mempolicy.c 65.0 KB
Newer Older
L
Linus Torvalds 已提交
1 2 3 4
/*
 * Simple NUMA memory policy for the Linux kernel.
 *
 * Copyright 2003,2004 Andi Kleen, SuSE Labs.
5
 * (C) Copyright 2005 Christoph Lameter, Silicon Graphics, Inc.
L
Linus Torvalds 已提交
6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
 * Subject to the GNU Public License, version 2.
 *
 * NUMA policy allows the user to give hints in which node(s) memory should
 * be allocated.
 *
 * Support four policies per VMA and per process:
 *
 * The VMA policy has priority over the process policy for a page fault.
 *
 * interleave     Allocate memory interleaved over a set of nodes,
 *                with normal fallback if it fails.
 *                For VMA based allocations this interleaves based on the
 *                offset into the backing object or offset into the mapping
 *                for anonymous memory. For process policy an process counter
 *                is used.
21
 *
L
Linus Torvalds 已提交
22 23
 * bind           Only allocate memory on a specific set of nodes,
 *                no fallback.
24 25 26 27
 *                FIXME: memory is allocated starting with the first node
 *                to the last. It would be better if bind would truly restrict
 *                the allocation to memory nodes instead
 *
L
Linus Torvalds 已提交
28 29 30 31 32
 * preferred       Try a specific node first before normal fallback.
 *                As a special case node -1 here means do the allocation
 *                on the local CPU. This is normally identical to default,
 *                but useful to set in a VMA when you have a non default
 *                process policy.
33
 *
L
Linus Torvalds 已提交
34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77
 * default        Allocate on the local node first, or when on a VMA
 *                use the process policy. This is what Linux always did
 *		  in a NUMA aware kernel and still does by, ahem, default.
 *
 * The process policy is applied for most non interrupt memory allocations
 * in that process' context. Interrupts ignore the policies and always
 * try to allocate on the local CPU. The VMA policy is only applied for memory
 * allocations for a VMA in the VM.
 *
 * Currently there are a few corner cases in swapping where the policy
 * is not applied, but the majority should be handled. When process policy
 * is used it is not remembered over swap outs/swap ins.
 *
 * Only the highest zone in the zone hierarchy gets policied. Allocations
 * requesting a lower zone just use default policy. This implies that
 * on systems with highmem kernel lowmem allocation don't get policied.
 * Same with GFP_DMA allocations.
 *
 * For shmfs/tmpfs/hugetlbfs shared memory the policy is shared between
 * all users and remembered even when nobody has memory mapped.
 */

/* Notebook:
   fix mmap readahead to honour policy and enable policy for any page cache
   object
   statistics for bigpages
   global policy for page cache? currently it uses process policy. Requires
   first item above.
   handle mremap for shared memory (currently ignored for the policy)
   grows down?
   make bind policy root only? It can trigger oom much faster and the
   kernel is not always grateful with that.
*/

#include <linux/mempolicy.h>
#include <linux/mm.h>
#include <linux/highmem.h>
#include <linux/hugetlb.h>
#include <linux/kernel.h>
#include <linux/sched.h>
#include <linux/nodemask.h>
#include <linux/cpuset.h>
#include <linux/slab.h>
#include <linux/string.h>
78
#include <linux/export.h>
79
#include <linux/nsproxy.h>
L
Linus Torvalds 已提交
80 81 82
#include <linux/interrupt.h>
#include <linux/init.h>
#include <linux/compat.h>
83
#include <linux/swap.h>
84 85
#include <linux/seq_file.h>
#include <linux/proc_fs.h>
C
Christoph Lameter 已提交
86
#include <linux/migrate.h>
87
#include <linux/ksm.h>
88
#include <linux/rmap.h>
89
#include <linux/security.h>
A
Adrian Bunk 已提交
90
#include <linux/syscalls.h>
91
#include <linux/ctype.h>
92
#include <linux/mm_inline.h>
93

L
Linus Torvalds 已提交
94 95
#include <asm/tlbflush.h>
#include <asm/uaccess.h>
96
#include <linux/random.h>
L
Linus Torvalds 已提交
97

98 99
#include "internal.h"

100
/* Internal flags */
101
#define MPOL_MF_DISCONTIG_OK (MPOL_MF_INTERNAL << 0)	/* Skip checks for continuous vmas */
102
#define MPOL_MF_INVERT (MPOL_MF_INTERNAL << 1)		/* Invert check for nodemask */
103

104 105
static struct kmem_cache *policy_cache;
static struct kmem_cache *sn_cache;
L
Linus Torvalds 已提交
106 107 108

/* Highest zone. An specific allocation for a zone below that is not
   policied. */
109
enum zone_type policy_zone = 0;
L
Linus Torvalds 已提交
110

111 112 113
/*
 * run-time system-wide default policy => local allocation
 */
114
static struct mempolicy default_policy = {
L
Linus Torvalds 已提交
115
	.refcnt = ATOMIC_INIT(1), /* never free it */
116
	.mode = MPOL_PREFERRED,
117
	.flags = MPOL_F_LOCAL,
L
Linus Torvalds 已提交
118 119
};

120 121
static const struct mempolicy_operations {
	int (*create)(struct mempolicy *pol, const nodemask_t *nodes);
122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137
	/*
	 * If read-side task has no lock to protect task->mempolicy, write-side
	 * task will rebind the task->mempolicy by two step. The first step is
	 * setting all the newly nodes, and the second step is cleaning all the
	 * disallowed nodes. In this way, we can avoid finding no node to alloc
	 * page.
	 * If we have a lock to protect task->mempolicy in read-side, we do
	 * rebind directly.
	 *
	 * step:
	 * 	MPOL_REBIND_ONCE - do rebind work at once
	 * 	MPOL_REBIND_STEP1 - set all the newly nodes
	 * 	MPOL_REBIND_STEP2 - clean all the disallowed nodes
	 */
	void (*rebind)(struct mempolicy *pol, const nodemask_t *nodes,
			enum mpol_rebind_step step);
138 139
} mpol_ops[MPOL_MAX];

140
/* Check that the nodemask contains at least one populated zone */
141
static int is_valid_nodemask(const nodemask_t *nodemask)
L
Linus Torvalds 已提交
142
{
143
	int nd, k;
L
Linus Torvalds 已提交
144

145 146 147 148 149 150 151
	for_each_node_mask(nd, *nodemask) {
		struct zone *z;

		for (k = 0; k <= policy_zone; k++) {
			z = &NODE_DATA(nd)->node_zones[k];
			if (z->present_pages > 0)
				return 1;
152
		}
153
	}
154 155

	return 0;
L
Linus Torvalds 已提交
156 157
}

158 159
static inline int mpol_store_user_nodemask(const struct mempolicy *pol)
{
B
Bob Liu 已提交
160
	return pol->flags & MPOL_MODE_FLAGS;
161 162 163 164 165 166 167 168
}

static void mpol_relative_nodemask(nodemask_t *ret, const nodemask_t *orig,
				   const nodemask_t *rel)
{
	nodemask_t tmp;
	nodes_fold(tmp, *orig, nodes_weight(*rel));
	nodes_onto(*ret, tmp, *rel);
169 170
}

171 172 173 174 175 176 177 178 179 180 181
static int mpol_new_interleave(struct mempolicy *pol, const nodemask_t *nodes)
{
	if (nodes_empty(*nodes))
		return -EINVAL;
	pol->v.nodes = *nodes;
	return 0;
}

static int mpol_new_preferred(struct mempolicy *pol, const nodemask_t *nodes)
{
	if (!nodes)
182
		pol->flags |= MPOL_F_LOCAL;	/* local allocation */
183 184 185 186 187 188 189 190 191 192 193 194 195 196 197
	else if (nodes_empty(*nodes))
		return -EINVAL;			/*  no allowed nodes */
	else
		pol->v.preferred_node = first_node(*nodes);
	return 0;
}

static int mpol_new_bind(struct mempolicy *pol, const nodemask_t *nodes)
{
	if (!is_valid_nodemask(nodes))
		return -EINVAL;
	pol->v.nodes = *nodes;
	return 0;
}

198 199 200 201 202 203 204 205 206
/*
 * mpol_set_nodemask is called after mpol_new() to set up the nodemask, if
 * any, for the new policy.  mpol_new() has already validated the nodes
 * parameter with respect to the policy mode and flags.  But, we need to
 * handle an empty nodemask with MPOL_PREFERRED here.
 *
 * Must be called holding task's alloc_lock to protect task's mems_allowed
 * and mempolicy.  May also be called holding the mmap_semaphore for write.
 */
207 208
static int mpol_set_nodemask(struct mempolicy *pol,
		     const nodemask_t *nodes, struct nodemask_scratch *nsc)
209 210 211 212 213 214
{
	int ret;

	/* if mode is MPOL_DEFAULT, pol is NULL. This is right. */
	if (pol == NULL)
		return 0;
215 216 217
	/* Check N_HIGH_MEMORY */
	nodes_and(nsc->mask1,
		  cpuset_current_mems_allowed, node_states[N_HIGH_MEMORY]);
218 219 220 221 222 223

	VM_BUG_ON(!nodes);
	if (pol->mode == MPOL_PREFERRED && nodes_empty(*nodes))
		nodes = NULL;	/* explicit local allocation */
	else {
		if (pol->flags & MPOL_F_RELATIVE_NODES)
224
			mpol_relative_nodemask(&nsc->mask2, nodes,&nsc->mask1);
225
		else
226 227
			nodes_and(nsc->mask2, *nodes, nsc->mask1);

228 229 230 231 232 233 234
		if (mpol_store_user_nodemask(pol))
			pol->w.user_nodemask = *nodes;
		else
			pol->w.cpuset_mems_allowed =
						cpuset_current_mems_allowed;
	}

235 236 237 238
	if (nodes)
		ret = mpol_ops[pol->mode].create(pol, &nsc->mask2);
	else
		ret = mpol_ops[pol->mode].create(pol, NULL);
239 240 241 242 243 244 245
	return ret;
}

/*
 * This function just creates a new policy, does some check and simple
 * initialization. You must invoke mpol_set_nodemask() to set nodes.
 */
246 247
static struct mempolicy *mpol_new(unsigned short mode, unsigned short flags,
				  nodemask_t *nodes)
L
Linus Torvalds 已提交
248 249 250
{
	struct mempolicy *policy;

251 252
	pr_debug("setting mode %d flags %d nodes[0] %lx\n",
		 mode, flags, nodes ? nodes_addr(*nodes)[0] : -1);
253

254 255
	if (mode == MPOL_DEFAULT) {
		if (nodes && !nodes_empty(*nodes))
256
			return ERR_PTR(-EINVAL);
257
		return NULL;	/* simply delete any existing policy */
258
	}
259 260 261 262 263 264 265 266 267 268 269 270 271 272 273
	VM_BUG_ON(!nodes);

	/*
	 * MPOL_PREFERRED cannot be used with MPOL_F_STATIC_NODES or
	 * MPOL_F_RELATIVE_NODES if the nodemask is empty (local allocation).
	 * All other modes require a valid pointer to a non-empty nodemask.
	 */
	if (mode == MPOL_PREFERRED) {
		if (nodes_empty(*nodes)) {
			if (((flags & MPOL_F_STATIC_NODES) ||
			     (flags & MPOL_F_RELATIVE_NODES)))
				return ERR_PTR(-EINVAL);
		}
	} else if (nodes_empty(*nodes))
		return ERR_PTR(-EINVAL);
L
Linus Torvalds 已提交
274 275 276 277
	policy = kmem_cache_alloc(policy_cache, GFP_KERNEL);
	if (!policy)
		return ERR_PTR(-ENOMEM);
	atomic_set(&policy->refcnt, 1);
278
	policy->mode = mode;
279
	policy->flags = flags;
280

L
Linus Torvalds 已提交
281
	return policy;
282 283
}

284 285 286 287 288 289 290 291
/* Slow path of a mpol destructor. */
void __mpol_put(struct mempolicy *p)
{
	if (!atomic_dec_and_test(&p->refcnt))
		return;
	kmem_cache_free(policy_cache, p);
}

292 293
static void mpol_rebind_default(struct mempolicy *pol, const nodemask_t *nodes,
				enum mpol_rebind_step step)
294 295 296
{
}

