS_MEM_EXCLUSIVE, &cs->flags);
Linus Torvalds's avatar
Linus Torvalds committed
145 146
}

Paul Jackson's avatar
Paul Jackson committed
147 148 149 150 151
static inline int is_sched_load_balance(const struct cpuset *cs)
{
	return test_bit(CS_SCHED_LOAD_BALANCE, &cs->flags);
}

152 153
static inline int is_memory_migrate(const struct cpuset *cs)
{
154
	return test_bit(CS_MEMORY_MIGRATE, &cs->flags);
155 156
}

157 158 159 160 161 162 163 164 165 166
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
167
/*
168
 * Increment this integer everytime any cpuset changes its
Linus Torvalds's avatar
Linus Torvalds committed
169 170 171 172
 * 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.
 *
173
 * A single, global generation is needed because cpuset_attach_task() could
Linus Torvalds's avatar
Linus Torvalds committed
174 175 176 177
 * 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
178
 * modify another's memory placement.  So we must enable every task,
Linus Torvalds's avatar
Linus Torvalds committed
179 180 181
 * 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.
182
 *
183
 * Since writes to cpuset_mems_generation are guarded by the cgroup lock
184
 * there is no need to mark it atomic.
Linus Torvalds's avatar
Linus Torvalds committed
185
 */
186
static int cpuset_mems_generation;
Linus Torvalds's avatar
Linus Torvalds committed
187 188 189 190 191 192 193 194

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

/*
195 196 197 198 199 200 201
 * There are two global mutexes guarding cpuset structures.  The first
 * is the main control groups cgroup_mutex, accessed via
 * cgroup_lock()/cgroup_unlock().  The second is the cpuset-specific
 * callback_mutex, below. They can nest.  It is ok to first take
 * cgroup_mutex, then nest callback_mutex.  We also require taking
 * task_lock() when dereferencing a task's cpuset pointer.  See "The
 * task_lock() exception", at the end of this comment.
202
 *
203
 * A task must hold both mutexes to modify cpusets.  If a task
204
 * holds cgroup_mutex, then it blocks others wanting that mutex,
205
 * ensuring that it is the only task able to also acquire callback_mutex
206 207
 * and be able to modify cpusets.  It can perform various checks on
 * the cpuset structure first, knowing nothing will change.  It can
208
 * also allocate memory while just holding cgroup_mutex.  While it is
209
 * performing these checks, various callback routines can briefly
210 211
 * acquire callback_mutex to query cpusets.  Once it is ready to make
 * the changes, it takes callback_mutex, blocking everyone else.
212 213
 *
 * Calls to the kernel memory allocator can not be made while holding
214
 * callback_mutex, as that would risk double tripping on callback_mutex
215 216 217
 * from one of the callbacks into the cpuset code from within
 * __alloc_pages().
 *
218
 * If a task is only holding callback_mutex, then it has read-only
219 220 221 222 223 224
 * 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.
 *
 * The cpuset_common_file_write handler for operations that modify
225
 * the cpuset hierarchy holds cgroup_mutex across the entire operation,
226 227
 * single threading all such cpuset modifications across the system.
 *
228
 * The cpuset_common_file_read() handlers only hold callback_mutex across
229 230 231
 * small pieces of code, such as when reading out possibly multi-word
 * cpumasks and nodemasks.
 *
232 233
 * Accessing a task's cpuset should be done in accordance with the
 * guidelines for accessing subsystem state in kernel/cgroup.c
Linus Torvalds's avatar
Linus Torvalds committed
234 235
 */

236
static DEFINE_MUTEX(callback_mutex);
237

238 239 240
/* 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 */
241 242 243
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
244
{
245 246 247 248 249 250 251 252 253 254 255
	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
256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273
}

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.
 *
274
 * Call with callback_mutex held.
Linus Torvalds's avatar
Linus Torvalds committed
275 276 277 278 279 280 281 282 283 284 285 286 287 288 289
 */

