ksm.c 45.1 KB
Newer Older
H
Hugh Dickins 已提交
1
/*
I
Izik Eidus 已提交
2 3 4 5 6
 * Memory merging support.
 *
 * This code enables dynamic sharing of identical pages found in different
 * memory areas, even if they are not shared by fork()
 *
I
Izik Eidus 已提交
7
 * Copyright (C) 2008-2009 Red Hat, Inc.
I
Izik Eidus 已提交
8 9 10 11
 * Authors:
 *	Izik Eidus
 *	Andrea Arcangeli
 *	Chris Wright
I
Izik Eidus 已提交
12
 *	Hugh Dickins
I
Izik Eidus 已提交
13 14
 *
 * This work is licensed under the terms of the GNU GPL, version 2.
H
Hugh Dickins 已提交
15 16 17
 */

#include <linux/errno.h>
I
Izik Eidus 已提交
18 19
#include <linux/mm.h>
#include <linux/fs.h>
H
Hugh Dickins 已提交
20
#include <linux/mman.h>
I
Izik Eidus 已提交
21 22 23 24 25 26 27 28 29 30 31 32
#include <linux/sched.h>
#include <linux/rwsem.h>
#include <linux/pagemap.h>
#include <linux/rmap.h>
#include <linux/spinlock.h>
#include <linux/jhash.h>
#include <linux/delay.h>
#include <linux/kthread.h>
#include <linux/wait.h>
#include <linux/slab.h>
#include <linux/rbtree.h>
#include <linux/mmu_notifier.h>
33
#include <linux/swap.h>
H
Hugh Dickins 已提交
34 35
#include <linux/ksm.h>

I
Izik Eidus 已提交
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 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153
#include <asm/tlbflush.h>

/*
 * A few notes about the KSM scanning process,
 * to make it easier to understand the data structures below:
 *
 * In order to reduce excessive scanning, KSM sorts the memory pages by their
 * contents into a data structure that holds pointers to the pages' locations.
 *
 * Since the contents of the pages may change at any moment, KSM cannot just
 * insert the pages into a normal sorted tree and expect it to find anything.
 * Therefore KSM uses two data structures - the stable and the unstable tree.
 *
 * The stable tree holds pointers to all the merged pages (ksm pages), sorted
 * by their contents.  Because each such page is write-protected, searching on
 * this tree is fully assured to be working (except when pages are unmapped),
 * and therefore this tree is called the stable tree.
 *
 * In addition to the stable tree, KSM uses a second data structure called the
 * unstable tree: this tree holds pointers to pages which have been found to
 * be "unchanged for a period of time".  The unstable tree sorts these pages
 * by their contents, but since they are not write-protected, KSM cannot rely
 * upon the unstable tree to work correctly - the unstable tree is liable to
 * be corrupted as its contents are modified, and so it is called unstable.
 *
 * KSM solves this problem by several techniques:
 *
 * 1) The unstable tree is flushed every time KSM completes scanning all
 *    memory areas, and then the tree is rebuilt again from the beginning.
 * 2) KSM will only insert into the unstable tree, pages whose hash value
 *    has not changed since the previous scan of all memory areas.
 * 3) The unstable tree is a RedBlack Tree - so its balancing is based on the
 *    colors of the nodes and not on their contents, assuring that even when
 *    the tree gets "corrupted" it won't get out of balance, so scanning time
 *    remains the same (also, searching and inserting nodes in an rbtree uses
 *    the same algorithm, so we have no overhead when we flush and rebuild).
 * 4) KSM never flushes the stable tree, which means that even if it were to
 *    take 10 attempts to find a page in the unstable tree, once it is found,
 *    it is secured in the stable tree.  (When we scan a new page, we first
 *    compare it against the stable tree, and then against the unstable tree.)
 */

/**
 * struct mm_slot - ksm information per mm that is being scanned
 * @link: link to the mm_slots hash list
 * @mm_list: link into the mm_slots list, rooted in ksm_mm_head
 * @rmap_list: head for this mm_slot's list of rmap_items
 * @mm: the mm that this information is valid for
 */
struct mm_slot {
	struct hlist_node link;
	struct list_head mm_list;
	struct list_head rmap_list;
	struct mm_struct *mm;
};