297 298 299 300 301 302 303 304
/*
 * step:
 * 	MPOL_REBIND_ONCE  - do rebind work at once
 * 	MPOL_REBIND_STEP1 - set all the newly nodes
 * 	MPOL_REBIND_STEP2 - clean all the disallowed nodes
 */
static void mpol_rebind_nodemask(struct mempolicy *pol, const nodemask_t *nodes,
				 enum mpol_rebind_step step)
305 306 307 308 309 310 311 312
{
	nodemask_t tmp;

	if (pol->flags & MPOL_F_STATIC_NODES)
		nodes_and(tmp, pol->w.user_nodemask, *nodes);
	else if (pol->flags & MPOL_F_RELATIVE_NODES)
		mpol_relative_nodemask(&tmp, &pol->w.user_nodemask, nodes);
	else {
313 314 315 316 317 318 319 320 321 322 323 324 325
		/*
		 * if step == 1, we use ->w.cpuset_mems_allowed to cache the
		 * result
		 */
		if (step == MPOL_REBIND_ONCE || step == MPOL_REBIND_STEP1) {
			nodes_remap(tmp, pol->v.nodes,
					pol->w.cpuset_mems_allowed, *nodes);
			pol->w.cpuset_mems_allowed = step ? tmp : *nodes;
		} else if (step == MPOL_REBIND_STEP2) {
			tmp = pol->w.cpuset_mems_allowed;
			pol->w.cpuset_mems_allowed = *nodes;
		} else
			BUG();
326
	}
327

328 329 330 331 332 333 334 335 336 337
	if (nodes_empty(tmp))
		tmp = *nodes;

	if (step == MPOL_REBIND_STEP1)
		nodes_or(pol->v.nodes, pol->v.nodes, tmp);
	else if (step == MPOL_REBIND_ONCE || step == MPOL_REBIND_STEP2)
		pol->v.nodes = tmp;
	else
		BUG();

338 339 340 341 342 343 344 345 346 347
	if (!node_isset(current->il_next, tmp)) {
		current->il_next = next_node(current->il_next, tmp);
		if (current->il_next >= MAX_NUMNODES)
			current->il_next = first_node(tmp);
		if (current->il_next >= MAX_NUMNODES)
			current->il_next = numa_node_id();
	}
}

static void mpol_rebind_preferred(struct mempolicy *pol,
348 349
				  const nodemask_t *nodes,
				  enum mpol_rebind_step step)
350 351 352 353 354 355
{
	nodemask_t tmp;

	if (pol->flags & MPOL_F_STATIC_NODES) {
		int node = first_node(pol->w.user_nodemask);

356
		if (node_isset(node, *nodes)) {
357
			pol->v.preferred_node = node;
358 359 360
			pol->flags &= ~MPOL_F_LOCAL;
		} else
			pol->flags |= MPOL_F_LOCAL;
361 362 363
	} else if (pol->flags & MPOL_F_RELATIVE_NODES) {
		mpol_relative_nodemask(&tmp, &pol->w.user_nodemask, nodes);
		pol->v.preferred_node = first_node(tmp);
364
	} else if (!(pol->flags & MPOL_F_LOCAL)) {
365 366 367 368 369
		pol->v.preferred_node = node_remap(pol->v.preferred_node,
						   pol->w.cpuset_mems_allowed,
						   *nodes);
		pol->w.cpuset_mems_allowed = *nodes;
	}
L
Linus Torvalds 已提交
370 371
}

372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389
/*
 * mpol_rebind_policy - Migrate a policy to a different set of nodes
 *
 * If read-side task has no lock to protect task->mempolicy, write-side
 * task will rebind the task->mempolicy by two step. The first step is
 * setting all the newly nodes, and the second step is cleaning all the
 * disallowed nodes. In this way, we can avoid finding no node to alloc
 * page.
 * If we have a lock to protect task->mempolicy in read-side, we do
 * rebind directly.
 *
 * step:
 * 	MPOL_REBIND_ONCE  - do rebind work at once
 * 	MPOL_REBIND_STEP1 - set all the newly nodes
 * 	MPOL_REBIND_STEP2 - clean all the disallowed nodes
 */
static void mpol_rebind_policy(struct mempolicy *pol, const nodemask_t *newmask,
				enum mpol_rebind_step step)
390 391 392
{
	if (!pol)
		return;
393
	if (!mpol_store_user_nodemask(pol) && step == 0 &&
394 395
	    nodes_equal(pol->w.cpuset_mems_allowed, *newmask))
		return;
396 397 398 399 400 401 402 403 404 405 406 407 408 409 410

	if (step == MPOL_REBIND_STEP1 && (pol->flags & MPOL_F_REBINDING))
		return;

	if (step == MPOL_REBIND_STEP2 && !(pol->flags & MPOL_F_REBINDING))
		BUG();

	if (step == MPOL_REBIND_STEP1)
		pol->flags |= MPOL_F_REBINDING;
	else if (step == MPOL_REBIND_STEP2)
		pol->flags &= ~MPOL_F_REBINDING;
	else if (step >= MPOL_REBIND_NSTEP)
		BUG();

	mpol_ops[pol->mode].rebind(pol, newmask, step);
411 412 413 414 415
}

/*
 * Wrapper for mpol_rebind_policy() that just requires task
 * pointer, and updates task mempolicy.
416 417
 *
 * Called with task's alloc_lock held.
418 419
 */

420 421
void mpol_rebind_task(struct task_struct *tsk, const nodemask_t *new,
			enum mpol_rebind_step step)
422
{
423
	mpol_rebind_policy(tsk->mempolicy, new, step);
424 425 426 427 428 429 430 431 432 433 434 435 436 437
}

/*
 * Rebind each vma in mm to new nodemask.
 *
 * Call holding a reference to mm.  Takes mm->mmap_sem during call.
 */

void mpol_rebind_mm(struct mm_struct *mm, nodemask_t *new)
{
	struct vm_area_struct *vma;

	down_write(&mm->mmap_sem);
	for (vma = mm->mmap; vma; vma = vma->vm_next)
438
		mpol_rebind_policy(vma->vm_policy, new, MPOL_REBIND_ONCE);
439 440 441
	up_write(&mm->mmap_sem);
}

442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459
static const struct mempolicy_operations mpol_ops[MPOL_MAX] = {
	[MPOL_DEFAULT] = {
		.rebind = mpol_rebind_default,
	},
	[MPOL_INTERLEAVE] = {
		.create = mpol_new_interleave,
		.rebind = mpol_rebind_nodemask,
	},
	[MPOL_PREFERRED] = {
		.create = mpol_new_preferred,
		.rebind = mpol_rebind_preferred,
	},
	[MPOL_BIND] = {
		.create = mpol_new_bind,
		.rebind = mpol_rebind_nodemask,
	},
};

460 461
static void migrate_page_add(struct page *page, struct list_head *pagelist,
				unsigned long flags);
462

463
/* Scan through pages checking if pages follow certain conditions. */
N
Nick Piggin 已提交
464
static int check_pte_range(struct vm_area_struct *vma, pmd_t *pmd,
465 466
		unsigned long addr, unsigned long end,
		const nodemask_t *nodes, unsigned long flags,
467
		void *private)
L
Linus Torvalds 已提交
468
{
469 470
	pte_t *orig_pte;
	pte_t *pte;
471
	spinlock_t *ptl;
472

473
	orig_pte = pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
474
	do {
475
		struct page *page;
476
		int nid;
477 478

		if (!pte_present(*pte))
L
Linus Torvalds 已提交
479
			continue;
480 481
		page = vm_normal_page(vma, addr, *pte);
		if (!page)
L
Linus Torvalds 已提交
482
			continue;
483
		/*
484 485 486
		 * vm_normal_page() filters out zero pages, but there might
		 * still be PageReserved pages to skip, perhaps in a VDSO.
		 * And we cannot move PageKsm pages sensibly or safely yet.
487
		 */
488
		if (PageReserved(page) || PageKsm(page))
489
			continue;
490
		nid = page_to_nid(page);
491 492 493
		if (node_isset(nid, *nodes) == !!(flags & MPOL_MF_INVERT))
			continue;

S
Stephen Wilson 已提交
494
		if (flags & (MPOL_MF_MOVE | MPOL_MF_MOVE_ALL))
495
			migrate_page_add(page, private, flags);
496 497
		else
			break;
498
	} while (pte++, addr += PAGE_SIZE, addr != end);
499
	pte_unmap_unlock(orig_pte, ptl);
500 501 502
	return addr != end;
}

N
Nick Piggin 已提交
503
static inline int check_pmd_range(struct vm_area_struct *vma, pud_t *pud,
504 505
		unsigned long addr, unsigned long end,
		const nodemask_t *nodes, unsigned long flags,
506
		void *private)
507 508 509 510 511 512 513
{
	pmd_t *pmd;
	unsigned long next;

	pmd = pmd_offset(pud, addr);
	do {
		next = pmd_addr_end(addr, end);
A
Andrea Arcangeli 已提交
514
		split_huge_page_pmd(vma->vm_mm, pmd);
515
		if (pmd_none_or_trans_huge_or_clear_bad(pmd))
516
			continue;
517
		if (check_pte_range(vma, pmd, addr, next, nodes,
518
				    flags, private))
519 520 521 522 523
			return -EIO;
	} while (pmd++, addr = next, addr != end);
	return 0;
}

N
Nick Piggin 已提交
524
static inline int check_pud_range(struct vm_area_struct *vma, pgd_t *pgd,
525 526
		unsigned long addr, unsigned long end,
		const nodemask_t *nodes, unsigned long flags,
527
		void *private)
528 529 530 531 532 533 534 535 536
{
	pud_t *pud;
	unsigned long next;

	pud = pud_offset(pgd, addr);
	do {
		next = pud_addr_end(addr, end);
		if (pud_none_or_clear_bad(pud))
			continue;
537
		if (check_pmd_range(vma, pud, addr, next, nodes,
538
				    flags, private))
539 540 541 542 543
			return -EIO;
	} while (pud++, addr = next, addr != end);
	return 0;
}

N
Nick Piggin 已提交
544
static inline int check_pgd_range(struct vm_area_struct *vma,
545 546
		unsigned long addr, unsigned long end,
		const nodemask_t *nodes, unsigned long flags,
547
		void *private)
548 549 550 551
{
	pgd_t *pgd;
	unsigned long next;

N
Nick Piggin 已提交
552
	pgd = pgd_offset(vma->vm_mm, addr);
553 554 555 556
	do {
		next = pgd_addr_end(addr, end);
		if (pgd_none_or_clear_bad(pgd))
			continue;
557
		if (check_pud_range(vma, pgd, addr, next, nodes,
558
				    flags, private))
559 560 561
			return -EIO;
	} while (pgd++, addr = next, addr != end);
	return 0;
L
Linus Torvalds 已提交
562 563
}

564 565 566 567 568
/*
 * Check if all pages in a range are on a set of nodes.
 * If pagelist != NULL then isolate pages from the LRU and
 * put them on the pagelist.
 */
L
Linus Torvalds 已提交
569 570
static struct vm_area_struct *
check_range(struct mm_struct *mm, unsigned long start, unsigned long end,
571
		const nodemask_t *nodes, unsigned long flags, void *private)
L
Linus Torvalds 已提交
572 573 574 575
{
	int err;
	struct vm_area_struct *first, *vma, *prev;

576

L
Linus Torvalds 已提交
577 578 579 580 581
	first = find_vma(mm, start);
	if (!first)
		return ERR_PTR(-EFAULT);
	prev = NULL;
	for (vma = first; vma && vma->vm_start < end; vma = vma->vm_next) {
582 583 584 585 586 587 588 589 590 591
		if (!(flags & MPOL_MF_DISCONTIG_OK)) {
			if (!vma->vm_next && vma->vm_end < end)
				return ERR_PTR(-EFAULT);
			if (prev && prev->vm_end < vma->vm_start)
				return ERR_PTR(-EFAULT);
		}
		if (!is_vm_hugetlb_page(vma) &&
		    ((flags & MPOL_MF_STRICT) ||
		     ((flags & (MPOL_MF_MOVE | MPOL_MF_MOVE_ALL)) &&
				vma_migratable(vma)))) {
A
Andi Kleen 已提交
592
			unsigned long endvma = vma->vm_end;
593

A
Andi Kleen 已提交
594 595 596 597
			if (endvma > end)
				endvma = end;
			if (vma->vm_start > start)
				start = vma->vm_start;
598
			err = check_pgd_range(vma, start, endvma, nodes,
599
						flags, private);
L
Linus Torvalds 已提交
600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615
			if (err) {
				first = ERR_PTR(err);
				break;
			}
		}
		prev = vma;
	}
	return first;
}

/* Apply policy to a single VMA */
static int policy_vma(struct vm_area_struct *vma, struct mempolicy *new)
{
	int err = 0;
	struct mempolicy *old = vma->vm_policy;

