mempolicy.c 51.8 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 78 79 80
 * 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.
   could replace all the switch()es with a mempolicy_ops structure.
*/

#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/gfp.h>
#include <linux/slab.h>
#include <linux/string.h>
#include <linux/module.h>
81
#include <linux/nsproxy.h>
L
Linus Torvalds 已提交
82 83 84
#include <linux/interrupt.h>
#include <linux/init.h>
#include <linux/compat.h>
85
#include <linux/swap.h>
86 87
#include <linux/seq_file.h>
#include <linux/proc_fs.h>
C
Christoph Lameter 已提交
88
#include <linux/migrate.h>
89
#include <linux/rmap.h>
90
#include <linux/security.h>
A
Adrian Bunk 已提交
91
#include <linux/syscalls.h>
92

L
Linus Torvalds 已提交
93 94 95
#include <asm/tlbflush.h>
#include <asm/uaccess.h>

96
/* Internal flags */
97
#define MPOL_MF_DISCONTIG_OK (MPOL_MF_INTERNAL << 0)	/* Skip checks for continuous vmas */
98
#define MPOL_MF_INVERT (MPOL_MF_INTERNAL << 1)		/* Invert check for nodemask */
99
#define MPOL_MF_STATS (MPOL_MF_INTERNAL << 2)		/* Gather statistics */
100

101 102
static struct kmem_cache *policy_cache;
static struct kmem_cache *sn_cache;
L
Linus Torvalds 已提交
103 104 105

/* Highest zone. An specific allocation for a zone below that is not
   policied. */
106
enum zone_type policy_zone = 0;
L
Linus Torvalds 已提交
107

108
struct mempolicy default_policy = {
L
Linus Torvalds 已提交
109 110 111 112
	.refcnt = ATOMIC_INIT(1), /* never free it */
	.policy = MPOL_DEFAULT,
};

A
Adrian Bunk 已提交
113 114 115
static void mpol_rebind_policy(struct mempolicy *pol,
                               const nodemask_t *newmask);

116 117
/* Check that the nodemask contains at least one populated zone */
static int is_valid_nodemask(nodemask_t *nodemask)
L
Linus Torvalds 已提交
118
{
119
	int nd, k;
L
Linus Torvalds 已提交
120

121 122 123 124 125 126 127 128 129 130
	/* Check that there is something useful in this mask */
	k = policy_zone;

	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;
131
		}
132
	}
133 134

	return 0;
L
Linus Torvalds 已提交
135 136
}

137 138 139 140 141
static inline int mpol_store_user_nodemask(const struct mempolicy *pol)
{
	return pol->flags & MPOL_F_STATIC_NODES;
}

L
Linus Torvalds 已提交
142
/* Create a new policy */
143 144
static struct mempolicy *mpol_new(unsigned short mode, unsigned short flags,
				  nodemask_t *nodes)
L
Linus Torvalds 已提交
145 146
{
	struct mempolicy *policy;
147
	nodemask_t cpuset_context_nmask;
L
Linus Torvalds 已提交
148

149 150
	pr_debug("setting mode %d flags %d nodes[0] %lx\n",
		 mode, flags, nodes ? nodes_addr(*nodes)[0] : -1);
151

L
Linus Torvalds 已提交
152
	if (mode == MPOL_DEFAULT)
153 154
		return (nodes && nodes_weight(*nodes)) ? ERR_PTR(-EINVAL) :
							 NULL;
L
Linus Torvalds 已提交
155 156 157 158
	policy = kmem_cache_alloc(policy_cache, GFP_KERNEL);
	if (!policy)
		return ERR_PTR(-ENOMEM);
	atomic_set(&policy->refcnt, 1);
159 160
	cpuset_update_task_memory_state();
	nodes_and(cpuset_context_nmask, *nodes, cpuset_current_mems_allowed);
L
Linus Torvalds 已提交
161 162
	switch (mode) {
	case MPOL_INTERLEAVE:
163 164 165
		if (nodes_empty(*nodes) || nodes_empty(cpuset_context_nmask))
			goto free;
		policy->v.nodes = cpuset_context_nmask;
L
Linus Torvalds 已提交
166 167
		break;
	case MPOL_PREFERRED:
168
		policy->v.preferred_node = first_node(cpuset_context_nmask);
L
Linus Torvalds 已提交
169
		if (policy->v.preferred_node >= MAX_NUMNODES)
170
			goto free;
L
Linus Torvalds 已提交
171 172
		break;
	case MPOL_BIND:
173 174 175
		if (!is_valid_nodemask(&cpuset_context_nmask))
			goto free;
		policy->v.nodes = cpuset_context_nmask;
L
Linus Torvalds 已提交
176
		break;
177 178
	default:
		BUG();
L
Linus Torvalds 已提交
179 180
	}
	policy->policy = mode;
181
	policy->flags = flags;
182 183 184 185
	if (mpol_store_user_nodemask(policy))
		policy->w.user_nodemask = *nodes;
	else
		policy->w.cpuset_mems_allowed = cpuset_mems_allowed(current);
L
Linus Torvalds 已提交
186
	return policy;
187 188 189 190

free:
	kmem_cache_free(policy_cache, policy);
	return ERR_PTR(-EINVAL);
L
Linus Torvalds 已提交
191 192
}

C
Christoph Lameter 已提交
193
static void gather_stats(struct page *, void *, int pte_dirty);
194 195
static void migrate_page_add(struct page *page, struct list_head *pagelist,
				unsigned long flags);
196

197
/* Scan through pages checking if pages follow certain conditions. */
N
Nick Piggin 已提交
198
static int check_pte_range(struct vm_area_struct *vma, pmd_t *pmd,
199 200
		unsigned long addr, unsigned long end,
		const nodemask_t *nodes, unsigned long flags,
201
		void *private)
L
Linus Torvalds 已提交
202
{
203 204
	pte_t *orig_pte;
	pte_t *pte;
205
	spinlock_t *ptl;
206

207
	orig_pte = pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
208
	do {
209
		struct page *page;
210
		int nid;
211 212

		if (!pte_present(*pte))
L
Linus Torvalds 已提交
213
			continue;
214 215
		page = vm_normal_page(vma, addr, *pte);
		if (!page)
L
Linus Torvalds 已提交
216
			continue;
217 218 219 220 221 222 223 224 225 226 227
		/*
		 * The check for PageReserved here is important to avoid
		 * handling zero pages and other pages that may have been
		 * marked special by the system.
		 *
		 * If the PageReserved would not be checked here then f.e.
		 * the location of the zero page could have an influence
		 * on MPOL_MF_STRICT, zero pages would be counted for
		 * the per node stats, and there would be useless attempts
		 * to put zero pages on the migration list.
		 */
228 229
		if (PageReserved(page))
			continue;
230
		nid = page_to_nid(page);
231 232 233
		if (node_isset(nid, *nodes) == !!(flags & MPOL_MF_INVERT))
			continue;