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
290 291 292 293
 * 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
294 295
 *
 * One way or another, we guarantee to return some non-empty subset
296
 * of node_states[N_HIGH_MEMORY].
Linus Torvalds's avatar
Linus Torvalds committed
297
 *
298
 * Call with callback_mutex held.
Linus Torvalds's avatar
Linus Torvalds committed
299 300 301 302
 */

static void guarantee_online_mems(const struct cpuset *cs, nodemask_t *pmask)
{
303 304
	while (cs && !nodes_intersects(cs->mems_allowed,
					node_states[N_HIGH_MEMORY]))
Linus Torvalds's avatar
Linus Torvalds committed
305 306
		cs = cs->parent;
	if (cs)
307 308
		nodes_and(*pmask, cs->mems_allowed,
					node_states[N_HIGH_MEMORY]);
Linus Torvalds's avatar
Linus Torvalds committed
309
	else
310 311
		*pmask = node_states[N_HIGH_MEMORY];
	BUG_ON(!nodes_intersects(*pmask, node_states[N_HIGH_MEMORY]));
Linus Torvalds's avatar
Linus Torvalds committed
312 313
}

314 315 316 317 318 319
/**
 * 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.
320
 *
321 322 323 324
 * 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().
 *
325
 * Call without callback_mutex or task_lock() held.  May be
326 327
 * called with or without cgroup_mutex held.  Thanks in part to
 * 'the_top_cpuset_hack', the task's cpuset pointer will never
David Rientjes's avatar
David Rientjes committed
328 329
 * be NULL.  This routine also might acquire callback_mutex during
 * call.
330
 *
331 332
 * Reading current->cpuset->mems_generation doesn't need task_lock
 * to guard the current->cpuset derefence, because it is guarded
333
 * from concurrent freeing of current->cpuset using RCU.
334 335 336 337 338 339 340 341 342 343 344 345 346 347
 *
 * 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.
348 349 350 351 352
 *
 * 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
353 354
 */

355
void cpuset_update_task_memory_state(void)
Linus Torvalds's avatar
Linus Torvalds committed
356
{
357
	int my_cpusets_mem_gen;
358
	struct task_struct *tsk = current;
359
	struct cpuset *cs;
360

361
	if (task_cs(tsk) == &top_cpuset) {
362 363 364 365
		/* Don't need rcu for top_cpuset.  It's never freed. */
		my_cpusets_mem_gen = top_cpuset.mems_generation;
	} else {
		rcu_read_lock();
366
		my_cpusets_mem_gen = task_cs(current)->mems_generation;
367 368
		rcu_read_unlock();
	}
Linus Torvalds's avatar
Linus Torvalds committed
369

370
	if (my_cpusets_mem_gen != tsk->cpuset_mems_generation) {
371
		mutex_lock(&callback_mutex);
372
		task_lock(tsk);
373
		cs = task_cs(tsk); /* Maybe changed when task not locked */
374 375
		guarantee_online_mems(cs, &tsk->mems_allowed);
		tsk->cpuset_mems_generation = cs->mems_generation;
376 377 378 379 380 381 382 383
		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;
384
		task_unlock(tsk);
385
		mutex_unlock(&callback_mutex);
386
		mpol_rebind_task(tsk, &tsk->mems_allowed);
Linus Torvalds's avatar
Linus Torvalds committed
387 388 389 390 391 392 393 394
	}
}

/*
 * 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
395
 * are only set if the other's are set.  Call holding cgroup_mutex.
Linus Torvalds's avatar
Linus Torvalds committed
396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412
 */

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
413
 * cgroup_mutex held.
Linus Torvalds's avatar
Linus Torvalds committed
414 415 416 417 418 419 420 421 422 423 424 425 426 427
 *
 * '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)
{
428
	struct cgroup *cont;
Linus Torvalds's avatar
Linus Torvalds committed
429 430 431
	struct cpuset *c, *par;

	/* Each of our child cpusets must be a subset of us */
432 433
	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
434 435 436 437
			return -EBUSY;
	}

	/* Remaining checks don't apply to root cpuset */
438
	if (cur == &top_cpuset)
Linus Torvalds's avatar
Linus Torvalds committed
439 440
		return 0;

441 442
	par = cur->parent;

Linus Torvalds's avatar
Linus Torvalds committed
443 444 445 446
	/* We must be a subset of our parent cpuset */
	if (!is_cpuset_subset(trial, par))
		return -EACCES;

447 448 449 450
	/*
	 * If either I or some sibling (!= me) is exclusive, we can't
	 * overlap
	 */
451 452
	list_for_each_entry(cont, &par->css.cgroup->children, sibling) {
		c = cgroup_cs(cont);
Linus Torvalds's avatar
Linus Torvalds committed
453 454 455 456 457 458 459 460 461 462
		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;
	}