616
	pr_debug("vma %lx-%lx/%lx vm_ops %p vm_file %p set_policy %p\n",
L
Linus Torvalds 已提交
617 618 619 620 621 622 623 624 625
		 vma->vm_start, vma->vm_end, vma->vm_pgoff,
		 vma->vm_ops, vma->vm_file,
		 vma->vm_ops ? vma->vm_ops->set_policy : NULL);

	if (vma->vm_ops && vma->vm_ops->set_policy)
		err = vma->vm_ops->set_policy(vma, new);
	if (!err) {
		mpol_get(new);
		vma->vm_policy = new;
626
		mpol_put(old);
L
Linus Torvalds 已提交
627 628 629 630 631
	}
	return err;
}

/* Step 2: apply policy to a range and do splits. */
632 633
static int mbind_range(struct mm_struct *mm, unsigned long start,
		       unsigned long end, struct mempolicy *new_pol)
L
Linus Torvalds 已提交
634 635
{
	struct vm_area_struct *next;
636 637 638
	struct vm_area_struct *prev;
	struct vm_area_struct *vma;
	int err = 0;
639
	pgoff_t pgoff;
640 641
	unsigned long vmstart;
	unsigned long vmend;
L
Linus Torvalds 已提交
642

643
	vma = find_vma(mm, start);
644 645 646
	if (!vma || vma->vm_start > start)
		return -EFAULT;

647
	prev = vma->vm_prev;
648 649 650
	if (start > vma->vm_start)
		prev = vma;

651
	for (; vma && vma->vm_start < end; prev = vma, vma = next) {
L
Linus Torvalds 已提交
652
		next = vma->vm_next;
653 654 655
		vmstart = max(start, vma->vm_start);
		vmend   = min(end, vma->vm_end);

656 657 658 659 660
		if (mpol_equal(vma_policy(vma), new_pol))
			continue;

		pgoff = vma->vm_pgoff +
			((vmstart - vma->vm_start) >> PAGE_SHIFT);
661
		prev = vma_merge(mm, prev, vmstart, vmend, vma->vm_flags,
662
				  vma->anon_vma, vma->vm_file, pgoff,
663
				  new_pol);
664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679
		if (prev) {
			vma = prev;
			next = vma->vm_next;
			continue;
		}
		if (vma->vm_start != vmstart) {
			err = split_vma(vma->vm_mm, vma, vmstart, 1);
			if (err)
				goto out;
		}
		if (vma->vm_end != vmend) {
			err = split_vma(vma->vm_mm, vma, vmend, 0);
			if (err)
				goto out;
		}
		err = policy_vma(vma, new_pol);
L
Linus Torvalds 已提交
680
		if (err)
681
			goto out;
L
Linus Torvalds 已提交
682
	}
683 684

 out:
L
Linus Torvalds 已提交
685 686 687
	return err;
}

688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717
/*
 * Update task->flags PF_MEMPOLICY bit: set iff non-default
 * mempolicy.  Allows more rapid checking of this (combined perhaps
 * with other PF_* flag bits) on memory allocation hot code paths.
 *
 * If called from outside this file, the task 'p' should -only- be
 * a newly forked child not yet visible on the task list, because
 * manipulating the task flags of a visible task is not safe.
 *
 * The above limitation is why this routine has the funny name
 * mpol_fix_fork_child_flag().
 *
 * It is also safe to call this with a task pointer of current,
 * which the static wrapper mpol_set_task_struct_flag() does,
 * for use within this file.
 */

void mpol_fix_fork_child_flag(struct task_struct *p)
{
	if (p->mempolicy)
		p->flags |= PF_MEMPOLICY;
	else
		p->flags &= ~PF_MEMPOLICY;
}

static void mpol_set_task_struct_flag(void)
{
	mpol_fix_fork_child_flag(current);
}

L
Linus Torvalds 已提交
718
/* Set the process memory policy */
719 720
static long do_set_mempolicy(unsigned short mode, unsigned short flags,
			     nodemask_t *nodes)
L
Linus Torvalds 已提交
721
{
722
	struct mempolicy *new, *old;
723
	struct mm_struct *mm = current->mm;
724
	NODEMASK_SCRATCH(scratch);
725
	int ret;
L
Linus Torvalds 已提交
726

727 728
	if (!scratch)
		return -ENOMEM;
729

730 731 732 733 734
	new = mpol_new(mode, flags, nodes);
	if (IS_ERR(new)) {
		ret = PTR_ERR(new);
		goto out;
	}
735 736 737 738 739 740 741 742
	/*
	 * prevent changing our mempolicy while show_numa_maps()
	 * is using it.
	 * Note:  do_set_mempolicy() can be called at init time
	 * with no 'mm'.
	 */
	if (mm)
		down_write(&mm->mmap_sem);
743
	task_lock(current);
744
	ret = mpol_set_nodemask(new, nodes, scratch);
745 746 747 748 749
	if (ret) {
		task_unlock(current);
		if (mm)
			up_write(&mm->mmap_sem);
		mpol_put(new);
750
		goto out;
751 752
	}
	old = current->mempolicy;
L
Linus Torvalds 已提交
753
	current->mempolicy = new;
754
	mpol_set_task_struct_flag();
755
	if (new && new->mode == MPOL_INTERLEAVE &&
756
	    nodes_weight(new->v.nodes))
757
		current->il_next = first_node(new->v.nodes);
758
	task_unlock(current);
759 760 761
	if (mm)
		up_write(&mm->mmap_sem);

762
	mpol_put(old);
763 764 765 766
	ret = 0;
out:
	NODEMASK_SCRATCH_FREE(scratch);
	return ret;
L
Linus Torvalds 已提交
767 768
}

769 770
/*
 * Return nodemask for policy for get_mempolicy() query
771 772
 *
 * Called with task's alloc_lock held
773 774
 */
static void get_policy_nodemask(struct mempolicy *p, nodemask_t *nodes)
L
Linus Torvalds 已提交
775
{
776
	nodes_clear(*nodes);
777 778 779
	if (p == &default_policy)
		return;

780
	switch (p->mode) {
781 782
	case MPOL_BIND:
		/* Fall through */
L
Linus Torvalds 已提交
783
	case MPOL_INTERLEAVE:
784
		*nodes = p->v.nodes;
L
Linus Torvalds 已提交
785 786
		break;
	case MPOL_PREFERRED:
787
		if (!(p->flags & MPOL_F_LOCAL))
788
			node_set(p->v.preferred_node, *nodes);
789
		/* else return empty node mask for local allocation */
L
Linus Torvalds 已提交
790 791 792 793 794 795 796 797 798 799 800 801 802 803 804 805 806 807 808 809
		break;
	default:
		BUG();
	}
}

static int lookup_node(struct mm_struct *mm, unsigned long addr)
{
	struct page *p;
	int err;

	err = get_user_pages(current, mm, addr & PAGE_MASK, 1, 0, 0, &p, NULL);
	if (err >= 0) {
		err = page_to_nid(p);
		put_page(p);
	}
	return err;
}

/* Retrieve NUMA policy */
A
Adrian Bunk 已提交
810 811
static long do_get_mempolicy(int *policy, nodemask_t *nmask,
			     unsigned long addr, unsigned long flags)
L
Linus Torvalds 已提交
812
{
813
	int err;
L
Linus Torvalds 已提交
814 815 816 817
	struct mm_struct *mm = current->mm;
	struct vm_area_struct *vma = NULL;
	struct mempolicy *pol = current->mempolicy;

818 819
	if (flags &
		~(unsigned long)(MPOL_F_NODE|MPOL_F_ADDR|MPOL_F_MEMS_ALLOWED))
L
Linus Torvalds 已提交
820
		return -EINVAL;
821 822 823 824 825

	if (flags & MPOL_F_MEMS_ALLOWED) {
		if (flags & (MPOL_F_NODE|MPOL_F_ADDR))
			return -EINVAL;
		*policy = 0;	/* just so it's initialized */
826
		task_lock(current);
827
		*nmask  = cpuset_current_mems_allowed;
828
		task_unlock(current);
829 830 831
		return 0;
	}

L
Linus Torvalds 已提交
832
	if (flags & MPOL_F_ADDR) {
833 834 835 836 837
		/*
		 * Do NOT fall back to task policy if the
		 * vma/shared policy at addr is NULL.  We
		 * want to return MPOL_DEFAULT in this case.
		 */
L
Linus Torvalds 已提交
838 839 840 841 842 843 844 845 846 847 848 849 850 851
		down_read(&mm->mmap_sem);
		vma = find_vma_intersection(mm, addr, addr+1);
		if (!vma) {
			up_read(&mm->mmap_sem);
			return -EFAULT;
		}
		if (vma->vm_ops && vma->vm_ops->get_policy)
			pol = vma->vm_ops->get_policy(vma, addr);
		else
			pol = vma->vm_policy;
	} else if (addr)
		return -EINVAL;

	if (!pol)
852
		pol = &default_policy;	/* indicates default behavior */
L
Linus Torvalds 已提交
853 854 855 856 857 858

	if (flags & MPOL_F_NODE) {
		if (flags & MPOL_F_ADDR) {
			err = lookup_node(mm, addr);
			if (err < 0)
				goto out;
859
			*policy = err;
L
Linus Torvalds 已提交
860
		} else if (pol == current->mempolicy &&
861
				pol->mode == MPOL_INTERLEAVE) {
862
			*policy = current->il_next;
L
Linus Torvalds 已提交
863 864 865 866
		} else {
			err = -EINVAL;
			goto out;
		}
867 868 869
	} else {
		*policy = pol == &default_policy ? MPOL_DEFAULT :
						pol->mode;
870 871 872 873 874
		/*
		 * Internal mempolicy flags must be masked off before exposing
		 * the policy to userspace.
		 */
		*policy |= (pol->flags & MPOL_MODE_FLAGS);
875
	}
L
Linus Torvalds 已提交
876 877 878 879 880 881 882

	if (vma) {
		up_read(&current->mm->mmap_sem);
		vma = NULL;
	}

	err = 0;
883
	if (nmask) {
884 885 886 887 888 889 890
		if (mpol_store_user_nodemask(pol)) {
			*nmask = pol->w.user_nodemask;
		} else {
			task_lock(current);
			get_policy_nodemask(pol, nmask);
			task_unlock(current);
		}
891
	}
L
Linus Torvalds 已提交
892 893

 out:
894
	mpol_cond_put(pol);
L
Linus Torvalds 已提交
895 896 897 898 899
	if (vma)
		up_read(&current->mm->mmap_sem);
	return err;
}

C
Christoph Lameter 已提交
900
#ifdef CONFIG_MIGRATION
901 902 903
/*
 * page migration
 */
904 905
static void migrate_page_add(struct page *page, struct list_head *pagelist,
				unsigned long flags)
906 907
{
	/*
908
	 * Avoid migrating a page that is shared with others.
909
	 */
910 911 912
	if ((flags & MPOL_MF_MOVE_ALL) || page_mapcount(page) == 1) {
		if (!isolate_lru_page(page)) {
			list_add_tail(&page->lru, pagelist);
913 914
			inc_zone_page_state(page, NR_ISOLATED_ANON +
					    page_is_file_cache(page));
915 916
		}
	}
917
}
918

919
static struct page *new_node_page(struct page *page, unsigned long node, int **x)
920
{
921
	return alloc_pages_exact_node(node, GFP_HIGHUSER_MOVABLE, 0);
922 923
}

924 925 926 927
/*
 * Migrate pages from one node to a target node.
 * Returns error or the number of pages not migrated.
 */
A
Adrian Bunk 已提交
928 929
static int migrate_to_node(struct mm_struct *mm, int source, int dest,
			   int flags)
930 931 932 933
{
	nodemask_t nmask;
	LIST_HEAD(pagelist);
	int err = 0;
934
	struct vm_area_struct *vma;
935 936 937

	nodes_clear(nmask);
	node_set(source, nmask);
938

939
	vma = check_range(mm, mm->mmap->vm_start, mm->task_size, &nmask,
940
			flags | MPOL_MF_DISCONTIG_OK, &pagelist);
941 942
	if (IS_ERR(vma))
		return PTR_ERR(vma);
943

944
	if (!list_empty(&pagelist)) {
945
		err = migrate_pages(&pagelist, new_node_page, dest,
946
							false, MIGRATE_SYNC);
947 948 949
		if (err)
			putback_lru_pages(&pagelist);
	}
950

951
	return err;
952 953
}

954
/*
955 956
 * Move pages between the two nodesets so as to preserve the physical
 * layout as much as possible.
957 958 959 960 961 962
 *
 * Returns the number of page that could not be moved.
 */
int do_migrate_pages(struct mm_struct *mm,
	const nodemask_t *from_nodes, const nodemask_t *to_nodes, int flags)
{
963
	int busy = 0;
964
	int err;
965
	nodemask_t tmp;
966

967 968 969 970
	err = migrate_prep();
	if (err)
		return err;