234
		if (flags & MPOL_MF_STATS)
C
Christoph Lameter 已提交
235
			gather_stats(page, private, pte_dirty(*pte));
236
		else if (flags & (MPOL_MF_MOVE | MPOL_MF_MOVE_ALL))
237
			migrate_page_add(page, private, flags);
238 239
		else
			break;
240
	} while (pte++, addr += PAGE_SIZE, addr != end);
241
	pte_unmap_unlock(orig_pte, ptl);
242 243 244
	return addr != end;
}

N
Nick Piggin 已提交
245
static inline int check_pmd_range(struct vm_area_struct *vma, pud_t *pud,
246 247
		unsigned long addr, unsigned long end,
		const nodemask_t *nodes, unsigned long flags,
248
		void *private)
249 250 251 252 253 254 255 256 257
{
	pmd_t *pmd;
	unsigned long next;

	pmd = pmd_offset(pud, addr);
	do {
		next = pmd_addr_end(addr, end);
		if (pmd_none_or_clear_bad(pmd))
			continue;
258
		if (check_pte_range(vma, pmd, addr, next, nodes,
259
				    flags, private))
260 261 262 263 264
			return -EIO;
	} while (pmd++, addr = next, addr != end);
	return 0;
}

N
Nick Piggin 已提交
265
static inline int check_pud_range(struct vm_area_struct *vma, pgd_t *pgd,
266 267
		unsigned long addr, unsigned long end,
		const nodemask_t *nodes, unsigned long flags,
268
		void *private)
269 270 271 272 273 274 275 276 277
{
	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;
278
		if (check_pmd_range(vma, pud, addr, next, nodes,
279
				    flags, private))
280 281 282 283 284
			return -EIO;
	} while (pud++, addr = next, addr != end);
	return 0;
}

N
Nick Piggin 已提交
285
static inline int check_pgd_range(struct vm_area_struct *vma,
286 287
		unsigned long addr, unsigned long end,
		const nodemask_t *nodes, unsigned long flags,
288
		void *private)
289 290 291 292
{
	pgd_t *pgd;
	unsigned long next;

N
Nick Piggin 已提交
293
	pgd = pgd_offset(vma->vm_mm, addr);
294 295 296 297
	do {
		next = pgd_addr_end(addr, end);
		if (pgd_none_or_clear_bad(pgd))
			continue;
298
		if (check_pud_range(vma, pgd, addr, next, nodes,
299
				    flags, private))
300 301 302
			return -EIO;
	} while (pgd++, addr = next, addr != end);
	return 0;
L
Linus Torvalds 已提交
303 304
}

305 306 307 308 309
/*
 * 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
310 311
static struct vm_area_struct *
check_range(struct mm_struct *mm, unsigned long start, unsigned long end,
312
		const nodemask_t *nodes, unsigned long flags, void *private)
L
Linus Torvalds 已提交
313 314 315 316
{
	int err;
	struct vm_area_struct *first, *vma, *prev;

317 318
	if (flags & (MPOL_MF_MOVE | MPOL_MF_MOVE_ALL)) {

C
Christoph Lameter 已提交
319 320 321
		err = migrate_prep();
		if (err)
			return ERR_PTR(err);
322
	}
323

L
Linus Torvalds 已提交
324 325 326 327 328
	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) {
329 330 331 332 333 334 335 336 337 338
		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
339
			unsigned long endvma = vma->vm_end;
340

A
Andi Kleen 已提交
341 342 343 344
			if (endvma > end)
				endvma = end;
			if (vma->vm_start > start)
				start = vma->vm_start;
345
			err = check_pgd_range(vma, start, endvma, nodes,
346
						flags, private);
L
Linus Torvalds 已提交
347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362
			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;

363
	pr_debug("vma %lx-%lx/%lx vm_ops %p vm_file %p set_policy %p\n",
L
Linus Torvalds 已提交
364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399
		 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;
		mpol_free(old);
	}
	return err;
}

/* Step 2: apply policy to a range and do splits. */
static int mbind_range(struct vm_area_struct *vma, unsigned long start,
		       unsigned long end, struct mempolicy *new)
{
	struct vm_area_struct *next;
	int err;

	err = 0;
	for (; vma && vma->vm_start < end; vma = next) {
		next = vma->vm_next;
		if (vma->vm_start < start)
			err = split_vma(vma->vm_mm, vma, start, 1);
		if (!err && vma->vm_end > end)
			err = split_vma(vma->vm_mm, vma, end, 0);
		if (!err)
			err = policy_vma(vma, new);
		if (err)
			break;
	}
	return err;
}

400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429
/*
 * 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
430
/* Set the process memory policy */
431 432
static long do_set_mempolicy(unsigned short mode, unsigned short flags,
			     nodemask_t *nodes)
L
Linus Torvalds 已提交
433 434 435
{
	struct mempolicy *new;

436
	new = mpol_new(mode, flags, nodes);
L
Linus Torvalds 已提交
437 438 439 440
	if (IS_ERR(new))
		return PTR_ERR(new);
	mpol_free(current->mempolicy);
	current->mempolicy = new;
441
	mpol_set_task_struct_flag();
442 443
	if (new && new->policy == MPOL_INTERLEAVE &&
	    nodes_weight(new->v.nodes))
444
		current->il_next = first_node(new->v.nodes);
L
Linus Torvalds 已提交
445 446 447 448
	return 0;
}

/* Fill a zone bitmap for a policy */
449
static void get_zonemask(struct mempolicy *p, nodemask_t *nodes)
L
Linus Torvalds 已提交
450
{
451
	nodes_clear(*nodes);
L
Linus Torvalds 已提交
452 453 454
	switch (p->policy) {
	case MPOL_DEFAULT:
		break;
455 456
	case MPOL_BIND:
		/* Fall through */
L
Linus Torvalds 已提交
457
	case MPOL_INTERLEAVE:
458
		*nodes = p->v.nodes;
L
Linus Torvalds 已提交
459 460
		break;
	case MPOL_PREFERRED:
461
		/* or use current node instead of memory_map? */
L
Linus Torvalds 已提交
462
		if (p->v.preferred_node < 0)
463
			*nodes = node_states[N_HIGH_MEMORY];
L
Linus Torvalds 已提交
464
		else
465
			node_set(p->v.preferred_node, *nodes);
L
Linus Torvalds 已提交
466 467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485
		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
486 487
static long do_get_mempolicy(int *policy, nodemask_t *nmask,
			     unsigned long addr, unsigned long flags)
L
Linus Torvalds 已提交
488
{
489
	int err;
L
Linus Torvalds 已提交
490 491 492 493
	struct mm_struct *mm = current->mm;
	struct vm_area_struct *vma = NULL;
	struct mempolicy *pol = current->mempolicy;

494
	cpuset_update_task_memory_state();
495 496
	if (flags &
		~(unsigned long)(MPOL_F_NODE|MPOL_F_ADDR|MPOL_F_MEMS_ALLOWED))
L
Linus Torvalds 已提交
497
		return -EINVAL;
498 499 500 501 502 503 504 505 506

	if (flags & MPOL_F_MEMS_ALLOWED) {
		if (flags & (MPOL_F_NODE|MPOL_F_ADDR))
			return -EINVAL;
		*policy = 0;	/* just so it's initialized */
		*nmask  = cpuset_current_mems_allowed;
		return 0;
	}