463 464 465 466 467 468 469 470
	/* Cpusets with tasks can't have empty cpus_allowed or mems_allowed */
	if (cgroup_task_count(cur->css.cgroup)) {
		if (cpus_empty(trial->cpus_allowed) ||
		    nodes_empty(trial->mems_allowed)) {
			return -ENOSPC;
		}
	}

Linus Torvalds's avatar
Linus Torvalds committed
471 472 473
	return 0;
}

Paul Jackson's avatar
Paul Jackson committed
474 475 476 477 478 479 480 481 482 483
/*
 * 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);
}

484 485 486 487 488 489 490 491 492 493
static void
update_domain_attr(struct sched_domain_attr *dattr, struct cpuset *c)
{
	if (!dattr)
		return;
	if (dattr->relax_domain_level < c->relax_domain_level)
		dattr->relax_domain_level = c->relax_domain_level;
	return;
}

Paul Jackson's avatar
Paul Jackson committed
494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520
/*
 * 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
521
 * a call to the get_online_cpus()/put_online_cpus() pair.
Paul Jackson's avatar
Paul Jackson committed
522
 * Must not be called holding callback_mutex, because we must not
523 524
 * call get_online_cpus() while holding callback_mutex.  Elsewhere
 * the kernel nests callback_mutex inside get_online_cpus() calls.
Paul Jackson's avatar
Paul Jackson committed
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
 * 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 */
569
	struct sched_domain_attr *dattr;  /* attributes for custom domains */
Paul Jackson's avatar
Paul Jackson committed
570 571 572 573 574 575
	int ndoms;		/* number of sched domains in result */
	int nslot;		/* next empty doms[] cpumask_t slot */

	q = NULL;
	csa = NULL;
	doms = NULL;
576
	dattr = NULL;
Paul Jackson's avatar
Paul Jackson committed
577 578 579 580 581 582 583

	/* 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;
584 585 586 587 588
		dattr = kmalloc(sizeof(struct sched_domain_attr), GFP_KERNEL);
		if (dattr) {
			*dattr = SD_ATTR_INIT;
			update_domain_attr(dattr, &top_cpuset);
		}
Paul Jackson's avatar
Paul Jackson committed
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
		*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;
645
	dattr = kmalloc(ndoms * sizeof(struct sched_domain_attr), GFP_KERNEL);
Paul Jackson's avatar
Paul Jackson committed
646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667

	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);
668 669
			if (dattr)
				*(dattr + nslot) = SD_ATTR_INIT;
Paul Jackson's avatar
Paul Jackson committed
670 671 672 673 674 675
			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;
676
					update_domain_attr(dattr, b);
Paul Jackson's avatar
Paul Jackson committed
677 678 679 680 681 682 683 684 685
				}
			}
			nslot++;
		}
	}
	BUG_ON(nslot != ndoms);

rebuild:
	/* Have scheduler rebuild sched domains */
686
	get_online_cpus();
687
	partition_sched_domains(ndoms, doms, dattr);
688
	put_online_cpus();
Paul Jackson's avatar
Paul Jackson committed
689 690 691 692 693 694

done:
	if (q && !IS_ERR(q))
		kfifo_free(q);
	kfree(csa);
	/* Don't kfree(doms) -- partition_sched_domains() does that. */
695
	/* Don't kfree(dattr) -- partition_sched_domains() does that. */
Paul Jackson's avatar
Paul Jackson committed
696 697
}

Paul Menage's avatar
Paul Menage committed
698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727
static inline int started_after_time(struct task_struct *t1,
				     struct timespec *time,
				     struct task_struct *t2)
{
	int start_diff = timespec_compare(&t1->start_time, time);
	if (start_diff > 0) {
		return 1;
	} else if (start_diff < 0) {
		return 0;
	} else {
		/*
		 * Arbitrarily, if two processes started at the same
		 * time, we'll say that the lower pointer value
		 * started first. Note that t2 may have exited by now
		 * so this may not be a valid pointer any longer, but
		 * that's fine - it still serves to distinguish
		 * between two tasks started (effectively)
		 * simultaneously.
		 */
		return t1 > t2;
	}
}