971
	down_read(&mm->mmap_sem);
972

973 974 975 976
	err = migrate_vmas(mm, from_nodes, to_nodes, flags);
	if (err)
		goto out;

977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001
	/*
	 * Find a 'source' bit set in 'tmp' whose corresponding 'dest'
	 * bit in 'to' is not also set in 'tmp'.  Clear the found 'source'
	 * bit in 'tmp', and return that <source, dest> pair for migration.
	 * The pair of nodemasks 'to' and 'from' define the map.
	 *
	 * If no pair of bits is found that way, fallback to picking some
	 * pair of 'source' and 'dest' bits that are not the same.  If the
	 * 'source' and 'dest' bits are the same, this represents a node
	 * that will be migrating to itself, so no pages need move.
	 *
	 * If no bits are left in 'tmp', or if all remaining bits left
	 * in 'tmp' correspond to the same bit in 'to', return false
	 * (nothing left to migrate).
	 *
	 * This lets us pick a pair of nodes to migrate between, such that
	 * if possible the dest node is not already occupied by some other
	 * source node, minimizing the risk of overloading the memory on a
	 * node that would happen if we migrated incoming memory to a node
	 * before migrating outgoing memory source that same node.
	 *
	 * A single scan of tmp is sufficient.  As we go, we remember the
	 * most recent <s, d> pair that moved (s != d).  If we find a pair
	 * that not only moved, but what's better, moved to an empty slot
	 * (d is not set in tmp), then we break out then, with that pair.
1002
	 * Otherwise when we finish scanning from_tmp, we at least have the
1003 1004 1005 1006
	 * most recent <s, d> pair that moved.  If we get all the way through
	 * the scan of tmp without finding any node that moved, much less
	 * moved to an empty node, then there is nothing left worth migrating.
	 */
1007

1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034
	tmp = *from_nodes;
	while (!nodes_empty(tmp)) {
		int s,d;
		int source = -1;
		int dest = 0;

		for_each_node_mask(s, tmp) {
			d = node_remap(s, *from_nodes, *to_nodes);
			if (s == d)
				continue;

			source = s;	/* Node moved. Memorize */
			dest = d;

			/* dest not in remaining from nodes? */
			if (!node_isset(dest, tmp))
				break;
		}
		if (source == -1)
			break;

		node_clear(source, tmp);
		err = migrate_to_node(mm, source, dest, flags);
		if (err > 0)
			busy += err;
		if (err < 0)
			break;
1035
	}
1036
out:
1037
	up_read(&mm->mmap_sem);
1038 1039 1040
	if (err < 0)
		return err;
	return busy;
C
Christoph Lameter 已提交
1041 1042 1043

}

1044 1045 1046 1047 1048 1049 1050
/*
 * Allocate a new page for page migration based on vma policy.
 * Start assuming that page is mapped by vma pointed to by @private.
 * Search forward from there, if not.  N.B., this assumes that the
 * list of pages handed to migrate_pages()--which is how we get here--
 * is in virtual address order.
 */
1051
static struct page *new_vma_page(struct page *page, unsigned long private, int **x)
1052 1053
{
	struct vm_area_struct *vma = (struct vm_area_struct *)private;
1054
	unsigned long uninitialized_var(address);
1055

1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066
	while (vma) {
		address = page_address_in_vma(page, vma);
		if (address != -EFAULT)
			break;
		vma = vma->vm_next;
	}

	/*
	 * if !vma, alloc_page_vma() will use task or system default policy
	 */
	return alloc_page_vma(GFP_HIGHUSER_MOVABLE, vma, address);
1067
}
C
Christoph Lameter 已提交
1068 1069 1070 1071 1072
#else

static void migrate_page_add(struct page *page, struct list_head *pagelist,
				unsigned long flags)
{
1073 1074
}

C
Christoph Lameter 已提交
1075 1076 1077 1078 1079
int do_migrate_pages(struct mm_struct *mm,
	const nodemask_t *from_nodes, const nodemask_t *to_nodes, int flags)
{
	return -ENOSYS;
}
1080

1081
static struct page *new_vma_page(struct page *page, unsigned long private, int **x)
1082 1083 1084
{
	return NULL;
}
C
Christoph Lameter 已提交
1085 1086
#endif

A
Adrian Bunk 已提交
1087
static long do_mbind(unsigned long start, unsigned long len,
1088 1089
		     unsigned short mode, unsigned short mode_flags,
		     nodemask_t *nmask, unsigned long flags)
1090 1091 1092 1093 1094 1095 1096 1097
{
	struct vm_area_struct *vma;
	struct mm_struct *mm = current->mm;
	struct mempolicy *new;
	unsigned long end;
	int err;
	LIST_HEAD(pagelist);

1098 1099
	if (flags & ~(unsigned long)(MPOL_MF_STRICT |
				     MPOL_MF_MOVE | MPOL_MF_MOVE_ALL))
1100
		return -EINVAL;
1101
	if ((flags & MPOL_MF_MOVE_ALL) && !capable(CAP_SYS_NICE))
1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115 1116 1117
		return -EPERM;

	if (start & ~PAGE_MASK)
		return -EINVAL;

	if (mode == MPOL_DEFAULT)
		flags &= ~MPOL_MF_STRICT;

	len = (len + PAGE_SIZE - 1) & PAGE_MASK;
	end = start + len;

	if (end < start)
		return -EINVAL;
	if (end == start)
		return 0;

1118
	new = mpol_new(mode, mode_flags, nmask);
1119 1120 1121 1122 1123 1124 1125 1126 1127 1128
	if (IS_ERR(new))
		return PTR_ERR(new);

	/*
	 * If we are using the default policy then operation
	 * on discontinuous address spaces is okay after all
	 */
	if (!new)
		flags |= MPOL_MF_DISCONTIG_OK;

1129 1130 1131
	pr_debug("mbind %lx-%lx mode:%d flags:%d nodes:%lx\n",
		 start, start + len, mode, mode_flags,
		 nmask ? nodes_addr(*nmask)[0] : -1);
1132

1133 1134 1135 1136
	if (flags & (MPOL_MF_MOVE | MPOL_MF_MOVE_ALL)) {

		err = migrate_prep();
		if (err)
K
KOSAKI Motohiro 已提交
1137
			goto mpol_out;
1138
	}
1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151
	{
		NODEMASK_SCRATCH(scratch);
		if (scratch) {
			down_write(&mm->mmap_sem);
			task_lock(current);
			err = mpol_set_nodemask(new, nmask, scratch);
			task_unlock(current);
			if (err)
				up_write(&mm->mmap_sem);
		} else
			err = -ENOMEM;
		NODEMASK_SCRATCH_FREE(scratch);
	}
K
KOSAKI Motohiro 已提交
1152 1153 1154
	if (err)
		goto mpol_out;

1155 1156 1157 1158 1159 1160 1161
	vma = check_range(mm, start, end, nmask,
			  flags | MPOL_MF_INVERT, &pagelist);

	err = PTR_ERR(vma);
	if (!IS_ERR(vma)) {
		int nr_failed = 0;

1162
		err = mbind_range(mm, start, end, new);
1163

1164
		if (!list_empty(&pagelist)) {
1165
			nr_failed = migrate_pages(&pagelist, new_vma_page,
1166 1167
						(unsigned long)vma,
						false, true);
1168 1169 1170
			if (nr_failed)
				putback_lru_pages(&pagelist);
		}
1171 1172 1173

		if (!err && nr_failed && (flags & MPOL_MF_STRICT))
			err = -EIO;
1174 1175
	} else
		putback_lru_pages(&pagelist);
C
Christoph Lameter 已提交
1176

1177
	up_write(&mm->mmap_sem);
K
KOSAKI Motohiro 已提交
1178
 mpol_out:
1179
	mpol_put(new);
1180 1181 1182
	return err;
}

1183 1184 1185 1186 1187
/*
 * User space interface with variable sized bitmaps for nodelists.
 */

/* Copy a node mask from user space. */
1188
static int get_nodes(nodemask_t *nodes, const unsigned long __user *nmask,
1189 1190 1191 1192 1193 1194 1195 1196 1197 1198
		     unsigned long maxnode)
{
	unsigned long k;
	unsigned long nlongs;
	unsigned long endmask;

	--maxnode;
	nodes_clear(*nodes);
	if (maxnode == 0 || !nmask)
		return 0;
A
Andi Kleen 已提交
1199
	if (maxnode > PAGE_SIZE*BITS_PER_BYTE)
C
Chris Wright 已提交
1200
		return -EINVAL;
1201 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 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249

	nlongs = BITS_TO_LONGS(maxnode);
	if ((maxnode % BITS_PER_LONG) == 0)
		endmask = ~0UL;
	else
		endmask = (1UL << (maxnode % BITS_PER_LONG)) - 1;

	/* When the user specified more nodes than supported just check
	   if the non supported part is all zero. */
	if (nlongs > BITS_TO_LONGS(MAX_NUMNODES)) {
		if (nlongs > PAGE_SIZE/sizeof(long))
			return -EINVAL;
		for (k = BITS_TO_LONGS(MAX_NUMNODES); k < nlongs; k++) {
			unsigned long t;
			if (get_user(t, nmask + k))
				return -EFAULT;
			if (k == nlongs - 1) {
				if (t & endmask)
					return -EINVAL;
			} else if (t)
				return -EINVAL;
		}
		nlongs = BITS_TO_LONGS(MAX_NUMNODES);
		endmask = ~0UL;
	}

	if (copy_from_user(nodes_addr(*nodes), nmask, nlongs*sizeof(unsigned long)))
		return -EFAULT;
	nodes_addr(*nodes)[nlongs-1] &= endmask;
	return 0;
}

/* Copy a kernel node mask to user space */
static int copy_nodes_to_user(unsigned long __user *mask, unsigned long maxnode,
			      nodemask_t *nodes)
{
	unsigned long copy = ALIGN(maxnode-1, 64) / 8;
	const int nbytes = BITS_TO_LONGS(MAX_NUMNODES) * sizeof(long);

	if (copy > nbytes) {
		if (copy > PAGE_SIZE)
			return -EINVAL;
		if (clear_user((char __user *)mask + nbytes, copy - nbytes))
			return -EFAULT;
		copy = nbytes;
	}
	return copy_to_user(mask, nodes_addr(*nodes), copy) ? -EFAULT : 0;
}

1250 1251 1252
SYSCALL_DEFINE6(mbind, unsigned long, start, unsigned long, len,
		unsigned long, mode, unsigned long __user *, nmask,
		unsigned long, maxnode, unsigned, flags)
1253 1254 1255
{
	nodemask_t nodes;
	int err;
1256
	unsigned short mode_flags;
1257

1258 1259
	mode_flags = mode & MPOL_MODE_FLAGS;
	mode &= ~MPOL_MODE_FLAGS;
1260 1261
	if (mode >= MPOL_MAX)
		return -EINVAL;
1262 1263 1264
	if ((mode_flags & MPOL_F_STATIC_NODES) &&
	    (mode_flags & MPOL_F_RELATIVE_NODES))
		return -EINVAL;
1265 1266 1267
	err = get_nodes(&nodes, nmask, maxnode);
	if (err)
		return err;
1268
	return do_mbind(start, len, mode, mode_flags, &nodes, flags);
1269 1270 1271
}

/* Set the process memory policy */
1272 1273
SYSCALL_DEFINE3(set_mempolicy, int, mode, unsigned long __user *, nmask,
		unsigned long, maxnode)
1274 1275 1276
{
	int err;
	nodemask_t nodes;
1277
	unsigned short flags;
1278

1279 1280 1281
	flags = mode & MPOL_MODE_FLAGS;
	mode &= ~MPOL_MODE_FLAGS;
	if ((unsigned int)mode >= MPOL_MAX)
1282
		return -EINVAL;
1283 1284
	if ((flags & MPOL_F_STATIC_NODES) && (flags & MPOL_F_RELATIVE_NODES))
		return -EINVAL;
1285 1286 1287
	err = get_nodes(&nodes, nmask, maxnode);
	if (err)
		return err;
1288
	return do_set_mempolicy(mode, flags, &nodes);
1289 1290
}

1291 1292 1293
SYSCALL_DEFINE4(migrate_pages, pid_t, pid, unsigned long, maxnode,
		const unsigned long __user *, old_nodes,
		const unsigned long __user *, new_nodes)
1294
{
1295
	const struct cred *cred = current_cred(), *tcred;
1296
	struct mm_struct *mm = NULL;
1297 1298 1299
	struct task_struct *task;
	nodemask_t task_nodes;
	int err;
1300 1301 1302 1303 1304 1305
	nodemask_t *old;
	nodemask_t *new;
	NODEMASK_SCRATCH(scratch);

	if (!scratch)
		return -ENOMEM;
1306

1307 1308 1309 1310
	old = &scratch->mask1;
	new = &scratch->mask2;

	err = get_nodes(old, old_nodes, maxnode);
1311
	if (err)
1312
		goto out;
1313

1314
	err = get_nodes(new, new_nodes, maxnode);
1315
	if (err)
1316
		goto out;
1317 1318