L
Linus Torvalds 已提交
507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525 526 527 528
	if (flags & MPOL_F_ADDR) {
		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)
		pol = &default_policy;

	if (flags & MPOL_F_NODE) {
		if (flags & MPOL_F_ADDR) {
			err = lookup_node(mm, addr);
			if (err < 0)
				goto out;
529
			*policy = err;
L
Linus Torvalds 已提交
530 531
		} else if (pol == current->mempolicy &&
				pol->policy == MPOL_INTERLEAVE) {
532
			*policy = current->il_next;
L
Linus Torvalds 已提交
533 534 535 536 537
		} else {
			err = -EINVAL;
			goto out;
		}
	} else
538
		*policy = pol->policy | pol->flags;
L
Linus Torvalds 已提交
539 540 541 542 543 544 545

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

	err = 0;
546 547
	if (nmask)
		get_zonemask(pol, nmask);
L
Linus Torvalds 已提交
548 549 550 551 552 553 554

 out:
	if (vma)
		up_read(&current->mm->mmap_sem);
	return err;
}

C
Christoph Lameter 已提交
555
#ifdef CONFIG_MIGRATION
556 557 558
/*
 * page migration
 */
559 560
static void migrate_page_add(struct page *page, struct list_head *pagelist,
				unsigned long flags)
561 562
{
	/*
563
	 * Avoid migrating a page that is shared with others.
564
	 */
C
Christoph Lameter 已提交
565 566
	if ((flags & MPOL_MF_MOVE_ALL) || page_mapcount(page) == 1)
		isolate_lru_page(page, pagelist);
567
}
568

569
static struct page *new_node_page(struct page *page, unsigned long node, int **x)
570
{
571
	return alloc_pages_node(node, GFP_HIGHUSER_MOVABLE, 0);
572 573
}

574 575 576 577
/*
 * Migrate pages from one node to a target node.
 * Returns error or the number of pages not migrated.
 */
A
Adrian Bunk 已提交
578 579
static int migrate_to_node(struct mm_struct *mm, int source, int dest,
			   int flags)
580 581 582 583 584 585 586
{
	nodemask_t nmask;
	LIST_HEAD(pagelist);
	int err = 0;

	nodes_clear(nmask);
	node_set(source, nmask);
587

588 589 590
	check_range(mm, mm->mmap->vm_start, TASK_SIZE, &nmask,
			flags | MPOL_MF_DISCONTIG_OK, &pagelist);

591
	if (!list_empty(&pagelist))
592 593
		err = migrate_pages(&pagelist, new_node_page, dest);

594
	return err;
595 596
}

597
/*
598 599
 * Move pages between the two nodesets so as to preserve the physical
 * layout as much as possible.
600 601 602 603 604 605 606
 *
 * 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)
{
	LIST_HEAD(pagelist);
607 608 609
	int busy = 0;
	int err = 0;
	nodemask_t tmp;
610

611
  	down_read(&mm->mmap_sem);
612

613 614 615 616
	err = migrate_vmas(mm, from_nodes, to_nodes, flags);
	if (err)
		goto out;

617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646
/*
 * 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.
 * Otherwise when we finish scannng from_tmp, we at least have the
 * 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.
 */
647

648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674
	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;
675
	}
676
out:
677
	up_read(&mm->mmap_sem);
678 679 680
	if (err < 0)
		return err;
	return busy;
C
Christoph Lameter 已提交
681 682 683

}

684 685 686 687 688 689 690
/*
 * 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.
 */
691
static struct page *new_vma_page(struct page *page, unsigned long private, int **x)
692 693
{
	struct vm_area_struct *vma = (struct vm_area_struct *)private;
694
	unsigned long uninitialized_var(address);
695

696 697 698 699 700 701 702 703 704 705 706
	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);
707
}
C
Christoph Lameter 已提交
708 709 710 711 712
#else

static void migrate_page_add(struct page *page, struct list_head *pagelist,
				unsigned long flags)
{
713 714
}

C
Christoph Lameter 已提交
715 716 717 718 719
int do_migrate_pages(struct mm_struct *mm,
	const nodemask_t *from_nodes, const nodemask_t *to_nodes, int flags)
{
	return -ENOSYS;
}
720

721
static struct page *new_vma_page(struct page *page, unsigned long private, int **x)
722 723 724
{
	return NULL;
}
C
Christoph Lameter 已提交
725 726
#endif

A
Adrian Bunk 已提交
727
static long do_mbind(unsigned long start, unsigned long len,
728 729
		     unsigned short mode, unsigned short mode_flags,
		     nodemask_t *nmask, unsigned long flags)
730 731 732 733 734 735 736 737
{
	struct vm_area_struct *vma;
	struct mm_struct *mm = current->mm;
	struct mempolicy *new;
	unsigned long end;
	int err;
	LIST_HEAD(pagelist);

738 739
	if (flags & ~(unsigned long)(MPOL_MF_STRICT |
				     MPOL_MF_MOVE | MPOL_MF_MOVE_ALL))
740
		return -EINVAL;
741
	if ((flags & MPOL_MF_MOVE_ALL) && !capable(CAP_SYS_NICE))
742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757
		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;

758
	new = mpol_new(mode, mode_flags, nmask);
759 760 761 762 763 764 765 766 767 768
	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;

769 770 771
	pr_debug("mbind %lx-%lx mode:%d flags:%d nodes:%lx\n",
		 start, start + len, mode, mode_flags,
		 nmask ? nodes_addr(*nmask)[0] : -1);
772 773 774 775 776 777 778 779 780 781

	down_write(&mm->mmap_sem);
	vma = check_range(mm, start, end, nmask,
			  flags | MPOL_MF_INVERT, &pagelist);

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

		err = mbind_range(vma, start, end, new);
782

783
		if (!list_empty(&pagelist))
784 785
			nr_failed = migrate_pages(&pagelist, new_vma_page,
						(unsigned long)vma);
786 787 788 789

		if (!err && nr_failed && (flags & MPOL_MF_STRICT))
			err = -EIO;
	}
C
Christoph Lameter 已提交
790

791 792 793 794 795
	up_write(&mm->mmap_sem);
	mpol_free(new);
	return err;
}

796 797 798 799 800
/*
 * User space interface with variable sized bitmaps for nodelists.
 */

/* Copy a node mask from user space. */
801
static int get_nodes(nodemask_t *nodes, const unsigned long __user *nmask,
802 803 804 805 806 807 808 809 810 811
		     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
812
	if (maxnode > PAGE_SIZE*BITS_PER_BYTE)
C
Chris Wright 已提交
813
		return -EINVAL;
814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869