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

728 729 730 731 732
/**
 * cpuset_test_cpumask - test a task's cpus_allowed versus its cpuset's
 * @tsk: task to test
 * @scan: struct cgroup_scanner contained in its struct cpuset_hotplug_scanner
 *
733
 * Call with cgroup_mutex held.  May take callback_mutex during call.
734 735 736
 * Called for each task in a cgroup by cgroup_scan_tasks().
 * Return nonzero if this tasks's cpus_allowed mask should be changed (in other
 * words, if its mask is not equal to its cpuset's mask).
737
 */
738 739 740 741 742
int cpuset_test_cpumask(struct task_struct *tsk, struct cgroup_scanner *scan)
{
	return !cpus_equal(tsk->cpus_allowed,
			(cgroup_cs(scan->cg))->cpus_allowed);
}
743

744 745 746 747 748 749 750 751 752 753 754 755 756
/**
 * cpuset_change_cpumask - make a task's cpus_allowed the same as its cpuset's
 * @tsk: task to test
 * @scan: struct cgroup_scanner containing the cgroup of the task
 *
 * Called by cgroup_scan_tasks() for each task in a cgroup whose
 * cpus_allowed mask needs to be changed.
 *
 * We don't need to re-check for the cgroup/cpuset membership, since we're
 * holding cgroup_lock() at this point.
 */
void cpuset_change_cpumask(struct task_struct *tsk, struct cgroup_scanner *scan)
{
757
	set_cpus_allowed_ptr(tsk, &((cgroup_cs(scan->cg))->cpus_allowed));
758 759 760 761 762 763 764
}

/**
 * update_cpumask - update the cpus_allowed mask of a cpuset and all tasks in it
 * @cs: the cpuset to consider
 * @buf: buffer of cpu numbers written to this cpuset
 */
Linus Torvalds's avatar
Linus Torvalds committed
765 766 767
static int update_cpumask(struct cpuset *cs, char *buf)
{
	struct cpuset trialcs;
768
	struct cgroup_scanner scan;
Paul Menage's avatar
Paul Menage committed
769
	struct ptr_heap heap;
770 771
	int retval;
	int is_load_balanced;
Linus Torvalds's avatar
Linus Torvalds committed
772

773 774 775 776
	/* 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
777
	trialcs = *cs;
778 779

	/*
780
	 * An empty cpus_allowed is ok only if the cpuset has no tasks.
781 782 783
	 * Since cpulist_parse() fails on an empty mask, we special case
	 * that parsing.  The validate_change() call ensures that cpusets
	 * with tasks have cpus.
784
	 */
785 786
	buf = strstrip(buf);
	if (!*buf) {
787 788 789 790 791 792
		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
793 794
	cpus_and(trialcs.cpus_allowed, trialcs.cpus_allowed, cpu_online_map);
	retval = validate_change(cs, &trialcs);
795 796
	if (retval < 0)
		return retval;
Paul Jackson's avatar
Paul Jackson committed
797

Paul Menage's avatar
Paul Menage committed
798 799 800
	/* Nothing to do if the cpus didn't change */
	if (cpus_equal(cs->cpus_allowed, trialcs.cpus_allowed))
		return 0;
801

Paul Menage's avatar
Paul Menage committed
802 803 804 805
	retval = heap_init(&heap, PAGE_SIZE, GFP_KERNEL, &started_after);
	if (retval)
		return retval;

Paul Jackson's avatar
Paul Jackson committed
806 807
	is_load_balanced = is_sched_load_balance(&trialcs);

808
	mutex_lock(&callback_mutex);
809
	cs->cpus_allowed = trialcs.cpus_allowed;
810
	mutex_unlock(&callback_mutex);
Paul Jackson's avatar
Paul Jackson committed
811

Paul Menage's avatar
Paul Menage committed
812 813
	/*
	 * Scan tasks in the cpuset, and update the cpumasks of any
814
	 * that need an update.
Paul Menage's avatar
Paul Menage committed
815
	 */
816 817 818 819 820
	scan.cg = cs->css.cgroup;
	scan.test_task = cpuset_test_cpumask;
	scan.process_task = cpuset_change_cpumask;
	scan.heap = &heap;
	cgroup_scan_tasks(&scan);
Paul Menage's avatar
Paul Menage committed
821
	heap_free(&heap);
822

Paul Menage's avatar
Paul Menage committed
823
	if (is_load_balanced)
Paul Jackson's avatar
Paul Jackson committed
824
		rebuild_sched_domains();
825
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
826 827
}