/**
 * struct ksm_scan - cursor for scanning
 * @mm_slot: the current mm_slot we are scanning
 * @address: the next address inside that to be scanned
 * @rmap_item: the current rmap that we are scanning inside the rmap_list
 * @seqnr: count of completed full scans (needed when removing unstable node)
 *
 * There is only the one ksm_scan instance of this cursor structure.
 */
struct ksm_scan {
	struct mm_slot *mm_slot;
	unsigned long address;
	struct rmap_item *rmap_item;
	unsigned long seqnr;
};

/**
 * struct rmap_item - reverse mapping item for virtual addresses
 * @link: link into mm_slot's rmap_list (rmap_list is per mm)
 * @mm: the memory structure this rmap_item is pointing into
 * @address: the virtual address this rmap_item tracks (+ flags in low bits)
 * @oldchecksum: previous checksum of the page at that virtual address
 * @node: rb_node of this rmap_item in either unstable or stable tree
 * @next: next rmap_item hanging off the same node of the stable tree
 * @prev: previous rmap_item hanging off the same node of the stable tree
 */
struct rmap_item {
	struct list_head link;
	struct mm_struct *mm;
	unsigned long address;		/* + low bits used for flags below */
	union {
		unsigned int oldchecksum;		/* when unstable */
		struct rmap_item *next;			/* when stable */
	};
	union {
		struct rb_node node;			/* when tree node */
		struct rmap_item *prev;			/* in stable list */
	};
};

#define SEQNR_MASK	0x0ff	/* low bits of unstable tree seqnr */
#define NODE_FLAG	0x100	/* is a node of unstable or stable tree */
#define STABLE_FLAG	0x200	/* is a node or list item of stable tree */

/* The stable and unstable tree heads */
static struct rb_root root_stable_tree = RB_ROOT;
static struct rb_root root_unstable_tree = RB_ROOT;

#define MM_SLOTS_HASH_HEADS 1024
static struct hlist_head *mm_slots_hash;

static struct mm_slot ksm_mm_head = {
	.mm_list = LIST_HEAD_INIT(ksm_mm_head.mm_list),
};
static struct ksm_scan ksm_scan = {
	.mm_slot = &ksm_mm_head,
};

static struct kmem_cache *rmap_item_cache;
static struct kmem_cache *mm_slot_cache;

/* The number of nodes in the stable tree */
154
static unsigned long ksm_pages_shared;
I
Izik Eidus 已提交
155

H
Hugh Dickins 已提交
156
/* The number of page slots additionally sharing those nodes */
157
static unsigned long ksm_pages_sharing;
I
Izik Eidus 已提交
158

159 160 161 162 163 164
/* The number of nodes in the unstable tree */
static unsigned long ksm_pages_unshared;

/* The number of rmap_items in use: to calculate pages_volatile */
static unsigned long ksm_rmap_items;

I
Izik Eidus 已提交
165
/* Limit on the number of unswappable pages used */
166
static unsigned long ksm_max_kernel_pages;
I
Izik Eidus 已提交
167 168

/* Number of pages ksmd should scan in one batch */
169
static unsigned int ksm_thread_pages_to_scan = 100;
I
Izik Eidus 已提交
170 171

/* Milliseconds ksmd should sleep between batches */
H
Hugh Dickins 已提交
172
static unsigned int ksm_thread_sleep_millisecs = 20;
I
Izik Eidus 已提交
173 174 175 176

#define KSM_RUN_STOP	0
#define KSM_RUN_MERGE	1
#define KSM_RUN_UNMERGE	2
177
static unsigned int ksm_run = KSM_RUN_STOP;
I
Izik Eidus 已提交
178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213

static DECLARE_WAIT_QUEUE_HEAD(ksm_thread_wait);
static DEFINE_MUTEX(ksm_thread_mutex);
static DEFINE_SPINLOCK(ksm_mmlist_lock);