	/* Find the mm_struct */
1319
	rcu_read_lock();
1320
	task = pid ? find_task_by_vpid(pid) : current;
1321
	if (!task) {
1322
		rcu_read_unlock();
1323 1324
		err = -ESRCH;
		goto out;
1325
	}
1326
	get_task_struct(task);
1327

1328
	err = -EINVAL;
1329 1330 1331 1332

	/*
	 * Check if this process has the right to modify the specified
	 * process. The right exists if the process has administrative
A
Alexey Dobriyan 已提交
1333
	 * capabilities, superuser privileges or the same
1334 1335
	 * userid as the target process.
	 */
1336
	tcred = __task_cred(task);
1337 1338
	if (cred->euid != tcred->suid && cred->euid != tcred->uid &&
	    cred->uid  != tcred->suid && cred->uid  != tcred->uid &&
1339
	    !capable(CAP_SYS_NICE)) {
1340
		rcu_read_unlock();
1341
		err = -EPERM;
1342
		goto out_put;
1343
	}
1344
	rcu_read_unlock();
1345 1346 1347

	task_nodes = cpuset_mems_allowed(task);
	/* Is the user allowed to access the target nodes? */
1348
	if (!nodes_subset(*new, task_nodes) && !capable(CAP_SYS_NICE)) {
1349
		err = -EPERM;
1350
		goto out_put;
1351 1352
	}

1353
	if (!nodes_subset(*new, node_states[N_HIGH_MEMORY])) {
1354
		err = -EINVAL;
1355
		goto out_put;
1356 1357
	}

1358 1359
	err = security_task_movememory(task);
	if (err)
1360
		goto out_put;
1361

1362 1363
	mm = get_task_mm(task);
	put_task_struct(task);
1364
	if (mm)
1365 1366 1367 1368 1369 1370 1371
		err = do_migrate_pages(mm, old, new,
			capable(CAP_SYS_NICE) ? MPOL_MF_MOVE_ALL : MPOL_MF_MOVE);
	else
		err = -EINVAL;

	mmput(mm);
out:
1372 1373
	NODEMASK_SCRATCH_FREE(scratch);

1374
	return err;
1375 1376 1377 1378 1379

out_put:
	put_task_struct(task);
	goto out;

1380 1381 1382
}


1383
/* Retrieve NUMA policy */
1384 1385 1386
SYSCALL_DEFINE5(get_mempolicy, int __user *, policy,
		unsigned long __user *, nmask, unsigned long, maxnode,
		unsigned long, addr, unsigned long, flags)
1387
{
A
Adrian Bunk 已提交
1388 1389
	int err;
	int uninitialized_var(pval);
1390 1391 1392 1393 1394 1395 1396 1397 1398 1399 1400 1401 1402 1403 1404 1405 1406 1407 1408
	nodemask_t nodes;

	if (nmask != NULL && maxnode < MAX_NUMNODES)
		return -EINVAL;

	err = do_get_mempolicy(&pval, &nodes, addr, flags);

	if (err)
		return err;

	if (policy && put_user(pval, policy))
		return -EFAULT;

	if (nmask)
		err = copy_nodes_to_user(nmask, maxnode, &nodes);

	return err;
}

L
Linus Torvalds 已提交
1409 1410 1411 1412 1413 1414 1415 1416 1417 1418 1419 1420 1421 1422 1423 1424 1425 1426 1427 1428 1429
#ifdef CONFIG_COMPAT

asmlinkage long compat_sys_get_mempolicy(int __user *policy,
				     compat_ulong_t __user *nmask,
				     compat_ulong_t maxnode,
				     compat_ulong_t addr, compat_ulong_t flags)
{
	long err;
	unsigned long __user *nm = NULL;
	unsigned long nr_bits, alloc_size;
	DECLARE_BITMAP(bm, MAX_NUMNODES);

	nr_bits = min_t(unsigned long, maxnode-1, MAX_NUMNODES);
	alloc_size = ALIGN(nr_bits, BITS_PER_LONG) / 8;

	if (nmask)
		nm = compat_alloc_user_space(alloc_size);

	err = sys_get_mempolicy(policy, nm, nr_bits+1, addr, flags);

	if (!err && nmask) {
1430 1431 1432
		unsigned long copy_size;
		copy_size = min_t(unsigned long, sizeof(bm), alloc_size);
		err = copy_from_user(bm, nm, copy_size);
L
Linus Torvalds 已提交
1433 1434 1435 1436 1437 1438 1439 1440 1441 1442 1443 1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470
		/* ensure entire bitmap is zeroed */
		err |= clear_user(nmask, ALIGN(maxnode-1, 8) / 8);
		err |= compat_put_bitmap(nmask, bm, nr_bits);
	}

	return err;
}

asmlinkage long compat_sys_set_mempolicy(int mode, compat_ulong_t __user *nmask,
				     compat_ulong_t maxnode)
{
	long err = 0;
	unsigned long __user *nm = NULL;
	unsigned long nr_bits, alloc_size;
	DECLARE_BITMAP(bm, MAX_NUMNODES);

	nr_bits = min_t(unsigned long, maxnode-1, MAX_NUMNODES);
	alloc_size = ALIGN(nr_bits, BITS_PER_LONG) / 8;

	if (nmask) {
		err = compat_get_bitmap(bm, nmask, nr_bits);
		nm = compat_alloc_user_space(alloc_size);
		err |= copy_to_user(nm, bm, alloc_size);
	}

	if (err)
		return -EFAULT;

	return sys_set_mempolicy(mode, nm, nr_bits+1);
}

asmlinkage long compat_sys_mbind(compat_ulong_t start, compat_ulong_t len,
			     compat_ulong_t mode, compat_ulong_t __user *nmask,
			     compat_ulong_t maxnode, compat_ulong_t flags)
{
	long err = 0;
	unsigned long __user *nm = NULL;
	unsigned long nr_bits, alloc_size;
1471
	nodemask_t bm;
L
Linus Torvalds 已提交
1472 1473 1474 1475 1476

	nr_bits = min_t(unsigned long, maxnode-1, MAX_NUMNODES);
	alloc_size = ALIGN(nr_bits, BITS_PER_LONG) / 8;

	if (nmask) {
1477
		err = compat_get_bitmap(nodes_addr(bm), nmask, nr_bits);
L
Linus Torvalds 已提交
1478
		nm = compat_alloc_user_space(alloc_size);
1479
		err |= copy_to_user(nm, nodes_addr(bm), alloc_size);
L
Linus Torvalds 已提交
1480 1481 1482 1483 1484 1485 1486 1487 1488 1489
	}

	if (err)
		return -EFAULT;

	return sys_mbind(start, len, mode, nm, nr_bits+1, flags);
}

#endif

1490 1491 1492 1493 1494 1495 1496 1497
/*
 * get_vma_policy(@task, @vma, @addr)
 * @task - task for fallback if vma policy == default
 * @vma   - virtual memory area whose policy is sought
 * @addr  - address in @vma for shared policy lookup
 *
 * Returns effective policy for a VMA at specified address.
 * Falls back to @task or system default policy, as necessary.
1498 1499 1500 1501 1502 1503 1504
 * Current or other task's task mempolicy and non-shared vma policies
 * are protected by the task's mmap_sem, which must be held for read by
 * the caller.
 * Shared policies [those marked as MPOL_F_SHARED] require an extra reference
 * count--added by the get_policy() vm_op, as appropriate--to protect against
 * freeing by another task.  It is the caller's responsibility to free the
 * extra reference for shared policies.
1505
 */
S
Stephen Wilson 已提交
1506
struct mempolicy *get_vma_policy(struct task_struct *task,
1507
		struct vm_area_struct *vma, unsigned long addr)
L
Linus Torvalds 已提交
1508
{
1509
	struct mempolicy *pol = task->mempolicy;
L
Linus Torvalds 已提交
1510 1511

	if (vma) {
1512
		if (vma->vm_ops && vma->vm_ops->get_policy) {
1513 1514 1515 1516
			struct mempolicy *vpol = vma->vm_ops->get_policy(vma,
									addr);
			if (vpol)
				pol = vpol;
1517
		} else if (vma->vm_policy)
L
Linus Torvalds 已提交
1518 1519 1520 1521 1522 1523 1524
			pol = vma->vm_policy;
	}
	if (!pol)
		pol = &default_policy;
	return pol;
}

1525 1526 1527 1528 1529
/*
 * Return a nodemask representing a mempolicy for filtering nodes for
 * page allocation
 */
static nodemask_t *policy_nodemask(gfp_t gfp, struct mempolicy *policy)
1530 1531
{
	/* Lower zones don't get a nodemask applied for MPOL_BIND */
1532
	if (unlikely(policy->mode == MPOL_BIND) &&
1533 1534 1535 1536 1537 1538 1539
			gfp_zone(gfp) >= policy_zone &&
			cpuset_nodemask_valid_mems_allowed(&policy->v.nodes))
		return &policy->v.nodes;

	return NULL;
}

1540
/* Return a zonelist indicated by gfp for node representing a mempolicy */
1541 1542
static struct zonelist *policy_zonelist(gfp_t gfp, struct mempolicy *policy,
	int nd)
L
Linus Torvalds 已提交
1543
{
1544
	switch (policy->mode) {
L
Linus Torvalds 已提交
1545
	case MPOL_PREFERRED:
1546 1547
		if (!(policy->flags & MPOL_F_LOCAL))
			nd = policy->v.preferred_node;
L
Linus Torvalds 已提交
1548 1549
		break;
	case MPOL_BIND:
1550
		/*
1551 1552
		 * Normally, MPOL_BIND allocations are node-local within the
		 * allowed nodemask.  However, if __GFP_THISNODE is set and the
1553
		 * current node isn't part of the mask, we use the zonelist for
1554
		 * the first node in the mask instead.
1555 1556 1557 1558 1559
		 */
		if (unlikely(gfp & __GFP_THISNODE) &&
				unlikely(!node_isset(nd, policy->v.nodes)))
			nd = first_node(policy->v.nodes);
		break;
L
Linus Torvalds 已提交
1560 1561 1562
	default:
		BUG();
	}
1563
	return node_zonelist(nd, gfp);
L
Linus Torvalds 已提交
1564 1565 1566 1567 1568 1569 1570 1571 1572
}

/* Do dynamic interleaving for a process */
static unsigned interleave_nodes(struct mempolicy *policy)
{
	unsigned nid, next;
	struct task_struct *me = current;

	nid = me->il_next;
1573
	next = next_node(nid, policy->v.nodes);
L
Linus Torvalds 已提交
1574
	if (next >= MAX_NUMNODES)
1575
		next = first_node(policy->v.nodes);
1576 1577
	if (next < MAX_NUMNODES)
		me->il_next = next;
L
Linus Torvalds 已提交
1578 1579 1580
	return nid;
}

1581 1582 1583
/*
 * Depending on the memory policy provide a node from which to allocate the
 * next slab entry.
1584 1585 1586 1587
 * @policy must be protected by freeing by the caller.  If @policy is
 * the current task's mempolicy, this protection is implicit, as only the
 * task can change it's policy.  The system default policy requires no
 * such protection.
1588 1589 1590
 */
unsigned slab_node(struct mempolicy *policy)
{
1591
	if (!policy || policy->flags & MPOL_F_LOCAL)
1592 1593 1594 1595
		return numa_node_id();

	switch (policy->mode) {
	case MPOL_PREFERRED:
1596 1597 1598 1599
		/*
		 * handled MPOL_F_LOCAL above
		 */
		return policy->v.preferred_node;
1600

1601 1602 1603
	case MPOL_INTERLEAVE:
		return interleave_nodes(policy);

1604
	case MPOL_BIND: {
1605 1606 1607 1608
		/*
		 * Follow bind policy behavior and start allocation at the
		 * first node.
		 */
1609 1610 1611 1612 1613 1614 1615
		struct zonelist *zonelist;
		struct zone *zone;
		enum zone_type highest_zoneidx = gfp_zone(GFP_KERNEL);
		zonelist = &NODE_DATA(numa_node_id())->node_zonelists[0];
		(void)first_zones_zonelist(zonelist, highest_zoneidx,
							&policy->v.nodes,
							&zone);
E
Eric Dumazet 已提交
1616
		return zone ? zone->node : numa_node_id();
1617
	}
1618 1619

	default:
1620
		BUG();
1621 1622 1623
	}
}

L
Linus Torvalds 已提交
1624 1625 1626 1627
/* Do static interleaving for a VMA with known offset. */
static unsigned offset_il_node(struct mempolicy *pol,
		struct vm_area_struct *vma, unsigned long off)
{
1628
	unsigned nnodes = nodes_weight(pol->v.nodes);
1629
	unsigned target;
L
Linus Torvalds 已提交
1630 1631 1632
	int c;
	int nid = -1;