	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;
}

asmlinkage long sys_mbind(unsigned long start, unsigned long len,
			unsigned long mode,
			unsigned long __user *nmask, unsigned long maxnode,
			unsigned flags)
{
	nodemask_t nodes;
	int err;
870
	unsigned short mode_flags;
871

872 873
	mode_flags = mode & MPOL_MODE_FLAGS;
	mode &= ~MPOL_MODE_FLAGS;
874 875
	if (mode >= MPOL_MAX)
		return -EINVAL;
876 877 878
	err = get_nodes(&nodes, nmask, maxnode);
	if (err)
		return err;
879
	return do_mbind(start, len, mode, mode_flags, &nodes, flags);
880 881 882 883 884 885 886 887
}

/* Set the process memory policy */
asmlinkage long sys_set_mempolicy(int mode, unsigned long __user *nmask,
		unsigned long maxnode)
{
	int err;
	nodemask_t nodes;
888
	unsigned short flags;
889

890 891 892
	flags = mode & MPOL_MODE_FLAGS;
	mode &= ~MPOL_MODE_FLAGS;
	if ((unsigned int)mode >= MPOL_MAX)
893 894 895 896
		return -EINVAL;
	err = get_nodes(&nodes, nmask, maxnode);
	if (err)
		return err;
897
	return do_set_mempolicy(mode, flags, &nodes);
898 899
}

900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920
asmlinkage long sys_migrate_pages(pid_t pid, unsigned long maxnode,
		const unsigned long __user *old_nodes,
		const unsigned long __user *new_nodes)
{
	struct mm_struct *mm;
	struct task_struct *task;
	nodemask_t old;
	nodemask_t new;
	nodemask_t task_nodes;
	int err;

	err = get_nodes(&old, old_nodes, maxnode);
	if (err)
		return err;

	err = get_nodes(&new, new_nodes, maxnode);
	if (err)
		return err;

	/* Find the mm_struct */
	read_lock(&tasklist_lock);
921
	task = pid ? find_task_by_vpid(pid) : current;
922 923 924 925 926 927 928 929 930 931 932 933 934
	if (!task) {
		read_unlock(&tasklist_lock);
		return -ESRCH;
	}
	mm = get_task_mm(task);
	read_unlock(&tasklist_lock);

	if (!mm)
		return -EINVAL;

	/*
	 * 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
935
	 * capabilities, superuser privileges or the same
936 937 938 939
	 * userid as the target process.
	 */
	if ((current->euid != task->suid) && (current->euid != task->uid) &&
	    (current->uid != task->suid) && (current->uid != task->uid) &&
940
	    !capable(CAP_SYS_NICE)) {
941 942 943 944 945 946
		err = -EPERM;
		goto out;
	}

	task_nodes = cpuset_mems_allowed(task);
	/* Is the user allowed to access the target nodes? */
947
	if (!nodes_subset(new, task_nodes) && !capable(CAP_SYS_NICE)) {
948 949 950 951
		err = -EPERM;
		goto out;
	}

952
	if (!nodes_subset(new, node_states[N_HIGH_MEMORY])) {
953 954 955 956
		err = -EINVAL;
		goto out;
	}

957 958 959 960
	err = security_task_movememory(task);
	if (err)
		goto out;

961
	err = do_migrate_pages(mm, &old, &new,
962
		capable(CAP_SYS_NICE) ? MPOL_MF_MOVE_ALL : MPOL_MF_MOVE);
963 964 965 966 967 968
out:
	mmput(mm);
	return err;
}


969 970 971 972 973 974
/* Retrieve NUMA policy */
asmlinkage long sys_get_mempolicy(int __user *policy,
				unsigned long __user *nmask,
				unsigned long maxnode,
				unsigned long addr, unsigned long flags)
{
A
Adrian Bunk 已提交
975 976
	int err;
	int uninitialized_var(pval);
977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993 994 995
	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
996 997 998 999 1000 1001 1002 1003 1004 1005 1006 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 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055
#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) {
		err = copy_from_user(bm, nm, alloc_size);
		/* 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;
1056
	nodemask_t bm;
L
Linus Torvalds 已提交
1057 1058 1059 1060 1061

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

	if (nmask) {
1062
		err = compat_get_bitmap(nodes_addr(bm), nmask, nr_bits);
L
Linus Torvalds 已提交
1063
		nm = compat_alloc_user_space(alloc_size);
1064
		err |= copy_to_user(nm, nodes_addr(bm), alloc_size);
L
Linus Torvalds 已提交
1065 1066 1067 1068 1069 1070 1071 1072 1073 1074
	}

	if (err)
		return -EFAULT;

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

#endif

1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087
/*
 * 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.
 * Returned policy has extra reference count if shared, vma,
 * or some other task's policy [show_numa_maps() can pass
 * @task != current].  It is the caller's responsibility to
 * free the reference in these cases.
 */
1088 1089
static struct mempolicy * get_vma_policy(struct task_struct *task,
		struct vm_area_struct *vma, unsigned long addr)
L
Linus Torvalds 已提交
1090
{
1091
	struct mempolicy *pol = task->mempolicy;
1092
	int shared_pol = 0;
L
Linus Torvalds 已提交
1093 1094

	if (vma) {
1095
		if (vma->vm_ops && vma->vm_ops->get_policy) {
1096
			pol = vma->vm_ops->get_policy(vma, addr);
1097 1098
			shared_pol = 1;	/* if pol non-NULL, add ref below */
		} else if (vma->vm_policy &&
L
Linus Torvalds 已提交
1099 1100 1101 1102 1103
				vma->vm_policy->policy != MPOL_DEFAULT)
			pol = vma->vm_policy;
	}
	if (!pol)
		pol = &default_policy;
1104 1105
	else if (!shared_pol && pol != current->mempolicy)
		mpol_get(pol);	/* vma or other task's policy */
L
Linus Torvalds 已提交
1106 1107 1108
	return pol;
}

1109 1110 1111 1112 1113 1114 1115 1116 1117 1118 1119 1120
/* Return a nodemask representing a mempolicy */
static nodemask_t *nodemask_policy(gfp_t gfp, struct mempolicy *policy)
{
	/* Lower zones don't get a nodemask applied for MPOL_BIND */
	if (unlikely(policy->policy == MPOL_BIND) &&
			gfp_zone(gfp) >= policy_zone &&
			cpuset_nodemask_valid_mems_allowed(&policy->v.nodes))
		return &policy->v.nodes;

	return NULL;
}

L
Linus Torvalds 已提交
1121
/* Return a zonelist representing a mempolicy */
A
Al Viro 已提交
1122
static struct zonelist *zonelist_policy(gfp_t gfp, struct mempolicy *policy)
L
Linus Torvalds 已提交
1123 1124 1125 1126 1127 1128 1129 1130 1131 1132
{
	int nd;

	switch (policy->policy) {
	case MPOL_PREFERRED:
		nd = policy->v.preferred_node;
		if (nd < 0)
			nd = numa_node_id();
		break;
	case MPOL_BIND:
1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143
		/*
		 * Normally, MPOL_BIND allocations node-local are node-local
		 * within the allowed nodemask. However, if __GFP_THISNODE is
		 * set and the current node is part of the mask, we use the
		 * the zonelist for the first node in the mask instead.
		 */
		nd = numa_node_id();
		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
1144 1145 1146 1147 1148 1149 1150 1151
	case MPOL_INTERLEAVE: /* should not happen */
	case MPOL_DEFAULT:
		nd = numa_node_id();
		break;
	default:
		nd = 0;
		BUG();
	}
1152
	return node_zonelist(nd, gfp);
L
Linus Torvalds 已提交
1153 1154 1155 1156 1157 1158 1159 1160 1161
}