828 829 830 831 832 833 834 835
/*
 * 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.
 *
836
 *    Call holding cgroup_mutex, so current's cpuset won't change
837
 *    during this call, as manage_mutex holds off any cpuset_attach()
838 839
 *    calls.  Therefore we don't need to take task_lock around the
 *    call to guarantee_online_mems(), as we know no one is changing
840
 *    our task's cpuset.
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
 *
 *    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);
873
	guarantee_online_mems(task_cs(tsk),&tsk->mems_allowed);
874 875 876
	mutex_unlock(&callback_mutex);
}

877
/*
878 879 880
 * 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
881 882 883
 * task in the cpuset, rebind any vma mempolicies and if
 * the cpuset is marked 'memory_migrate', migrate the tasks
 * pages to the new memory.
884
 *
885
 * Call with cgroup_mutex held.  May take callback_mutex during call.
886 887 888
 * 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.
889 890
 */

891 892
static void *cpuset_being_rebound;

Linus Torvalds's avatar
Linus Torvalds committed
893 894 895
static int update_nodemask(struct cpuset *cs, char *buf)
{
	struct cpuset trialcs;
896
	nodemask_t oldmem;
897
	struct task_struct *p;
898 899
	struct mm_struct **mmarray;
	int i, n, ntasks;
900
	int migrate;
901
	int fudge;
Linus Torvalds's avatar
Linus Torvalds committed
902
	int retval;
903
	struct cgroup_iter it;
Linus Torvalds's avatar
Linus Torvalds committed
904

905 906 907 908
	/*
	 * top_cpuset.mems_allowed tracks node_stats[N_HIGH_MEMORY];
	 * it's read-only
	 */
909 910 911
	if (cs == &top_cpuset)
		return -EACCES;

Linus Torvalds's avatar
Linus Torvalds committed
912
	trialcs = *cs;
913 914

	/*
915 916 917 918
	 * 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.
919
	 */
920 921
	buf = strstrip(buf);
	if (!*buf) {
922 923 924 925 926 927
		nodes_clear(trialcs.mems_allowed);
	} else {
		retval = nodelist_parse(buf, trialcs.mems_allowed);
		if (retval < 0)
			goto done;
	}
928 929
	nodes_and(trialcs.mems_allowed, trialcs.mems_allowed,
						node_states[N_HIGH_MEMORY]);
930 931 932 933 934
	oldmem = cs->mems_allowed;
	if (nodes_equal(oldmem, trialcs.mems_allowed)) {
		retval = 0;		/* Too easy - nothing to do */
		goto done;
	}
935 936 937 938
	retval = validate_change(cs, &trialcs);
	if (retval < 0)
		goto done;

939
	mutex_lock(&callback_mutex);
940
	cs->mems_allowed = trialcs.mems_allowed;
941
	cs->mems_generation = cpuset_mems_generation++;
942
	mutex_unlock(&callback_mutex);
943

944
	cpuset_being_rebound = cs;		/* causes mpol_copy() rebind */
945 946 947 948 949 950 951 952 953 954 955 956 957

	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) {
958
		ntasks = cgroup_task_count(cs->css.cgroup);  /* guess */
959 960 961 962
		ntasks += fudge;
		mmarray = kmalloc(ntasks * sizeof(*mmarray), GFP_KERNEL);
		if (!mmarray)
			goto done;
963
		read_lock(&tasklist_lock);		/* block fork */
964
		if (cgroup_task_count(cs->css.cgroup) <= ntasks)
965
			break;				/* got enough */
966
		read_unlock(&tasklist_lock);		/* try again */
967 968 969 970 971 972
		kfree(mmarray);
	}

	n = 0;

	/* Load up mmarray[] with mm reference for each task in cpuset. */
973 974
	cgroup_iter_start(cs->css.cgroup, &it);
	while ((p = cgroup_iter_next(cs->css.cgroup, &it))) {
975 976 977 978 979
		struct mm_struct *mm;

		if (n >= ntasks) {
			printk(KERN_WARNING
				"Cpuset mempolicy rebind incomplete.\n");
980
			break;
981 982 983 984 985
		}
		mm = get_task_mm(p);
		if (!mm)
			continue;
		mmarray[n++] = mm;
986 987
	}
	cgroup_iter_end(cs->css.cgroup, &it);
988
	read_unlock(&tasklist_lock);
989 990 991 992 993 994 995 996 997