#define KSM_KMEM_CACHE(__struct, __flags) kmem_cache_create("ksm_"#__struct,\
		sizeof(struct __struct), __alignof__(struct __struct),\
		(__flags), NULL)

static int __init ksm_slab_init(void)
{
	rmap_item_cache = KSM_KMEM_CACHE(rmap_item, 0);
	if (!rmap_item_cache)
		goto out;

	mm_slot_cache = KSM_KMEM_CACHE(mm_slot, 0);
	if (!mm_slot_cache)
		goto out_free;

	return 0;

out_free:
	kmem_cache_destroy(rmap_item_cache);
out:
	return -ENOMEM;
}

static void __init ksm_slab_free(void)
{
	kmem_cache_destroy(mm_slot_cache);
	kmem_cache_destroy(rmap_item_cache);
	mm_slot_cache = NULL;
}

static inline struct rmap_item *alloc_rmap_item(void)
{
214 215 216 217 218 219
	struct rmap_item *rmap_item;

	rmap_item = kmem_cache_zalloc(rmap_item_cache, GFP_KERNEL);
	if (rmap_item)
		ksm_rmap_items++;
	return rmap_item;
I
Izik Eidus 已提交
220 221 222 223
}

static inline void free_rmap_item(struct rmap_item *rmap_item)
{
224
	ksm_rmap_items--;
I
Izik Eidus 已提交
225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286
	rmap_item->mm = NULL;	/* debug safety */
	kmem_cache_free(rmap_item_cache, rmap_item);
}

static inline struct mm_slot *alloc_mm_slot(void)
{
	if (!mm_slot_cache)	/* initialization failed */
		return NULL;
	return kmem_cache_zalloc(mm_slot_cache, GFP_KERNEL);
}

static inline void free_mm_slot(struct mm_slot *mm_slot)
{
	kmem_cache_free(mm_slot_cache, mm_slot);
}

static int __init mm_slots_hash_init(void)
{
	mm_slots_hash = kzalloc(MM_SLOTS_HASH_HEADS * sizeof(struct hlist_head),
				GFP_KERNEL);
	if (!mm_slots_hash)
		return -ENOMEM;
	return 0;
}

static void __init mm_slots_hash_free(void)
{
	kfree(mm_slots_hash);
}

static struct mm_slot *get_mm_slot(struct mm_struct *mm)
{
	struct mm_slot *mm_slot;
	struct hlist_head *bucket;
	struct hlist_node *node;

	bucket = &mm_slots_hash[((unsigned long)mm / sizeof(struct mm_struct))
				% MM_SLOTS_HASH_HEADS];
	hlist_for_each_entry(mm_slot, node, bucket, link) {
		if (mm == mm_slot->mm)
			return mm_slot;
	}
	return NULL;
}

static void insert_to_mm_slots_hash(struct mm_struct *mm,
				    struct mm_slot *mm_slot)
{
	struct hlist_head *bucket;

	bucket = &mm_slots_hash[((unsigned long)mm / sizeof(struct mm_struct))
				% MM_SLOTS_HASH_HEADS];
	mm_slot->mm = mm;
	INIT_LIST_HEAD(&mm_slot->rmap_list);
	hlist_add_head(&mm_slot->link, bucket);
}

static inline int in_stable_tree(struct rmap_item *rmap_item)
{
	return rmap_item->address & STABLE_FLAG;
}

H
Hugh Dickins 已提交
287 288 289 290 291 292 293 294 295 296 297 298 299
/*
 * ksmd, and unmerge_and_remove_all_rmap_items(), must not touch an mm's
 * page tables after it has passed through ksm_exit() - which, if necessary,
 * takes mmap_sem briefly to serialize against them.  ksm_exit() does not set
 * a special flag: they can just back out as soon as mm_users goes to zero.
 * ksm_test_exit() is used throughout to make this test for exit: in some
 * places for correctness, in some places just to avoid unnecessary work.
 */
static inline bool ksm_test_exit(struct mm_struct *mm)
{
	return atomic_read(&mm->mm_users) == 0;
}

I
Izik Eidus 已提交
300 301 302 303 304 305 306 307 308 309 310
/*
 * We use break_ksm to break COW on a ksm page: it's a stripped down
 *
 *	if (get_user_pages(current, mm, addr, 1, 1, 1, &page, NULL) == 1)
 *		put_page(page);
 *
 * but taking great care only to touch a ksm page, in a VM_MERGEABLE vma,
 * in case the application has unmapped and remapped mm,addr meanwhile.
 * Could a ksm page appear anywhere else?  Actually yes, in a VM_PFNMAP
 * mmap of /dev/mem or /dev/kmem, where we would not want to touch it.
 */
H
Hugh Dickins 已提交
311
static int break_ksm(struct vm_area_struct *vma, unsigned long addr)
I
Izik Eidus 已提交
312 313
{
	struct page *page;
H
Hugh Dickins 已提交
314
	int ret = 0;
I
Izik Eidus 已提交
315 316 317 318 319 320 321 322 323 324 325 326