/* 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;
1162
	next = next_node(nid, policy->v.nodes);
L
Linus Torvalds 已提交
1163
	if (next >= MAX_NUMNODES)
1164
		next = first_node(policy->v.nodes);
1165 1166
	if (next < MAX_NUMNODES)
		me->il_next = next;
L
Linus Torvalds 已提交
1167 1168 1169
	return nid;
}

1170 1171 1172 1173 1174 1175
/*
 * Depending on the memory policy provide a node from which to allocate the
 * next slab entry.
 */
unsigned slab_node(struct mempolicy *policy)
{
1176
	unsigned short pol = policy ? policy->policy : MPOL_DEFAULT;
1177 1178

	switch (pol) {
1179 1180 1181
	case MPOL_INTERLEAVE:
		return interleave_nodes(policy);

1182
	case MPOL_BIND: {
1183 1184 1185 1186
		/*
		 * Follow bind policy behavior and start allocation at the
		 * first node.
		 */
1187 1188 1189 1190 1191 1192 1193 1194
		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);
		return zone->node;
1195
	}
1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206

	case MPOL_PREFERRED:
		if (policy->v.preferred_node >= 0)
			return policy->v.preferred_node;
		/* Fall through */

	default:
		return numa_node_id();
	}
}

L
Linus Torvalds 已提交
1207 1208 1209 1210
/* 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)
{
1211
	unsigned nnodes = nodes_weight(pol->v.nodes);
1212
	unsigned target;
L
Linus Torvalds 已提交
1213 1214 1215
	int c;
	int nid = -1;

1216 1217 1218
	if (!nnodes)
		return numa_node_id();
	target = (unsigned int)off % nnodes;
L
Linus Torvalds 已提交
1219 1220
	c = 0;
	do {
1221
		nid = next_node(nid, pol->v.nodes);
L
Linus Torvalds 已提交
1222 1223 1224 1225 1226
		c++;
	} while (c <= target);
	return nid;
}

1227 1228 1229 1230 1231 1232 1233
/* 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;

1234 1235 1236 1237 1238 1239 1240 1241 1242
		/*
		 * 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);
1243 1244 1245 1246 1247 1248
		off += (addr - vma->vm_start) >> shift;
		return offset_il_node(pol, vma, off);
	} else
		return interleave_nodes(pol);
}

1249
#ifdef CONFIG_HUGETLBFS
1250 1251 1252 1253 1254
/*
 * 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
1255 1256
 * @mpol = pointer to mempolicy pointer for reference counted mempolicy
 * @nodemask = pointer to nodemask pointer for MPOL_BIND nodemask
1257 1258
 *
 * Returns a zonelist suitable for a huge page allocation.
1259 1260
 * If the effective policy is 'BIND, returns pointer to local node's zonelist,
 * and a pointer to the mempolicy's @nodemask for filtering the zonelist.
1261
 * If it is also a policy for which get_vma_policy() returns an extra
1262
 * reference, we must hold that reference until after the allocation.
1263
 * In that case, return policy via @mpol so hugetlb allocation can drop
1264
 * the reference. For non-'BIND referenced policies, we can/do drop the
1265 1266 1267
 * reference here, so the caller doesn't need to know about the special case
 * for default and current task policy.
 */
1268
struct zonelist *huge_zonelist(struct vm_area_struct *vma, unsigned long addr,
1269 1270
				gfp_t gfp_flags, struct mempolicy **mpol,
				nodemask_t **nodemask)
1271 1272
{
	struct mempolicy *pol = get_vma_policy(current, vma, addr);
1273
	struct zonelist *zl;
1274

1275
	*mpol = NULL;		/* probably no unref needed */
1276 1277 1278 1279
	*nodemask = NULL;	/* assume !MPOL_BIND */
	if (pol->policy == MPOL_BIND) {
			*nodemask = &pol->v.nodes;
	} else if (pol->policy == MPOL_INTERLEAVE) {
1280 1281 1282
		unsigned nid;

		nid = interleave_nid(pol, vma, addr, HPAGE_SHIFT);
1283 1284 1285
		if (unlikely(pol != &default_policy &&
				pol != current->mempolicy))
			__mpol_free(pol);	/* finished with pol */
1286
		return node_zonelist(nid, gfp_flags);
1287
	}
1288 1289 1290 1291 1292 1293 1294 1295 1296

	zl = zonelist_policy(GFP_HIGHUSER, pol);
	if (unlikely(pol != &default_policy && pol != current->mempolicy)) {
		if (pol->policy != MPOL_BIND)
			__mpol_free(pol);	/* finished with pol */
		else
			*mpol = pol;	/* unref needed after allocation */
	}
	return zl;
1297
}
1298
#endif
1299

L
Linus Torvalds 已提交
1300 1301
/* Allocate a page in interleaved policy.
   Own path because it needs to do special accounting. */
1302 1303
static struct page *alloc_page_interleave(gfp_t gfp, unsigned order,
					unsigned nid)
L
Linus Torvalds 已提交
1304 1305 1306 1307
{
	struct zonelist *zl;
	struct page *page;

1308
	zl = node_zonelist(nid, gfp);
L
Linus Torvalds 已提交
1309
	page = __alloc_pages(gfp, order, zl);
1310
	if (page && page_zone(page) == zonelist_zone(&zl->_zonerefs[0]))
1311
		inc_zone_page_state(page, NUMA_INTERLEAVE_HIT);
L
Linus Torvalds 已提交
1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322 1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337
	return page;
}

/**
 * 	alloc_page_vma	- Allocate a page for a VMA.
 *
 * 	@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.
 *
 * 	@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 *
A
Al Viro 已提交
1338
alloc_page_vma(gfp_t gfp, struct vm_area_struct *vma, unsigned long addr)
L
Linus Torvalds 已提交
1339
{
1340
	struct mempolicy *pol = get_vma_policy(current, vma, addr);
1341
	struct zonelist *zl;
L
Linus Torvalds 已提交
1342