1633 1634 1635
	if (!nnodes)
		return numa_node_id();
	target = (unsigned int)off % nnodes;
L
Linus Torvalds 已提交
1636 1637
	c = 0;
	do {
1638
		nid = next_node(nid, pol->v.nodes);
L
Linus Torvalds 已提交
1639 1640 1641 1642 1643
		c++;
	} while (c <= target);
	return nid;
}

1644 1645 1646 1647 1648 1649 1650
/* Determine a node number for interleave */
static inline unsigned interleave_nid(struct mempolicy *pol,
		 struct vm_area_struct *vma, unsigned long addr, int shift)
{
	if (vma) {
		unsigned long off;

1651 1652 1653 1654 1655 1656 1657 1658 1659
		/*
		 * for small pages, there is no difference between
		 * shift and PAGE_SHIFT, so the bit-shift is safe.
		 * for huge pages, since vm_pgoff is in units of small
		 * pages, we need to shift off the always 0 bits to get
		 * a useful offset.
		 */
		BUG_ON(shift < PAGE_SHIFT);
		off = vma->vm_pgoff >> (shift - PAGE_SHIFT);
1660 1661 1662 1663 1664 1665
		off += (addr - vma->vm_start) >> shift;
		return offset_il_node(pol, vma, off);
	} else
		return interleave_nodes(pol);
}

1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680
/*
 * Return the bit number of a random bit set in the nodemask.
 * (returns -1 if nodemask is empty)
 */
int node_random(const nodemask_t *maskp)
{
	int w, bit = -1;

	w = nodes_weight(*maskp);
	if (w)
		bit = bitmap_ord_to_pos(maskp->bits,
			get_random_int() % w, MAX_NUMNODES);
	return bit;
}

1681
#ifdef CONFIG_HUGETLBFS
1682 1683 1684 1685 1686
/*
 * huge_zonelist(@vma, @addr, @gfp_flags, @mpol)
 * @vma = virtual memory area whose policy is sought
 * @addr = address in @vma for shared policy lookup and interleave policy
 * @gfp_flags = for requested zone
1687 1688
 * @mpol = pointer to mempolicy pointer for reference counted mempolicy
 * @nodemask = pointer to nodemask pointer for MPOL_BIND nodemask
1689
 *
1690 1691 1692 1693
 * Returns a zonelist suitable for a huge page allocation and a pointer
 * to the struct mempolicy for conditional unref after allocation.
 * If the effective policy is 'BIND, returns a pointer to the mempolicy's
 * @nodemask for filtering the zonelist.
1694 1695
 *
 * Must be protected by get_mems_allowed()
1696
 */
1697
struct zonelist *huge_zonelist(struct vm_area_struct *vma, unsigned long addr,
1698 1699
				gfp_t gfp_flags, struct mempolicy **mpol,
				nodemask_t **nodemask)
1700
{
1701
	struct zonelist *zl;
1702

1703
	*mpol = get_vma_policy(current, vma, addr);
1704
	*nodemask = NULL;	/* assume !MPOL_BIND */
1705

1706 1707
	if (unlikely((*mpol)->mode == MPOL_INTERLEAVE)) {
		zl = node_zonelist(interleave_nid(*mpol, vma, addr,
1708
				huge_page_shift(hstate_vma(vma))), gfp_flags);
1709
	} else {
1710
		zl = policy_zonelist(gfp_flags, *mpol, numa_node_id());
1711 1712
		if ((*mpol)->mode == MPOL_BIND)
			*nodemask = &(*mpol)->v.nodes;
1713 1714
	}
	return zl;
1715
}
1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730 1731 1732 1733 1734 1735 1736 1737 1738 1739 1740

/*
 * init_nodemask_of_mempolicy
 *
 * If the current task's mempolicy is "default" [NULL], return 'false'
 * to indicate default policy.  Otherwise, extract the policy nodemask
 * for 'bind' or 'interleave' policy into the argument nodemask, or
 * initialize the argument nodemask to contain the single node for
 * 'preferred' or 'local' policy and return 'true' to indicate presence
 * of non-default mempolicy.
 *
 * We don't bother with reference counting the mempolicy [mpol_get/put]
 * because the current task is examining it's own mempolicy and a task's
 * mempolicy is only ever changed by the task itself.
 *
 * N.B., it is the caller's responsibility to free a returned nodemask.
 */
bool init_nodemask_of_mempolicy(nodemask_t *mask)
{
	struct mempolicy *mempolicy;
	int nid;

	if (!(mask && current->mempolicy))
		return false;

1741
	task_lock(current);
1742 1743 1744 1745 1746 1747 1748 1749 1750 1751 1752 1753 1754 1755 1756 1757 1758 1759 1760
	mempolicy = current->mempolicy;
	switch (mempolicy->mode) {
	case MPOL_PREFERRED:
		if (mempolicy->flags & MPOL_F_LOCAL)
			nid = numa_node_id();
		else
			nid = mempolicy->v.preferred_node;
		init_nodemask_of_node(mask, nid);
		break;

	case MPOL_BIND:
		/* Fall through */
	case MPOL_INTERLEAVE:
		*mask =  mempolicy->v.nodes;
		break;

	default:
		BUG();
	}
1761
	task_unlock(current);
1762 1763 1764

	return true;
}
1765
#endif
1766

1767 1768 1769 1770 1771 1772 1773 1774 1775 1776 1777 1778 1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790 1791 1792 1793 1794 1795 1796 1797 1798 1799 1800 1801 1802 1803 1804 1805 1806 1807 1808 1809 1810
/*
 * mempolicy_nodemask_intersects
 *
 * If tsk's mempolicy is "default" [NULL], return 'true' to indicate default
 * policy.  Otherwise, check for intersection between mask and the policy
 * nodemask for 'bind' or 'interleave' policy.  For 'perferred' or 'local'
 * policy, always return true since it may allocate elsewhere on fallback.
 *
 * Takes task_lock(tsk) to prevent freeing of its mempolicy.
 */
bool mempolicy_nodemask_intersects(struct task_struct *tsk,
					const nodemask_t *mask)
{
	struct mempolicy *mempolicy;
	bool ret = true;

	if (!mask)
		return ret;
	task_lock(tsk);
	mempolicy = tsk->mempolicy;
	if (!mempolicy)
		goto out;

	switch (mempolicy->mode) {
	case MPOL_PREFERRED:
		/*
		 * MPOL_PREFERRED and MPOL_F_LOCAL are only preferred nodes to
		 * allocate from, they may fallback to other nodes when oom.
		 * Thus, it's possible for tsk to have allocated memory from
		 * nodes in mask.
		 */
		break;
	case MPOL_BIND:
	case MPOL_INTERLEAVE:
		ret = nodes_intersects(mempolicy->v.nodes, *mask);
		break;
	default:
		BUG();
	}
out:
	task_unlock(tsk);
	return ret;
}

L
Linus Torvalds 已提交
1811 1812
/* Allocate a page in interleaved policy.
   Own path because it needs to do special accounting. */
1813 1814
static struct page *alloc_page_interleave(gfp_t gfp, unsigned order,
					unsigned nid)
L
Linus Torvalds 已提交
1815 1816 1817 1818
{
	struct zonelist *zl;
	struct page *page;

1819
	zl = node_zonelist(nid, gfp);
L
Linus Torvalds 已提交
1820
	page = __alloc_pages(gfp, order, zl);
1821
	if (page && page_zone(page) == zonelist_zone(&zl->_zonerefs[0]))
1822
		inc_zone_page_state(page, NUMA_INTERLEAVE_HIT);
L
Linus Torvalds 已提交
1823 1824 1825 1826
	return page;
}

/**
1827
 * 	alloc_pages_vma	- Allocate a page for a VMA.
L
Linus Torvalds 已提交
1828 1829 1830 1831 1832 1833 1834 1835
 *
 * 	@gfp:
 *      %GFP_USER    user allocation.
 *      %GFP_KERNEL  kernel allocations,
 *      %GFP_HIGHMEM highmem/user allocations,
 *      %GFP_FS      allocation should not call back into a file system.
 *      %GFP_ATOMIC  don't sleep.
 *
1836
 *	@order:Order of the GFP allocation.
L
Linus Torvalds 已提交
1837 1838 1839 1840 1841 1842 1843 1844 1845 1846 1847 1848 1849
 * 	@vma:  Pointer to VMA or NULL if not available.
 *	@addr: Virtual Address of the allocation. Must be inside the VMA.
 *
 * 	This function allocates a page from the kernel page pool and applies
 *	a NUMA policy associated with the VMA or the current process.
 *	When VMA is not NULL caller must hold down_read on the mmap_sem of the
 *	mm_struct of the VMA to prevent it from going away. Should be used for
 *	all allocations for pages that will be mapped into
 * 	user space. Returns NULL when no page can be allocated.
 *
 *	Should be called with the mm_sem of the vma hold.
 */
struct page *
1850
alloc_pages_vma(gfp_t gfp, int order, struct vm_area_struct *vma,
1851
		unsigned long addr, int node)
L
Linus Torvalds 已提交
1852
{
1853
	struct mempolicy *pol;
1854
	struct zonelist *zl;
1855
	struct page *page;
1856 1857 1858 1859 1860
	unsigned int cpuset_mems_cookie;

retry_cpuset:
	pol = get_vma_policy(current, vma, addr);
	cpuset_mems_cookie = get_mems_allowed();
L
Linus Torvalds 已提交
1861

1862
	if (unlikely(pol->mode == MPOL_INTERLEAVE)) {
L
Linus Torvalds 已提交
1863
		unsigned nid;
1864

1865
		nid = interleave_nid(pol, vma, addr, PAGE_SHIFT + order);
1866
		mpol_cond_put(pol);
1867
		page = alloc_page_interleave(gfp, order, nid);
1868 1869 1870
		if (unlikely(!put_mems_allowed(cpuset_mems_cookie) && !page))
			goto retry_cpuset;

1871
		return page;
L
Linus Torvalds 已提交
1872
	}
1873
	zl = policy_zonelist(gfp, pol, node);
1874
	if (unlikely(mpol_needs_cond_ref(pol))) {
1875
		/*
1876
		 * slow path: ref counted shared policy
1877
		 */
1878
		struct page *page =  __alloc_pages_nodemask(gfp, order,
1879
						zl, policy_nodemask(gfp, pol));
1880
		__mpol_put(pol);
1881 1882
		if (unlikely(!put_mems_allowed(cpuset_mems_cookie) && !page))
			goto retry_cpuset;
1883 1884 1885 1886 1887
		return page;
	}
	/*
	 * fast path:  default or task policy
	 */
1888 1889
	page = __alloc_pages_nodemask(gfp, order, zl,
				      policy_nodemask(gfp, pol));
1890 1891
	if (unlikely(!put_mems_allowed(cpuset_mems_cookie) && !page))
		goto retry_cpuset;
1892
	return page;
L
Linus Torvalds 已提交
1893 1894 1895 1896 1897 1898 1899 1900 1901 1902 1903 1904 1905 1906 1907 1908 1909
}

/**
 * 	alloc_pages_current - Allocate pages.
 *
 *	@gfp:
 *		%GFP_USER   user allocation,
 *      	%GFP_KERNEL kernel allocation,
 *      	%GFP_HIGHMEM highmem allocation,
 *      	%GFP_FS     don't call back into a file system.
 *      	%GFP_ATOMIC don't sleep.
 *	@order: Power of two of allocation size in pages. 0 is a single page.
 *
 *	Allocate a page from the kernel page pool.  When not in
 *	interrupt context and apply the current process NUMA policy.
 *	Returns NULL when no page can be allocated.
 *
1910
 *	Don't call cpuset_update_task_memory_state() unless
L
Linus Torvalds 已提交
1911 1912 1913
 *	1) it's ok to take cpuset_sem (can WAIT), and
 *	2) allocating for current task (not interrupt).
 */
A
Al Viro 已提交
1914
struct page *alloc_pages_current(gfp_t gfp, unsigned order)
L
Linus Torvalds 已提交
1915 1916
{
	struct mempolicy *pol = current->mempolicy;
1917
	struct page *page;
1918
	unsigned int cpuset_mems_cookie;
L
Linus Torvalds 已提交
1919

1920
	if (!pol || in_interrupt() || (gfp & __GFP_THISNODE))
L
Linus Torvalds 已提交
1921
		pol = &default_policy;
1922