	do {
		cond_resched();
		page = follow_page(vma, addr, FOLL_GET);
		if (!page)
			break;
		if (PageKsm(page))
			ret = handle_mm_fault(vma->vm_mm, vma, addr,
							FAULT_FLAG_WRITE);
		else
			ret = VM_FAULT_WRITE;
		put_page(page);
H
Hugh Dickins 已提交
327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356
	} while (!(ret & (VM_FAULT_WRITE | VM_FAULT_SIGBUS | VM_FAULT_OOM)));
	/*
	 * We must loop because handle_mm_fault() may back out if there's
	 * any difficulty e.g. if pte accessed bit gets updated concurrently.
	 *
	 * VM_FAULT_WRITE is what we have been hoping for: it indicates that
	 * COW has been broken, even if the vma does not permit VM_WRITE;
	 * but note that a concurrent fault might break PageKsm for us.
	 *
	 * VM_FAULT_SIGBUS could occur if we race with truncation of the
	 * backing file, which also invalidates anonymous pages: that's
	 * okay, that truncation will have unmapped the PageKsm for us.
	 *
	 * VM_FAULT_OOM: at the time of writing (late July 2009), setting
	 * aside mem_cgroup limits, VM_FAULT_OOM would only be set if the
	 * current task has TIF_MEMDIE set, and will be OOM killed on return
	 * to user; and ksmd, having no mm, would never be chosen for that.
	 *
	 * But if the mm is in a limited mem_cgroup, then the fault may fail
	 * with VM_FAULT_OOM even if the current task is not TIF_MEMDIE; and
	 * even ksmd can fail in this way - though it's usually breaking ksm
	 * just to undo a merge it made a moment before, so unlikely to oom.
	 *
	 * That's a pity: we might therefore have more kernel pages allocated
	 * than we're counting as nodes in the stable tree; but ksm_do_scan
	 * will retry to break_cow on each pass, so should recover the page
	 * in due course.  The important thing is to not let VM_MERGEABLE
	 * be cleared while any such pages might remain in the area.
	 */
	return (ret & VM_FAULT_OOM) ? -ENOMEM : 0;
I
Izik Eidus 已提交
357 358
}

H
Hugh Dickins 已提交
359
static void break_cow(struct mm_struct *mm, unsigned long addr)
I
Izik Eidus 已提交
360 361 362
{
	struct vm_area_struct *vma;

H
Hugh Dickins 已提交
363
	down_read(&mm->mmap_sem);
H
Hugh Dickins 已提交
364 365
	if (ksm_test_exit(mm))
		goto out;
I
Izik Eidus 已提交
366 367
	vma = find_vma(mm, addr);
	if (!vma || vma->vm_start > addr)
H
Hugh Dickins 已提交
368
		goto out;
I
Izik Eidus 已提交
369
	if (!(vma->vm_flags & VM_MERGEABLE) || !vma->anon_vma)
H
Hugh Dickins 已提交
370
		goto out;
I
Izik Eidus 已提交
371
	break_ksm(vma, addr);
H
Hugh Dickins 已提交
372
out:
I
Izik Eidus 已提交
373 374 375 376 377 378 379 380 381 382 383
	up_read(&mm->mmap_sem);
}

static struct page *get_mergeable_page(struct rmap_item *rmap_item)
{
	struct mm_struct *mm = rmap_item->mm;
	unsigned long addr = rmap_item->address;
	struct vm_area_struct *vma;
	struct page *page;

	down_read(&mm->mmap_sem);
H
Hugh Dickins 已提交
384 385
	if (ksm_test_exit(mm))
		goto out;
I
Izik Eidus 已提交
386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437
	vma = find_vma(mm, addr);
	if (!vma || vma->vm_start > addr)
		goto out;
	if (!(vma->vm_flags & VM_MERGEABLE) || !vma->anon_vma)
		goto out;

	page = follow_page(vma, addr, FOLL_GET);
	if (!page)
		goto out;
	if (PageAnon(page)) {
		flush_anon_page(vma, page, addr);
		flush_dcache_page(page);
	} else {
		put_page(page);
out:		page = NULL;
	}
	up_read(&mm->mmap_sem);
	return page;
}