1343
	cpuset_update_task_memory_state();
L
Linus Torvalds 已提交
1344 1345 1346

	if (unlikely(pol->policy == MPOL_INTERLEAVE)) {
		unsigned nid;
1347 1348

		nid = interleave_nid(pol, vma, addr, PAGE_SHIFT);
1349 1350 1351
		if (unlikely(pol != &default_policy &&
				pol != current->mempolicy))
			__mpol_free(pol);	/* finished with pol */
L
Linus Torvalds 已提交
1352 1353
		return alloc_page_interleave(gfp, 0, nid);
	}
1354 1355 1356 1357 1358
	zl = zonelist_policy(gfp, pol);
	if (pol != &default_policy && pol != current->mempolicy) {
		/*
		 * slow path: ref counted policy -- shared or vma
		 */
1359 1360
		struct page *page =  __alloc_pages_nodemask(gfp, 0,
						zl, nodemask_policy(gfp, pol));
1361 1362 1363 1364 1365 1366
		__mpol_free(pol);
		return page;
	}
	/*
	 * fast path:  default or task policy
	 */
1367
	return __alloc_pages_nodemask(gfp, 0, zl, nodemask_policy(gfp, pol));
L
Linus Torvalds 已提交
1368 1369 1370 1371 1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384
}

/**
 * 	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.
 *
1385
 *	Don't call cpuset_update_task_memory_state() unless
L
Linus Torvalds 已提交
1386 1387 1388
 *	1) it's ok to take cpuset_sem (can WAIT), and
 *	2) allocating for current task (not interrupt).
 */
A
Al Viro 已提交
1389
struct page *alloc_pages_current(gfp_t gfp, unsigned order)
L
Linus Torvalds 已提交
1390 1391 1392 1393
{
	struct mempolicy *pol = current->mempolicy;

	if ((gfp & __GFP_WAIT) && !in_interrupt())
1394
		cpuset_update_task_memory_state();
1395
	if (!pol || in_interrupt() || (gfp & __GFP_THISNODE))
L
Linus Torvalds 已提交
1396 1397 1398
		pol = &default_policy;
	if (pol->policy == MPOL_INTERLEAVE)
		return alloc_page_interleave(gfp, order, interleave_nodes(pol));
1399 1400
	return __alloc_pages_nodemask(gfp, order,
			zonelist_policy(gfp, pol), nodemask_policy(gfp, pol));
L
Linus Torvalds 已提交
1401 1402 1403
}
EXPORT_SYMBOL(alloc_pages_current);

1404 1405 1406 1407 1408 1409 1410 1411
/*
 * If mpol_copy() sees current->cpuset == cpuset_being_rebound, then it
 * 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().
 */

L
Linus Torvalds 已提交
1412 1413 1414 1415 1416 1417 1418
/* Slow path of a mempolicy copy */
struct mempolicy *__mpol_copy(struct mempolicy *old)
{
	struct mempolicy *new = kmem_cache_alloc(policy_cache, GFP_KERNEL);

	if (!new)
		return ERR_PTR(-ENOMEM);
1419 1420 1421 1422
	if (current_cpuset_is_being_rebound()) {
		nodemask_t mems = cpuset_mems_allowed(current);
		mpol_rebind_policy(old, &mems);
	}
L
Linus Torvalds 已提交
1423 1424 1425 1426 1427
	*new = *old;
	atomic_set(&new->refcnt, 1);
	return new;
}

1428 1429 1430 1431 1432 1433 1434 1435 1436 1437
static int mpol_match_intent(const struct mempolicy *a,
			     const struct mempolicy *b)
{
	if (a->flags != b->flags)
		return 0;
	if (!mpol_store_user_nodemask(a))
		return 1;
	return nodes_equal(a->w.user_nodemask, b->w.user_nodemask);
}

L
Linus Torvalds 已提交
1438 1439 1440 1441 1442 1443 1444
/* Slow path of a mempolicy comparison */
int __mpol_equal(struct mempolicy *a, struct mempolicy *b)
{
	if (!a || !b)
		return 0;
	if (a->policy != b->policy)
		return 0;
1445 1446
	if (a->policy != MPOL_DEFAULT && !mpol_match_intent(a, b))
		return 0;
L
Linus Torvalds 已提交
1447 1448 1449
	switch (a->policy) {
	case MPOL_DEFAULT:
		return 1;
1450 1451
	case MPOL_BIND:
		/* Fall through */
L
Linus Torvalds 已提交
1452
	case MPOL_INTERLEAVE:
1453
		return nodes_equal(a->v.nodes, b->v.nodes);
L
Linus Torvalds 已提交
1454 1455 1456 1457 1458 1459 1460 1461 1462 1463 1464 1465 1466 1467 1468 1469 1470 1471 1472 1473 1474 1475 1476 1477 1478 1479 1480 1481 1482 1483 1484 1485 1486 1487 1488 1489 1490 1491 1492 1493 1494 1495 1496 1497 1498 1499 1500 1501 1502 1503 1504 1505 1506 1507 1508 1509 1510 1511 1512 1513 1514 1515 1516 1517 1518 1519 1520 1521 1522 1523 1524 1525 1526 1527 1528 1529 1530 1531
	case MPOL_PREFERRED:
		return a->v.preferred_node == b->v.preferred_node;
	default:
		BUG();
		return 0;
	}
}

/* Slow path of a mpol destructor. */
void __mpol_free(struct mempolicy *p)
{
	if (!atomic_dec_and_test(&p->refcnt))
		return;
	p->policy = MPOL_DEFAULT;
	kmem_cache_free(policy_cache, p);
}

/*
 * 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);
1532
	pr_debug("inserting %lx-%lx: %d\n", new->start, new->end,
L
Linus Torvalds 已提交
1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548 1549 1550 1551 1552 1553 1554 1555 1556
		 new->policy ? new->policy->policy : 0);
}

/* 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)
{
1557
	pr_debug("deleting %lx-l%lx\n", n->start, n->end);
L
Linus Torvalds 已提交
1558 1559 1560 1561 1562
	rb_erase(&n->nd, &sp->root);
	mpol_free(n->policy);
	kmem_cache_free(sn_cache, n);
}

A
Adrian Bunk 已提交
1563 1564
static struct sp_node *sp_alloc(unsigned long start, unsigned long end,
				struct mempolicy *pol)
L
Linus Torvalds 已提交
1565 1566 1567 1568 1569 1570 1571 1572 1573 1574 1575 1576 1577 1578 1579 1580 1581 1582 1583 1584 1585 1586 1587 1588 1589 1590 1591 1592 1593 1594 1595 1596 1597 1598 1599 1600 1601 1602 1603 1604 1605 1606 1607 1608 1609 1610 1611 1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624
{
	struct sp_node *n = kmem_cache_alloc(sn_cache, GFP_KERNEL);

	if (!n)
		return NULL;
	n->start = start;
	n->end = end;
	mpol_get(pol);
	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) {
		mpol_free(new2->policy);
		kmem_cache_free(sn_cache, new2);
	}
	return 0;
}

1625
void mpol_shared_policy_init(struct shared_policy *info, unsigned short policy,
1626
			unsigned short flags, nodemask_t *policy_nodes)
1627 1628 1629 1630 1631 1632 1633 1634
{
	info->root = RB_ROOT;
	spin_lock_init(&info->lock);

	if (policy != MPOL_DEFAULT) {
		struct mempolicy *newpol;