	/*
	 * 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
998
	 * cgroup_mutex, we know that no other rebind effort will
999 1000
	 * be contending for the global variable cpuset_being_rebound.
	 * It's ok if we rebind the same mm twice; mpol_rebind_mm()
1001
	 * is idempotent.  Also migrate pages in each mm to new nodes.
1002
	 */
1003
	migrate = is_memory_migrate(cs);
1004 1005 1006 1007
	for (i = 0; i < n; i++) {
		struct mm_struct *mm = mmarray[i];

		mpol_rebind_mm(mm, &cs->mems_allowed);
1008 1009
		if (migrate)
			cpuset_migrate_mm(mm, &oldmem, &cs->mems_allowed);
1010 1011 1012
		mmput(mm);
	}

1013
	/* We're done rebinding vmas to this cpuset's new mems_allowed. */
1014
	kfree(mmarray);
1015
	cpuset_being_rebound = NULL;
1016
	retval = 0;
1017
done:
Linus Torvalds's avatar
Linus Torvalds committed
1018 1019 1020
	return retval;
}

1021 1022 1023 1024 1025
int current_cpuset_is_being_rebound(void)
{
	return task_cs(current) == cpuset_being_rebound;
}

1026
/*
1027
 * Call with cgroup_mutex held.
1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038
 */

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;
}

1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053
static int update_relax_domain_level(struct cpuset *cs, char *buf)
{
	int val = simple_strtol(buf, NULL, 10);

	if (val < 0)
		val = -1;

	if (val != cs->relax_domain_level) {
		cs->relax_domain_level = val;
		rebuild_sched_domains();
	}

	return 0;
}

Linus Torvalds's avatar
Linus Torvalds committed
1054 1055 1056
/*
 * update_flag - read a 0 or a 1 in a file and update associated flag
 * bit:	the bit to update (CS_CPU_EXCLUSIVE, CS_MEM_EXCLUSIVE,
Paul Jackson's avatar
Paul Jackson committed
1057
 *				CS_SCHED_LOAD_BALANCE,
1058 1059
 *				CS_NOTIFY_ON_RELEASE, CS_MEMORY_MIGRATE,
 *				CS_SPREAD_PAGE, CS_SPREAD_SLAB)
Linus Torvalds's avatar
Linus Torvalds committed
1060 1061
 * cs:	the cpuset to update
 * buf:	the buffer where we read the 0 or 1
1062
 *
1063
 * Call with cgroup_mutex held.
Linus Torvalds's avatar
Linus Torvalds committed
1064 1065 1066 1067 1068 1069
 */

static int update_flag(cpuset_flagbits_t bit, struct cpuset *cs, char *buf)
{
	int turning_on;
	struct cpuset trialcs;
1070
	int err;
Paul Jackson's avatar
Paul Jackson committed
1071
	int cpus_nonempty, balance_flag_changed;
Linus Torvalds's avatar
Linus Torvalds committed
1072 1073 1074 1075 1076 1077 1078 1079 1080 1081

	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);
1082 1083
	if (err < 0)
		return err;
Paul Jackson's avatar
Paul Jackson committed
1084 1085 1086 1087 1088

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

1089
	mutex_lock(&callback_mutex);
1090
	cs->flags = trialcs.flags;
1091
	mutex_unlock(&callback_mutex);
1092

Paul Jackson's avatar
Paul Jackson committed
1093 1094 1095
	if (cpus_nonempty && balance_flag_changed)
		rebuild_sched_domains();

1096
	return 0;
Linus Torvalds's avatar
Linus Torvalds committed
1097 1098
}

1099
/*
1100
 * Frequency meter - How fast is some event occurring?
1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174 1175 1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196
 *
 * 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;
}

1197
/* Called by cgroups to determine if a cpuset is usable; cgroup_mutex held */
1198 1199
static int cpuset_can_attach(struct cgroup_subsys *ss,
			     struct cgroup *cont, struct task_struct *tsk)
Linus Torvalds's avatar
Linus Torvalds committed
1200
{
1201
	struct cpuset *cs = cgroup_cs(cont);
Linus Torvalds's avatar
Linus Torvalds committed
1202 1203 1204 1205

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

1206 1207
	return security_task_setscheduler(tsk, 0, NULL);
}
Linus Torvalds's avatar
Linus Torvalds committed
1208