/*
 * get_ksm_page: checks if the page at the virtual address in rmap_item
 * is still PageKsm, in which case we can trust the content of the page,
 * and it returns the gotten page; but NULL if the page has been zapped.
 */
static struct page *get_ksm_page(struct rmap_item *rmap_item)
{
	struct page *page;

	page = get_mergeable_page(rmap_item);
	if (page && !PageKsm(page)) {
		put_page(page);
		page = NULL;
	}
	return page;
}

/*
 * Removing rmap_item from stable or unstable tree.
 * This function will clean the information from the stable/unstable tree.
 */
static void remove_rmap_item_from_tree(struct rmap_item *rmap_item)
{
	if (in_stable_tree(rmap_item)) {
		struct rmap_item *next_item = rmap_item->next;

		if (rmap_item->address & NODE_FLAG) {
			if (next_item) {
				rb_replace_node(&rmap_item->node,
						&next_item->node,
						&root_stable_tree);
				next_item->address |= NODE_FLAG;
H
Hugh Dickins 已提交
438
				ksm_pages_sharing--;
I
Izik Eidus 已提交
439 440
			} else {
				rb_erase(&rmap_item->node, &root_stable_tree);
441
				ksm_pages_shared--;
I
Izik Eidus 已提交
442 443 444 445 446 447 448 449 450 451
			}
		} else {
			struct rmap_item *prev_item = rmap_item->prev;

			BUG_ON(prev_item->next != rmap_item);
			prev_item->next = next_item;
			if (next_item) {
				BUG_ON(next_item->prev != rmap_item);
				next_item->prev = rmap_item->prev;
			}
H
Hugh Dickins 已提交
452
			ksm_pages_sharing--;
I
Izik Eidus 已提交
453 454 455
		}

		rmap_item->next = NULL;
456
		rmap_item->address &= PAGE_MASK;
I
Izik Eidus 已提交
457 458 459 460

	} else if (rmap_item->address & NODE_FLAG) {
		unsigned char age;
		/*
H
Hugh Dickins 已提交
461
		 * Usually ksmd can and must skip the rb_erase, because
I
Izik Eidus 已提交
462
		 * root_unstable_tree was already reset to RB_ROOT.
H
Hugh Dickins 已提交
463 464 465
		 * But be careful when an mm is exiting: do the rb_erase
		 * if this rmap_item was inserted by this scan, rather
		 * than left over from before.
I
Izik Eidus 已提交
466 467
		 */
		age = (unsigned char)(ksm_scan.seqnr - rmap_item->address);
468
		BUG_ON(age > 1);
I
Izik Eidus 已提交
469 470
		if (!age)
			rb_erase(&rmap_item->node, &root_unstable_tree);
471

472
		ksm_pages_unshared--;
473
		rmap_item->address &= PAGE_MASK;
I
Izik Eidus 已提交
474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499
	}

	cond_resched();		/* we're called from many long loops */
}

static void remove_trailing_rmap_items(struct mm_slot *mm_slot,
				       struct list_head *cur)
{
	struct rmap_item *rmap_item;

	while (cur != &mm_slot->rmap_list) {
		rmap_item = list_entry(cur, struct rmap_item, link);
		cur = cur->next;
		remove_rmap_item_from_tree(rmap_item);
		list_del(&rmap_item->link);
		free_rmap_item(rmap_item);
	}
}

/*
 * Though it's very tempting to unmerge in_stable_tree(rmap_item)s rather
 * than check every pte of a given vma, the locking doesn't quite work for
 * that - an rmap_item is assigned to the stable tree after inserting ksm
 * page and upping mmap_sem.  Nor does it fit with the way we skip dup'ing
 * rmap_items from parent to child at fork time (so as not to waste time
 * if exit comes before the next scan reaches it).
H
Hugh Dickins 已提交
500 501 502 503 504
 *
 * Similarly, although we'd like to remove rmap_items (so updating counts
 * and freeing memory) when unmerging an area, it's easier to leave that
 * to the next pass of ksmd - consider, for example, how ksmd might be
 * in cmp_and_merge_page on one of the rmap_items we would be removing.
I
Izik Eidus 已提交
505
 */
H
Hugh Dickins 已提交
506 507
static int unmerge_ksm_pages(struct vm_area_struct *vma,
			     unsigned long start, unsigned long end)
I
Izik Eidus 已提交
508 509
{
	unsigned long addr;
H
Hugh Dickins 已提交
510
	int err = 0;
I
Izik Eidus 已提交
511