		/* Falls back to MPOL_DEFAULT on any error */
1635
		newpol = mpol_new(policy, flags, policy_nodes);
1636 1637 1638 1639 1640 1641 1642 1643 1644 1645 1646 1647 1648
		if (!IS_ERR(newpol)) {
			/* Create pseudo-vma that contains just the policy */
			struct vm_area_struct pvma;

			memset(&pvma, 0, sizeof(struct vm_area_struct));
			/* Policy covers entire file */
			pvma.vm_end = TASK_SIZE;
			mpol_set_shared_policy(info, &pvma, newpol);
			mpol_free(newpol);
		}
	}
}

L
Linus Torvalds 已提交
1649 1650 1651 1652 1653 1654 1655
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);

1656
	pr_debug("set_shared_policy %lx sz %lu %d %d %lx\n",
L
Linus Torvalds 已提交
1657
		 vma->vm_pgoff,
1658 1659
		 sz, npol ? npol->policy : -1,
		 npol ? npol->flags : -1,
1660
		 npol ? nodes_addr(npol->v.nodes)[0] : -1);
L
Linus Torvalds 已提交
1661 1662 1663 1664 1665 1666 1667 1668 1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683 1684 1685

	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);
1686
		rb_erase(&n->nd, &p->root);
L
Linus Torvalds 已提交
1687 1688 1689 1690 1691 1692 1693 1694 1695
		mpol_free(n->policy);
		kmem_cache_free(sn_cache, n);
	}
	spin_unlock(&p->lock);
}

/* assumes fs == KERNEL_DS */
void __init numa_policy_init(void)
{
1696 1697 1698 1699
	nodemask_t interleave_nodes;
	unsigned long largest = 0;
	int nid, prefer = 0;

L
Linus Torvalds 已提交
1700 1701
	policy_cache = kmem_cache_create("numa_policy",
					 sizeof(struct mempolicy),
1702
					 0, SLAB_PANIC, NULL);
L
Linus Torvalds 已提交
1703 1704 1705

	sn_cache = kmem_cache_create("shared_policy_node",
				     sizeof(struct sp_node),
1706
				     0, SLAB_PANIC, NULL);
L
Linus Torvalds 已提交
1707

1708 1709 1710 1711 1712 1713
	/*
	 * 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);
1714
	for_each_node_state(nid, N_HIGH_MEMORY) {
1715 1716 1717 1718 1719 1720 1721 1722 1723 1724 1725 1726 1727 1728 1729 1730
		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
1731

1732
	if (do_set_mempolicy(MPOL_INTERLEAVE, 0, &interleave_nodes))
L
Linus Torvalds 已提交
1733 1734 1735
		printk("numa_policy_init: interleaving failed\n");
}

1736
/* Reset policy of current process to default */
L
Linus Torvalds 已提交
1737 1738
void numa_default_policy(void)
{
1739
	do_set_mempolicy(MPOL_DEFAULT, 0, NULL);
L
Linus Torvalds 已提交
1740
}
1741 1742

/* Migrate a policy to a different set of nodes */
A
Adrian Bunk 已提交
1743 1744
static void mpol_rebind_policy(struct mempolicy *pol,
			       const nodemask_t *newmask)
1745 1746
{
	nodemask_t tmp;
1747
	int static_nodes;
1748 1749 1750

	if (!pol)
		return;
1751 1752 1753
	static_nodes = pol->flags & MPOL_F_STATIC_NODES;
	if (!mpol_store_user_nodemask(pol) &&
	    nodes_equal(pol->w.cpuset_mems_allowed, *newmask))
1754
		return;
1755 1756 1757 1758

	switch (pol->policy) {
	case MPOL_DEFAULT:
		break;
1759 1760
	case MPOL_BIND:
		/* Fall through */
1761
	case MPOL_INTERLEAVE:
1762 1763 1764 1765 1766 1767 1768
		if (static_nodes)
			nodes_and(tmp, pol->w.user_nodemask, *newmask);
		else {
			nodes_remap(tmp, pol->v.nodes,
				    pol->w.cpuset_mems_allowed, *newmask);
			pol->w.cpuset_mems_allowed = *newmask;
		}
1769
		pol->v.nodes = tmp;
1770 1771 1772 1773 1774 1775 1776
		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();
		}
1777 1778
		break;
	case MPOL_PREFERRED:
1779 1780 1781 1782 1783 1784 1785 1786 1787 1788 1789 1790
		if (static_nodes) {
			int node = first_node(pol->w.user_nodemask);

			if (node_isset(node, *newmask))
				pol->v.preferred_node = node;
			else
				pol->v.preferred_node = -1;
		} else {
			pol->v.preferred_node = node_remap(pol->v.preferred_node,
					pol->w.cpuset_mems_allowed, *newmask);
			pol->w.cpuset_mems_allowed = *newmask;
		}
1791 1792 1793 1794 1795 1796 1797 1798
		break;
	default:
		BUG();
		break;
	}
}

/*
1799 1800
 * Wrapper for mpol_rebind_policy() that just requires task
 * pointer, and updates task mempolicy.
1801
 */
1802 1803

void mpol_rebind_task(struct task_struct *tsk, const nodemask_t *new)
1804
{
1805
	mpol_rebind_policy(tsk->mempolicy, new);
1806
}
1807

1808 1809 1810 1811 1812 1813 1814 1815 1816 1817 1818 1819 1820 1821 1822 1823
/*
 * 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)
		mpol_rebind_policy(vma->vm_policy, new);
	up_write(&mm->mmap_sem);
}

1824 1825 1826 1827
/*
 * Display pages allocated per node and memory policy via /proc.
 */

1828 1829
static const char * const policy_types[] =
	{ "default", "prefer", "bind", "interleave" };
1830 1831 1832 1833 1834 1835 1836 1837 1838 1839 1840

/*
 * Convert a mempolicy into a string.
 * Returns the number of characters in buffer (if positive)
 * or an error (negative)
 */
static inline int mpol_to_str(char *buffer, int maxlen, struct mempolicy *pol)
{
	char *p = buffer;
	int l;
	nodemask_t nodes;
1841
	unsigned short mode = pol ? pol->policy : MPOL_DEFAULT;
1842
	unsigned short flags = pol ? pol->flags : 0;
1843 1844 1845 1846 1847 1848 1849 1850 1851 1852 1853 1854

	switch (mode) {
	case MPOL_DEFAULT:
		nodes_clear(nodes);
		break;

	case MPOL_PREFERRED:
		nodes_clear(nodes);
		node_set(pol->v.preferred_node, nodes);
		break;

	case MPOL_BIND:
1855
		/* Fall through */
1856 1857 1858 1859 1860 1861 1862 1863 1864 1865 1866 1867 1868 1869 1870 1871
	case MPOL_INTERLEAVE:
		nodes = pol->v.nodes;
		break;

	default:
		BUG();
		return -EFAULT;
	}

	l = strlen(policy_types[mode]);
 	if (buffer + maxlen < p + l + 1)
 		return -ENOSPC;

	strcpy(p, policy_types[mode]);
	p += l;