1209 1210 1211 1212 1213 1214 1215 1216 1217
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);
1218

1219
	mutex_lock(&callback_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
1220
	guarantee_online_cpus(cs, &cpus);
1221
	set_cpus_allowed_ptr(tsk, &cpus);
1222
	mutex_unlock(&callback_mutex);
Linus Torvalds's avatar
Linus Torvalds committed
1223

1224 1225
	from = oldcs->mems_allowed;
	to = cs->mems_allowed;
1226 1227 1228
	mm = get_task_mm(tsk);
	if (mm) {
		mpol_rebind_mm(mm, &to);
1229
		if (is_memory_migrate(cs))
1230
			cpuset_migrate_mm(mm, &from, &to);
1231 1232 1233
		mmput(mm);
	}

Linus Torvalds's avatar
Linus Torvalds committed
1234 1235 1236 1237 1238
}

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

typedef enum {
1239
	FILE_MEMORY_MIGRATE,
Linus Torvalds's avatar
Linus Torvalds committed
1240 1241 1242 1243
	FILE_CPULIST,
	FILE_MEMLIST,
	FILE_CPU_EXCLUSIVE,
	FILE_MEM_EXCLUSIVE,
Paul Jackson's avatar
Paul Jackson committed
1244
	FILE_SCHED_LOAD_BALANCE,
1245
	FILE_SCHED_RELAX_DOMAIN_LEVEL,
1246 1247
	FILE_MEMORY_PRESSURE_ENABLED,
	FILE_MEMORY_PRESSURE,
1248 1249
	FILE_SPREAD_PAGE,
	FILE_SPREAD_SLAB,
Linus Torvalds's avatar
Linus Torvalds committed
1250 1251
} cpuset_filetype_t;

1252 1253 1254
static ssize_t cpuset_common_file_write(struct cgroup *cont,
					struct cftype *cft,
					struct file *file,
1255
					const char __user *userbuf,
Linus Torvalds's avatar
Linus Torvalds committed
1256 1257
					size_t nbytes, loff_t *unused_ppos)
{
1258
	struct cpuset *cs = cgroup_cs(cont);
Linus Torvalds's avatar
Linus Torvalds committed
1259 1260 1261 1262 1263
	cpuset_filetype_t type = cft->private;
	char *buffer;
	int retval = 0;

	/* Crude upper limit on largest legitimate cpulist user might write. */
Paul Jackson's avatar
Paul Jackson committed
1264
	if (nbytes > 100U + 6 * max(NR_CPUS, MAX_NUMNODES))
Linus Torvalds's avatar
Linus Torvalds committed
1265 1266 1267 1268 1269 1270 1271 1272 1273 1274 1275 1276
		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 */

1277
	cgroup_lock();
Linus Torvalds's avatar
Linus Torvalds committed
1278

1279
	if (cgroup_is_removed(cont)) {
Linus Torvalds's avatar
Linus Torvalds committed
1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296
		retval = -ENODEV;
		goto out2;
	}

	switch (type) {
	case FILE_CPULIST:
		retval = update_cpumask(cs, buffer);
		break;
	case FILE_MEMLIST:
		retval = update_nodemask(cs, buffer);
		break;
	case FILE_CPU_EXCLUSIVE:
		retval = update_flag(CS_CPU_EXCLUSIVE, cs, buffer);
		break;
	case FILE_MEM_EXCLUSIVE:
		retval = update_flag(CS_MEM_EXCLUSIVE, cs, buffer);
		break;
Paul Jackson's avatar
Paul Jackson committed
1297 1298 1299
	case FILE_SCHED_LOAD_BALANCE:
		retval = update_flag(CS_SCHED_LOAD_BALANCE, cs, buffer);
		break;
1300 1301 1302
	case FILE_SCHED_RELAX_DOMAIN_LEVEL:
		retval = update_relax_domain_level(cs, buffer);
		break;
1303 1304 1305
	case FILE_MEMORY_MIGRATE:
		retval = update_flag(CS_MEMORY_MIGRATE, cs, buffer);
		break;
1306 1307 1308 1309 1310 1311
	case FILE_MEMORY_PRESSURE_ENABLED:
		retval = update_memory_pressure_enabled(cs, buffer);
		break;
	case FILE_MEMORY_PRESSURE:
		retval = -EACCES;
		break;