H
Hugh Dickins 已提交
512
	for (addr = start; addr < end && !err; addr += PAGE_SIZE) {
H
Hugh Dickins 已提交
513 514
		if (ksm_test_exit(vma->vm_mm))
			break;
H
Hugh Dickins 已提交
515 516 517 518 519 520
		if (signal_pending(current))
			err = -ERESTARTSYS;
		else
			err = break_ksm(vma, addr);
	}
	return err;
I
Izik Eidus 已提交
521 522
}

H
Hugh Dickins 已提交
523 524 525 526
#ifdef CONFIG_SYSFS
/*
 * Only called through the sysfs control interface:
 */
H
Hugh Dickins 已提交
527
static int unmerge_and_remove_all_rmap_items(void)
I
Izik Eidus 已提交
528 529 530 531
{
	struct mm_slot *mm_slot;
	struct mm_struct *mm;
	struct vm_area_struct *vma;
H
Hugh Dickins 已提交
532 533 534
	int err = 0;

	spin_lock(&ksm_mmlist_lock);
H
Hugh Dickins 已提交
535
	ksm_scan.mm_slot = list_entry(ksm_mm_head.mm_list.next,
H
Hugh Dickins 已提交
536 537
						struct mm_slot, mm_list);
	spin_unlock(&ksm_mmlist_lock);
I
Izik Eidus 已提交
538

H
Hugh Dickins 已提交
539 540
	for (mm_slot = ksm_scan.mm_slot;
			mm_slot != &ksm_mm_head; mm_slot = ksm_scan.mm_slot) {
I
Izik Eidus 已提交
541 542 543
		mm = mm_slot->mm;
		down_read(&mm->mmap_sem);
		for (vma = mm->mmap; vma; vma = vma->vm_next) {
H
Hugh Dickins 已提交
544 545
			if (ksm_test_exit(mm))
				break;
I
Izik Eidus 已提交
546 547
			if (!(vma->vm_flags & VM_MERGEABLE) || !vma->anon_vma)
				continue;
H
Hugh Dickins 已提交
548 549
			err = unmerge_ksm_pages(vma,
						vma->vm_start, vma->vm_end);
H
Hugh Dickins 已提交
550 551
			if (err)
				goto error;
I
Izik Eidus 已提交
552
		}
H
Hugh Dickins 已提交
553

H
Hugh Dickins 已提交
554
		remove_trailing_rmap_items(mm_slot, mm_slot->rmap_list.next);
H
Hugh Dickins 已提交
555 556

		spin_lock(&ksm_mmlist_lock);
H
Hugh Dickins 已提交
557
		ksm_scan.mm_slot = list_entry(mm_slot->mm_list.next,
H
Hugh Dickins 已提交
558
						struct mm_slot, mm_list);
H
Hugh Dickins 已提交
559 560 561 562 563 564 565 566 567 568 569 570 571
		if (ksm_test_exit(mm)) {
			hlist_del(&mm_slot->link);
			list_del(&mm_slot->mm_list);
			spin_unlock(&ksm_mmlist_lock);

			free_mm_slot(mm_slot);
			clear_bit(MMF_VM_MERGEABLE, &mm->flags);
			up_read(&mm->mmap_sem);
			mmdrop(mm);
		} else {
			spin_unlock(&ksm_mmlist_lock);
			up_read(&mm->mmap_sem);
		}
I
Izik Eidus 已提交
572 573
	}