1923 1924 1925
retry_cpuset:
	cpuset_mems_cookie = get_mems_allowed();

1926 1927 1928 1929
	/*
	 * No reference counting needed for current->mempolicy
	 * nor system default_policy
	 */
1930
	if (pol->mode == MPOL_INTERLEAVE)
1931 1932 1933
		page = alloc_page_interleave(gfp, order, interleave_nodes(pol));
	else
		page = __alloc_pages_nodemask(gfp, order,
1934 1935
				policy_zonelist(gfp, pol, numa_node_id()),
				policy_nodemask(gfp, pol));
1936 1937 1938 1939

	if (unlikely(!put_mems_allowed(cpuset_mems_cookie) && !page))
		goto retry_cpuset;

1940
	return page;
L
Linus Torvalds 已提交
1941 1942 1943
}
EXPORT_SYMBOL(alloc_pages_current);

1944
/*
1945
 * If mpol_dup() sees current->cpuset == cpuset_being_rebound, then it
1946 1947 1948 1949
 * rebinds the mempolicy its copying by calling mpol_rebind_policy()
 * with the mems_allowed returned by cpuset_mems_allowed().  This
 * keeps mempolicies cpuset relative after its cpuset moves.  See
 * further kernel/cpuset.c update_nodemask().
1950 1951 1952
 *
 * current's mempolicy may be rebinded by the other task(the task that changes
 * cpuset's mems), so we needn't do rebind work for current task.
1953 1954
 */

1955 1956
/* Slow path of a mempolicy duplicate */
struct mempolicy *__mpol_dup(struct mempolicy *old)
L
Linus Torvalds 已提交
1957 1958 1959 1960 1961
{
	struct mempolicy *new = kmem_cache_alloc(policy_cache, GFP_KERNEL);

	if (!new)
		return ERR_PTR(-ENOMEM);
1962 1963 1964 1965 1966 1967 1968 1969 1970

	/* task's mempolicy is protected by alloc_lock */
	if (old == current->mempolicy) {
		task_lock(current);
		*new = *old;
		task_unlock(current);
	} else
		*new = *old;

1971
	rcu_read_lock();
1972 1973
	if (current_cpuset_is_being_rebound()) {
		nodemask_t mems = cpuset_mems_allowed(current);
1974 1975 1976 1977
		if (new->flags & MPOL_F_REBINDING)
			mpol_rebind_policy(new, &mems, MPOL_REBIND_STEP2);
		else
			mpol_rebind_policy(new, &mems, MPOL_REBIND_ONCE);
1978
	}
1979
	rcu_read_unlock();
L
Linus Torvalds 已提交
1980 1981 1982 1983
	atomic_set(&new->refcnt, 1);
	return new;
}

1984 1985 1986 1987 1988 1989 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
/*
 * If *frompol needs [has] an extra ref, copy *frompol to *tompol ,
 * eliminate the * MPOL_F_* flags that require conditional ref and
 * [NOTE!!!] drop the extra ref.  Not safe to reference *frompol directly
 * after return.  Use the returned value.
 *
 * Allows use of a mempolicy for, e.g., multiple allocations with a single
 * policy lookup, even if the policy needs/has extra ref on lookup.
 * shmem_readahead needs this.
 */
struct mempolicy *__mpol_cond_copy(struct mempolicy *tompol,
						struct mempolicy *frompol)
{
	if (!mpol_needs_cond_ref(frompol))
		return frompol;

	*tompol = *frompol;
	tompol->flags &= ~MPOL_F_SHARED;	/* copy doesn't need unref */
	__mpol_put(frompol);
	return tompol;
}

L
Linus Torvalds 已提交
2006
/* Slow path of a mempolicy comparison */
2007
bool __mpol_equal(struct mempolicy *a, struct mempolicy *b)
L
Linus Torvalds 已提交
2008 2009
{
	if (!a || !b)
2010
		return false;
2011
	if (a->mode != b->mode)
2012
		return false;
B
Bob Liu 已提交
2013
	if (a->flags != b->flags)
2014
		return false;
B
Bob Liu 已提交
2015 2016
	if (mpol_store_user_nodemask(a))
		if (!nodes_equal(a->w.user_nodemask, b->w.user_nodemask))
2017
			return false;
B
Bob Liu 已提交
2018

2019
	switch (a->mode) {
2020 2021
	case MPOL_BIND:
		/* Fall through */
L
Linus Torvalds 已提交
2022
	case MPOL_INTERLEAVE:
2023
		return !!nodes_equal(a->v.nodes, b->v.nodes);
L
Linus Torvalds 已提交
2024
	case MPOL_PREFERRED:
2025
		return a->v.preferred_node == b->v.preferred_node;
L
Linus Torvalds 已提交
2026 2027
	default:
		BUG();
2028
		return false;
L
Linus Torvalds 已提交
2029 2030 2031 2032 2033 2034 2035 2036 2037 2038 2039 2040 2041 2042 2043 2044 2045 2046 2047 2048 2049 2050 2051 2052 2053 2054 2055 2056 2057 2058 2059 2060 2061 2062 2063 2064 2065 2066 2067 2068 2069 2070 2071 2072 2073 2074 2075 2076 2077 2078 2079 2080 2081 2082 2083 2084 2085 2086 2087 2088 2089 2090 2091 2092
	}
}

/*
 * Shared memory backing store policy support.
 *
 * Remember policies even when nobody has shared memory mapped.
 * The policies are kept in Red-Black tree linked from the inode.
 * They are protected by the sp->lock spinlock, which should be held
 * for any accesses to the tree.
 */

/* lookup first element intersecting start-end */
/* Caller holds sp->lock */
static struct sp_node *
sp_lookup(struct shared_policy *sp, unsigned long start, unsigned long end)
{
	struct rb_node *n = sp->root.rb_node;

	while (n) {
		struct sp_node *p = rb_entry(n, struct sp_node, nd);

		if (start >= p->end)
			n = n->rb_right;
		else if (end <= p->start)
			n = n->rb_left;
		else
			break;
	}
	if (!n)
		return NULL;
	for (;;) {
		struct sp_node *w = NULL;
		struct rb_node *prev = rb_prev(n);
		if (!prev)
			break;
		w = rb_entry(prev, struct sp_node, nd);
		if (w->end <= start)
			break;
		n = prev;
	}
	return rb_entry(n, struct sp_node, nd);
}

/* Insert a new shared policy into the list. */
/* Caller holds sp->lock */
static void sp_insert(struct shared_policy *sp, struct sp_node *new)
{
	struct rb_node **p = &sp->root.rb_node;
	struct rb_node *parent = NULL;
	struct sp_node *nd;

	while (*p) {
		parent = *p;
		nd = rb_entry(parent, struct sp_node, nd);
		if (new->start < nd->start)
			p = &(*p)->rb_left;
		else if (new->end > nd->end)
			p = &(*p)->rb_right;
		else
			BUG();
	}
	rb_link_node(&new->nd, parent, p);
	rb_insert_color(&new->nd, &sp->root);
2093
	pr_debug("inserting %lx-%lx: %d\n", new->start, new->end,
2094
		 new->policy ? new->policy->mode : 0);
L
Linus Torvalds 已提交
2095 2096 2097 2098 2099 2100 2101 2102 2103 2104 2105 2106 2107 2108 2109 2110 2111 2112 2113 2114 2115 2116 2117
}

/* Find shared policy intersecting idx */
struct mempolicy *
mpol_shared_policy_lookup(struct shared_policy *sp, unsigned long idx)
{
	struct mempolicy *pol = NULL;
	struct sp_node *sn;

	if (!sp->root.rb_node)
		return NULL;
	spin_lock(&sp->lock);
	sn = sp_lookup(sp, idx, idx+1);
	if (sn) {
		mpol_get(sn->policy);
		pol = sn->policy;
	}
	spin_unlock(&sp->lock);
	return pol;
}

static void sp_delete(struct shared_policy *sp, struct sp_node *n)
{
2118
	pr_debug("deleting %lx-l%lx\n", n->start, n->end);
L
Linus Torvalds 已提交
2119
	rb_erase(&n->nd, &sp->root);
2120
	mpol_put(n->policy);
L
Linus Torvalds 已提交
2121 2122 2123
	kmem_cache_free(sn_cache, n);
}

A
Adrian Bunk 已提交
2124 2125
static struct sp_node *sp_alloc(unsigned long start, unsigned long end,
				struct mempolicy *pol)
L
Linus Torvalds 已提交
2126 2127 2128 2129 2130 2131 2132 2133
{
	struct sp_node *n = kmem_cache_alloc(sn_cache, GFP_KERNEL);

	if (!n)
		return NULL;
	n->start = start;
	n->end = end;
	mpol_get(pol);
2134
	pol->flags |= MPOL_F_SHARED;	/* for unref */
L
Linus Torvalds 已提交
2135 2136 2137 2138 2139 2140 2141 2142 2143 2144 2145 2146 2147 2148 2149 2150 2151 2152 2153 2154 2155 2156 2157 2158 2159 2160 2161 2162 2163 2164 2165 2166 2167 2168 2169 2170 2171 2172 2173 2174 2175 2176 2177 2178 2179 2180
	n->policy = pol;
	return n;
}

/* Replace a policy range. */
static int shared_policy_replace(struct shared_policy *sp, unsigned long start,
				 unsigned long end, struct sp_node *new)
{
	struct sp_node *n, *new2 = NULL;

restart:
	spin_lock(&sp->lock);
	n = sp_lookup(sp, start, end);
	/* Take care of old policies in the same range. */
	while (n && n->start < end) {
		struct rb_node *next = rb_next(&n->nd);
		if (n->start >= start) {
			if (n->end <= end)
				sp_delete(sp, n);
			else
				n->start = end;
		} else {
			/* Old policy spanning whole new range. */
			if (n->end > end) {
				if (!new2) {
					spin_unlock(&sp->lock);
					new2 = sp_alloc(end, n->end, n->policy);
					if (!new2)
						return -ENOMEM;
					goto restart;
				}
				n->end = start;
				sp_insert(sp, new2);
				new2 = NULL;
				break;
			} else
				n->end = start;
		}
		if (!next)
			break;
		n = rb_entry(next, struct sp_node, nd);
	}
	if (new)
		sp_insert(sp, new);
	spin_unlock(&sp->lock);
	if (new2) {
2181
		mpol_put(new2->policy);
L
Linus Torvalds 已提交
2182 2183 2184 2185 2186
		kmem_cache_free(sn_cache, new2);
	}
	return 0;
}

2187 2188 2189 2190 2191 2192 2193 2194
/**
 * mpol_shared_policy_init - initialize shared policy for inode
 * @sp: pointer to inode shared policy
 * @mpol:  struct mempolicy to install
 *
 * Install non-NULL @mpol in inode's shared policy rb-tree.
 * On entry, the current task has a reference on a non-NULL @mpol.
 * This must be released on exit.
2195
 * This is called at get_inode() calls and we can use GFP_KERNEL.
2196 2197 2198
 */
void mpol_shared_policy_init(struct shared_policy *sp, struct mempolicy *mpol)
{
2199 2200
	int ret;

2201 2202 2203 2204 2205 2206
	sp->root = RB_ROOT;		/* empty tree == default mempolicy */
	spin_lock_init(&sp->lock);

	if (mpol) {
		struct vm_area_struct pvma;
		struct mempolicy *new;
2207
		NODEMASK_SCRATCH(scratch);
2208

2209
		if (!scratch)
2210
			goto put_mpol;
2211 2212
		/* contextualize the tmpfs mount point mempolicy */
		new = mpol_new(mpol->mode, mpol->flags, &mpol->w.user_nodemask);
2213
		if (IS_ERR(new))
2214
			goto free_scratch; /* no valid nodemask intersection */
2215 2216

		task_lock(current);
2217
		ret = mpol_set_nodemask(new, &mpol->w.user_nodemask, scratch);
2218
		task_unlock(current);
2219
		if (ret)
2220
			goto put_new;
2221 2222 2223 2224 2225

		/* Create pseudo-vma that contains just the policy */
		memset(&pvma, 0, sizeof(struct vm_area_struct));
		pvma.vm_end = TASK_SIZE;	/* policy covers entire file */
		mpol_set_shared_policy(sp, &pvma, new); /* adds ref */
2226

2227
put_new:
2228
		mpol_put(new);			/* drop initial ref */
2229
free_scratch:
2230
		NODEMASK_SCRATCH_FREE(scratch);
2231 2232
put_mpol:
		mpol_put(mpol);	/* drop our incoming ref on sb mpol */
2233 2234 2235
	}
}