1872 1873 1874 1875 1876 1877 1878 1879 1880 1881 1882
	if (flags) {
		int need_bar = 0;

		if (buffer + maxlen < p + 2)
			return -ENOSPC;
		*p++ = '=';

		if (flags & MPOL_F_STATIC_NODES)
			p += sprintf(p, "%sstatic", need_bar++ ? "|" : "");
	}

1883 1884 1885 1886 1887 1888 1889 1890 1891 1892 1893 1894
	if (!nodes_empty(nodes)) {
		if (buffer + maxlen < p + 2)
			return -ENOSPC;
		*p++ = '=';
	 	p += nodelist_scnprintf(p, buffer + maxlen - p, nodes);
	}
	return p - buffer;
}

struct numa_maps {
	unsigned long pages;
	unsigned long anon;
C
Christoph Lameter 已提交
1895 1896
	unsigned long active;
	unsigned long writeback;
1897
	unsigned long mapcount_max;
C
Christoph Lameter 已提交
1898 1899
	unsigned long dirty;
	unsigned long swapcache;
1900 1901 1902
	unsigned long node[MAX_NUMNODES];
};

C
Christoph Lameter 已提交
1903
static void gather_stats(struct page *page, void *private, int pte_dirty)
1904 1905 1906 1907
{
	struct numa_maps *md = private;
	int count = page_mapcount(page);

C
Christoph Lameter 已提交
1908 1909 1910
	md->pages++;
	if (pte_dirty || PageDirty(page))
		md->dirty++;
1911

C
Christoph Lameter 已提交
1912 1913
	if (PageSwapCache(page))
		md->swapcache++;
1914

C
Christoph Lameter 已提交
1915 1916 1917 1918 1919
	if (PageActive(page))
		md->active++;

	if (PageWriteback(page))
		md->writeback++;
1920 1921 1922 1923

	if (PageAnon(page))
		md->anon++;

C
Christoph Lameter 已提交
1924 1925 1926
	if (count > md->mapcount_max)
		md->mapcount_max = count;

1927 1928 1929
	md->node[page_to_nid(page)]++;
}

A
Andrew Morton 已提交
1930
#ifdef CONFIG_HUGETLB_PAGE
C
Christoph Lameter 已提交
1931 1932 1933 1934 1935 1936 1937 1938 1939 1940 1941 1942 1943 1944 1945 1946 1947 1948 1949 1950 1951 1952 1953 1954 1955
static void check_huge_range(struct vm_area_struct *vma,
		unsigned long start, unsigned long end,
		struct numa_maps *md)
{
	unsigned long addr;
	struct page *page;

	for (addr = start; addr < end; addr += HPAGE_SIZE) {
		pte_t *ptep = huge_pte_offset(vma->vm_mm, addr & HPAGE_MASK);
		pte_t pte;

		if (!ptep)
			continue;

		pte = *ptep;
		if (pte_none(pte))
			continue;

		page = pte_page(pte);
		if (!page)
			continue;

		gather_stats(page, md, pte_dirty(*ptep));
	}
}
A
Andrew Morton 已提交
1956 1957 1958 1959 1960 1961 1962
#else
static inline void check_huge_range(struct vm_area_struct *vma,
		unsigned long start, unsigned long end,
		struct numa_maps *md)
{
}
#endif
C
Christoph Lameter 已提交
1963

1964 1965
int show_numa_map(struct seq_file *m, void *v)
{
1966
	struct proc_maps_private *priv = m->private;
1967 1968
	struct vm_area_struct *vma = v;
	struct numa_maps *md;
C
Christoph Lameter 已提交
1969 1970
	struct file *file = vma->vm_file;
	struct mm_struct *mm = vma->vm_mm;
1971
	struct mempolicy *pol;
1972 1973 1974
	int n;
	char buffer[50];

C
Christoph Lameter 已提交
1975
	if (!mm)
1976 1977 1978 1979 1980 1981
		return 0;

	md = kzalloc(sizeof(struct numa_maps), GFP_KERNEL);
	if (!md)
		return 0;

1982 1983 1984 1985 1986 1987 1988
	pol = get_vma_policy(priv->task, vma, vma->vm_start);
	mpol_to_str(buffer, sizeof(buffer), pol);
	/*
	 * unref shared or other task's mempolicy
	 */
	if (pol != &default_policy && pol != current->mempolicy)
		__mpol_free(pol);
C
Christoph Lameter 已提交
1989 1990 1991 1992 1993

	seq_printf(m, "%08lx %s", vma->vm_start, buffer);

	if (file) {
		seq_printf(m, " file=");
1994
		seq_path(m, &file->f_path, "\n\t= ");
C
Christoph Lameter 已提交
1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 2005
	} else if (vma->vm_start <= mm->brk && vma->vm_end >= mm->start_brk) {
		seq_printf(m, " heap");
	} else if (vma->vm_start <= mm->start_stack &&
			vma->vm_end >= mm->start_stack) {
		seq_printf(m, " stack");
	}

	if (is_vm_hugetlb_page(vma)) {
		check_huge_range(vma, vma->vm_start, vma->vm_end, md);
		seq_printf(m, " huge");
	} else {
2006
		check_pgd_range(vma, vma->vm_start, vma->vm_end,
2007
			&node_states[N_HIGH_MEMORY], MPOL_MF_STATS, md);
C
Christoph Lameter 已提交
2008 2009 2010 2011
	}

	if (!md->pages)
		goto out;
2012

C
Christoph Lameter 已提交
2013 2014
	if (md->anon)
		seq_printf(m," anon=%lu",md->anon);
2015

C
Christoph Lameter 已提交
2016 2017
	if (md->dirty)
		seq_printf(m," dirty=%lu",md->dirty);
2018

C
Christoph Lameter 已提交
2019 2020
	if (md->pages != md->anon && md->pages != md->dirty)
		seq_printf(m, " mapped=%lu", md->pages);
2021

C
Christoph Lameter 已提交
2022 2023
	if (md->mapcount_max > 1)
		seq_printf(m, " mapmax=%lu", md->mapcount_max);
2024

C
Christoph Lameter 已提交
2025 2026 2027 2028 2029 2030 2031 2032 2033
	if (md->swapcache)
		seq_printf(m," swapcache=%lu", md->swapcache);

	if (md->active < md->pages && !is_vm_hugetlb_page(vma))
		seq_printf(m," active=%lu", md->active);

	if (md->writeback)
		seq_printf(m," writeback=%lu", md->writeback);

2034
	for_each_node_state(n, N_HIGH_MEMORY)
C
Christoph Lameter 已提交
2035 2036 2037 2038
		if (md->node[n])
			seq_printf(m, " N%d=%lu", n, md->node[n]);
out:
	seq_putc(m, '\n');
2039 2040 2041
	kfree(md);

	if (m->count < m->size)
2042
		m->version = (vma != priv->tail_vma) ? vma->vm_start : 0;
2043 2044
	return 0;
}