H
Hugh Dickins 已提交
574
	ksm_scan.seqnr = 0;
H
Hugh Dickins 已提交
575 576 577 578
	return 0;

error:
	up_read(&mm->mmap_sem);
I
Izik Eidus 已提交
579
	spin_lock(&ksm_mmlist_lock);
H
Hugh Dickins 已提交
580
	ksm_scan.mm_slot = &ksm_mm_head;
I
Izik Eidus 已提交
581
	spin_unlock(&ksm_mmlist_lock);
H
Hugh Dickins 已提交
582
	return err;
I
Izik Eidus 已提交
583
}
H
Hugh Dickins 已提交
584
#endif /* CONFIG_SYSFS */
I
Izik Eidus 已提交
585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746 747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785 786

static u32 calc_checksum(struct page *page)
{
	u32 checksum;
	void *addr = kmap_atomic(page, KM_USER0);
	checksum = jhash2(addr, PAGE_SIZE / 4, 17);
	kunmap_atomic(addr, KM_USER0);
	return checksum;
}

static int memcmp_pages(struct page *page1, struct page *page2)
{
	char *addr1, *addr2;
	int ret;

	addr1 = kmap_atomic(page1, KM_USER0);
	addr2 = kmap_atomic(page2, KM_USER1);
	ret = memcmp(addr1, addr2, PAGE_SIZE);
	kunmap_atomic(addr2, KM_USER1);
	kunmap_atomic(addr1, KM_USER0);
	return ret;
}

static inline int pages_identical(struct page *page1, struct page *page2)
{
	return !memcmp_pages(page1, page2);
}

static int write_protect_page(struct vm_area_struct *vma, struct page *page,
			      pte_t *orig_pte)
{
	struct mm_struct *mm = vma->vm_mm;
	unsigned long addr;
	pte_t *ptep;
	spinlock_t *ptl;
	int swapped;
	int err = -EFAULT;

	addr = page_address_in_vma(page, vma);
	if (addr == -EFAULT)
		goto out;

	ptep = page_check_address(page, mm, addr, &ptl, 0);
	if (!ptep)
		goto out;

	if (pte_write(*ptep)) {
		pte_t entry;

		swapped = PageSwapCache(page);
		flush_cache_page(vma, addr, page_to_pfn(page));
		/*
		 * Ok this is tricky, when get_user_pages_fast() run it doesnt
		 * take any lock, therefore the check that we are going to make
		 * with the pagecount against the mapcount is racey and
		 * O_DIRECT can happen right after the check.
		 * So we clear the pte and flush the tlb before the check
		 * this assure us that no O_DIRECT can happen after the check
		 * or in the middle of the check.
		 */
		entry = ptep_clear_flush(vma, addr, ptep);
		/*
		 * Check that no O_DIRECT or similar I/O is in progress on the
		 * page
		 */
		if ((page_mapcount(page) + 2 + swapped) != page_count(page)) {
			set_pte_at_notify(mm, addr, ptep, entry);
			goto out_unlock;
		}
		entry = pte_wrprotect(entry);
		set_pte_at_notify(mm, addr, ptep, entry);
	}
	*orig_pte = *ptep;
	err = 0;

out_unlock:
	pte_unmap_unlock(ptep, ptl);
out:
	return err;
}

/**
 * replace_page - replace page in vma by new ksm page
 * @vma:      vma that holds the pte pointing to oldpage
 * @oldpage:  the page we are replacing by newpage
 * @newpage:  the ksm page we replace oldpage by
 * @orig_pte: the original value of the pte
 *
 * Returns 0 on success, -EFAULT on failure.
 */
static int replace_page(struct vm_area_struct *vma, struct page *oldpage,
			struct page *newpage, pte_t orig_pte)
{
	struct mm_struct *mm = vma->vm_mm;
	pgd_t *pgd;
	pud_t *pud;
	pmd_t *pmd;
	pte_t *ptep;
	spinlock_t *ptl;
	unsigned long addr;
	pgprot_t prot;
	int err = -EFAULT;

	prot = vm_get_page_prot(vma->vm_flags & ~VM_WRITE);

	addr = page_address_in_vma(oldpage, vma);
	if (addr == -EFAULT)
		goto out;

	pgd = pgd_offset(mm, addr);
	if (!pgd_present(*pgd))
		goto out;

	pud = pud_offset(pgd, addr);
	if (!pud_present(*pud))
		goto out;

	pmd = pmd_offset(pud, addr);
	if (!pmd_present(*pmd))
		goto out;

	ptep = pte_offset_map_lock(mm, pmd, addr, &ptl);
	if (!pte_same(*ptep, orig_pte)) {
		pte_unmap_unlock(ptep, ptl);
		goto out;
	}

	get_page(newpage);
	page_add_ksm_rmap(newpage);