L
Linus Torvalds 已提交
2236 2237 2238 2239 2240 2241 2242
int mpol_set_shared_policy(struct shared_policy *info,
			struct vm_area_struct *vma, struct mempolicy *npol)
{
	int err;
	struct sp_node *new = NULL;
	unsigned long sz = vma_pages(vma);

2243
	pr_debug("set_shared_policy %lx sz %lu %d %d %lx\n",
L
Linus Torvalds 已提交
2244
		 vma->vm_pgoff,
2245
		 sz, npol ? npol->mode : -1,
2246
		 npol ? npol->flags : -1,
2247
		 npol ? nodes_addr(npol->v.nodes)[0] : -1);
L
Linus Torvalds 已提交
2248 2249 2250 2251 2252 2253 2254 2255 2256 2257 2258 2259 2260 2261 2262 2263 2264 2265 2266 2267 2268 2269 2270 2271 2272

	if (npol) {
		new = sp_alloc(vma->vm_pgoff, vma->vm_pgoff + sz, npol);
		if (!new)
			return -ENOMEM;
	}
	err = shared_policy_replace(info, vma->vm_pgoff, vma->vm_pgoff+sz, new);
	if (err && new)
		kmem_cache_free(sn_cache, new);
	return err;
}

/* Free a backing policy store on inode delete. */
void mpol_free_shared_policy(struct shared_policy *p)
{
	struct sp_node *n;
	struct rb_node *next;

	if (!p->root.rb_node)
		return;
	spin_lock(&p->lock);
	next = rb_first(&p->root);
	while (next) {
		n = rb_entry(next, struct sp_node, nd);
		next = rb_next(&n->nd);
2273
		rb_erase(&n->nd, &p->root);
2274
		mpol_put(n->policy);
L
Linus Torvalds 已提交
2275 2276 2277 2278 2279 2280 2281 2282
		kmem_cache_free(sn_cache, n);
	}
	spin_unlock(&p->lock);
}

/* assumes fs == KERNEL_DS */
void __init numa_policy_init(void)
{
2283 2284 2285 2286
	nodemask_t interleave_nodes;
	unsigned long largest = 0;
	int nid, prefer = 0;

L
Linus Torvalds 已提交
2287 2288
	policy_cache = kmem_cache_create("numa_policy",
					 sizeof(struct mempolicy),
2289
					 0, SLAB_PANIC, NULL);
L
Linus Torvalds 已提交
2290 2291 2292

	sn_cache = kmem_cache_create("shared_policy_node",
				     sizeof(struct sp_node),
2293
				     0, SLAB_PANIC, NULL);
L
Linus Torvalds 已提交
2294

2295 2296 2297 2298 2299 2300
	/*
	 * Set interleaving policy for system init. Interleaving is only
	 * enabled across suitably sized nodes (default is >= 16MB), or
	 * fall back to the largest node if they're all smaller.
	 */
	nodes_clear(interleave_nodes);
2301
	for_each_node_state(nid, N_HIGH_MEMORY) {
2302 2303 2304 2305 2306 2307 2308 2309 2310 2311 2312 2313 2314 2315 2316 2317
		unsigned long total_pages = node_present_pages(nid);

		/* Preserve the largest node */
		if (largest < total_pages) {
			largest = total_pages;
			prefer = nid;
		}

		/* Interleave this node? */
		if ((total_pages << PAGE_SHIFT) >= (16 << 20))
			node_set(nid, interleave_nodes);
	}

	/* All too small, use the largest */
	if (unlikely(nodes_empty(interleave_nodes)))
		node_set(prefer, interleave_nodes);
L
Linus Torvalds 已提交
2318

2319
	if (do_set_mempolicy(MPOL_INTERLEAVE, 0, &interleave_nodes))
L
Linus Torvalds 已提交
2320 2321 2322
		printk("numa_policy_init: interleaving failed\n");
}

2323
/* Reset policy of current process to default */
L
Linus Torvalds 已提交
2324 2325
void numa_default_policy(void)
{
2326
	do_set_mempolicy(MPOL_DEFAULT, 0, NULL);
L
Linus Torvalds 已提交
2327
}
2328

2329 2330 2331 2332
/*
 * Parse and format mempolicy from/to strings
 */

2333
/*
2334
 * "local" is pseudo-policy:  MPOL_PREFERRED with MPOL_F_LOCAL flag
2335
 * Used only for mpol_parse_str() and mpol_to_str()
2336
 */
2337 2338 2339 2340 2341 2342 2343 2344 2345
#define MPOL_LOCAL MPOL_MAX
static const char * const policy_modes[] =
{
	[MPOL_DEFAULT]    = "default",
	[MPOL_PREFERRED]  = "prefer",
	[MPOL_BIND]       = "bind",
	[MPOL_INTERLEAVE] = "interleave",
	[MPOL_LOCAL]      = "local"
};
2346

2347 2348 2349 2350 2351

#ifdef CONFIG_TMPFS
/**
 * mpol_parse_str - parse string to mempolicy
 * @str:  string containing mempolicy to parse
2352 2353
 * @mpol:  pointer to struct mempolicy pointer, returned on success.
 * @no_context:  flag whether to "contextualize" the mempolicy
2354 2355 2356 2357
 *
 * Format of input:
 *	<mode>[=<flags>][:<nodelist>]
 *
2358 2359 2360 2361 2362 2363 2364 2365
 * if @no_context is true, save the input nodemask in w.user_nodemask in
 * the returned mempolicy.  This will be used to "clone" the mempolicy in
 * a specific context [cpuset] at a later time.  Used to parse tmpfs mpol
 * mount option.  Note that if 'static' or 'relative' mode flags were
 * specified, the input nodemask will already have been saved.  Saving
 * it again is redundant, but safe.
 *
 * On success, returns 0, else 1
2366
 */
2367
int mpol_parse_str(char *str, struct mempolicy **mpol, int no_context)
2368
{
2369
	struct mempolicy *new = NULL;
2370
	unsigned short mode;
2371 2372
	unsigned short uninitialized_var(mode_flags);
	nodemask_t nodes;
2373 2374 2375 2376 2377 2378 2379
	char *nodelist = strchr(str, ':');
	char *flags = strchr(str, '=');
	int err = 1;

	if (nodelist) {
		/* NUL-terminate mode or flags string */
		*nodelist++ = '\0';
2380
		if (nodelist_parse(nodelist, nodes))
2381
			goto out;
2382
		if (!nodes_subset(nodes, node_states[N_HIGH_MEMORY]))
2383
			goto out;
2384 2385 2386
	} else
		nodes_clear(nodes);

2387 2388 2389
	if (flags)
		*flags++ = '\0';	/* terminate mode string */

2390
	for (mode = 0; mode <= MPOL_LOCAL; mode++) {
2391
		if (!strcmp(str, policy_modes[mode])) {
2392 2393 2394
			break;
		}
	}
2395
	if (mode > MPOL_LOCAL)
2396 2397
		goto out;

2398
	switch (mode) {
2399
	case MPOL_PREFERRED:
2400 2401 2402
		/*
		 * Insist on a nodelist of one node only
		 */
2403 2404 2405 2406
		if (nodelist) {
			char *rest = nodelist;
			while (isdigit(*rest))
				rest++;
2407 2408
			if (*rest)
				goto out;
2409 2410 2411 2412 2413 2414 2415
		}
		break;
	case MPOL_INTERLEAVE:
		/*
		 * Default to online nodes with memory if no nodelist
		 */
		if (!nodelist)
2416
			nodes = node_states[N_HIGH_MEMORY];
2417
		break;
2418
	case MPOL_LOCAL:
2419
		/*
2420
		 * Don't allow a nodelist;  mpol_new() checks flags
2421
		 */
2422
		if (nodelist)
2423
			goto out;
2424
		mode = MPOL_PREFERRED;
2425
		break;
2426 2427 2428 2429 2430 2431 2432
	case MPOL_DEFAULT:
		/*
		 * Insist on a empty nodelist
		 */
		if (!nodelist)
			err = 0;
		goto out;
2433 2434 2435 2436 2437 2438
	case MPOL_BIND:
		/*
		 * Insist on a nodelist
		 */
		if (!nodelist)
			goto out;
2439 2440
	}

2441
	mode_flags = 0;
2442 2443 2444 2445 2446 2447
	if (flags) {
		/*
		 * Currently, we only support two mutually exclusive
		 * mode flags.
		 */
		if (!strcmp(flags, "static"))
2448
			mode_flags |= MPOL_F_STATIC_NODES;
2449
		else if (!strcmp(flags, "relative"))
2450
			mode_flags |= MPOL_F_RELATIVE_NODES;
2451
		else
2452
			goto out;
2453
	}
2454 2455 2456

	new = mpol_new(mode, mode_flags, &nodes);
	if (IS_ERR(new))
2457 2458
		goto out;

2459 2460 2461 2462
	if (no_context) {
		/* save for contextualization */
		new->w.user_nodemask = nodes;
	} else {
2463
		int ret;
2464 2465 2466 2467 2468 2469 2470 2471 2472 2473
		NODEMASK_SCRATCH(scratch);
		if (scratch) {
			task_lock(current);
			ret = mpol_set_nodemask(new, &nodes, scratch);
			task_unlock(current);
		} else
			ret = -ENOMEM;
		NODEMASK_SCRATCH_FREE(scratch);
		if (ret) {
			mpol_put(new);
2474
			goto out;
2475 2476
		}
	}
2477
	err = 0;
2478

2479 2480 2481 2482 2483 2484
out:
	/* Restore string for error message */
	if (nodelist)
		*--nodelist = ':';
	if (flags)
		*--flags = '=';
2485 2486
	if (!err)
		*mpol = new;
2487 2488 2489 2490
	return err;
}
#endif /* CONFIG_TMPFS */

2491 2492 2493 2494 2495 2496 2497
/**
 * mpol_to_str - format a mempolicy structure for printing
 * @buffer:  to contain formatted mempolicy string
 * @maxlen:  length of @buffer
 * @pol:  pointer to mempolicy to be formatted
 * @no_context:  "context free" mempolicy - use nodemask in w.user_nodemask
 *
2498 2499 2500 2501
 * Convert a mempolicy into a string.
 * Returns the number of characters in buffer (if positive)
 * or an error (negative)
 */
2502
int mpol_to_str(char *buffer, int maxlen, struct mempolicy *pol, int no_context)
2503 2504 2505 2506
{
	char *p = buffer;
	int l;
	nodemask_t nodes;
2507
	unsigned short mode;
2508
	unsigned short flags = pol ? pol->flags : 0;
2509

2510 2511 2512 2513 2514
	/*
	 * Sanity check:  room for longest mode, flag and some nodes
	 */
	VM_BUG_ON(maxlen < strlen("interleave") + strlen("relative") + 16);

2515 2516 2517 2518 2519
	if (!pol || pol == &default_policy)
		mode = MPOL_DEFAULT;
	else
		mode = pol->mode;

2520 2521 2522 2523 2524 2525 2526
	switch (mode) {
	case MPOL_DEFAULT:
		nodes_clear(nodes);
		break;

	case MPOL_PREFERRED:
		nodes_clear(nodes);
2527
		if (flags & MPOL_F_LOCAL)
2528 2529
			mode = MPOL_LOCAL;	/* pseudo-policy */
		else
2530
			node_set(pol->v.preferred_node, nodes);
2531 2532 2533
		break;

	case MPOL_BIND:
2534
		/* Fall through */
2535
	case MPOL_INTERLEAVE:
2536 2537 2538 2539
		if (no_context)
			nodes = pol->w.user_nodemask;
		else
			nodes = pol->v.nodes;
2540 2541 2542 2543 2544 2545
		break;

	default:
		BUG();
	}

2546
	l = strlen(policy_modes[mode]);
2547 2548
	if (buffer + maxlen < p + l + 1)
		return -ENOSPC;
2549

2550
	strcpy(p, policy_modes[mode]);
2551 2552
	p += l;

2553
	if (flags & MPOL_MODE_FLAGS) {
2554 2555 2556 2557
		if (buffer + maxlen < p + 2)
			return -ENOSPC;
		*p++ = '=';

2558 2559 2560
		/*
		 * Currently, the only defined flags are mutually exclusive
		 */
2561
		if (flags & MPOL_F_STATIC_NODES)
2562 2563 2564
			p += snprintf(p, buffer + maxlen - p, "static");
		else if (flags & MPOL_F_RELATIVE_NODES)
			p += snprintf(p, buffer + maxlen - p, "relative");
2565 2566
	}

2567 2568 2569
	if (!nodes_empty(nodes)) {
		if (buffer + maxlen < p + 2)
			return -ENOSPC;
2570
		*p++ = ':';
2571 2572 2573 2574
	 	p += nodelist_scnprintf(p, buffer + maxlen - p, nodes);
	}
	return p - buffer;
}