	flush_cache_page(vma, addr, pte_pfn(*ptep));
	ptep_clear_flush(vma, addr, ptep);
	set_pte_at_notify(mm, addr, ptep, mk_pte(newpage, prot));

	page_remove_rmap(oldpage);
	put_page(oldpage);

	pte_unmap_unlock(ptep, ptl);
	err = 0;
out:
	return err;
}

/*
 * try_to_merge_one_page - take two pages and merge them into one
 * @vma: the vma that hold the pte pointing into oldpage
 * @oldpage: the page that we want to replace with newpage
 * @newpage: the page that we want to map instead of oldpage
 *
 * Note:
 * oldpage should be a PageAnon page, while newpage should be a PageKsm page,
 * or a newly allocated kernel page which page_add_ksm_rmap will make PageKsm.
 *
 * This function returns 0 if the pages were merged, -EFAULT otherwise.
 */
static int try_to_merge_one_page(struct vm_area_struct *vma,
				 struct page *oldpage,
				 struct page *newpage)
{
	pte_t orig_pte = __pte(0);
	int err = -EFAULT;

	if (!(vma->vm_flags & VM_MERGEABLE))
		goto out;

	if (!PageAnon(oldpage))
		goto out;

	get_page(newpage);
	get_page(oldpage);

	/*
	 * We need the page lock to read a stable PageSwapCache in
	 * write_protect_page().  We use trylock_page() instead of
	 * lock_page() because we don't want to wait here - we
	 * prefer to continue scanning and merging different pages,
	 * then come back to this page when it is unlocked.
	 */
	if (!trylock_page(oldpage))
		goto out_putpage;
	/*
	 * If this anonymous page is mapped only here, its pte may need
	 * to be write-protected.  If it's mapped elsewhere, all of its
	 * ptes are necessarily already write-protected.  But in either
	 * case, we need to lock and check page_count is not raised.
	 */
	if (write_protect_page(vma, oldpage, &orig_pte)) {
		unlock_page(oldpage);
		goto out_putpage;
	}
	unlock_page(oldpage);

	if (pages_identical(oldpage, newpage))
		err = replace_page(vma, oldpage, newpage, orig_pte);

out_putpage:
	put_page(oldpage);
	put_page(newpage);
out:
	return err;
}

H
Hugh Dickins 已提交
787 788 789 790 791 792 793 794 795 796 797 798 799
/*
 * try_to_merge_with_ksm_page - like try_to_merge_two_pages,
 * but no new kernel page is allocated: kpage must already be a ksm page.
 */
static int try_to_merge_with_ksm_page(struct mm_struct *mm1,
				      unsigned long addr1,
				      struct page *page1,
				      struct page *kpage)
{
	struct vm_area_struct *vma;
	int err = -EFAULT;

	down_read(&mm1->mmap_sem);
H
Hugh Dickins 已提交
800 801 802
	if (ksm_test_exit(mm1))
		goto out;

H
Hugh Dickins 已提交
803 804 805 806 807 808 809 810 811 812
	vma = find_vma(mm1, addr1);
	if (!vma || vma->vm_start > addr1)
		goto out;

	err = try_to_merge_one_page(vma, page1, kpage);
out:
	up_read(&mm1->mmap_sem);
	return err;
}

I
Izik Eidus 已提交
813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835
/*
 * try_to_merge_two_pages - take two identical pages and prepare them
 * to be merged into one page.
 *
 * This function returns 0 if we successfully mapped two identical pages
 * into one page, -EFAULT otherwise.
 *
 * Note that this function allocates a new kernel page: if one of the pages
 * is already a ksm page, try_to_merge_with_ksm_page should be used.
 */
static int try_to_merge_two_pages(struct mm_struct *mm1, unsigned long addr1,
				  struct page *page1, struct mm_struct *mm2,
				  unsigned long addr2, struct page *page2)
{
	struct vm_area_struct *vma;
	struct page *kpage;
	int err = -EFAULT;

	/*
	 * The number of nodes in the stable tree
	 * is the number of kernel pages that we hold.
	 */
	if (ksm_max_kernel_pages &&
836
	    ksm_max_kernel_pages <= ksm_pages_shared)
I
Izik Eidus 已提交
837 838 839 840 841 842 843
		return err;

	kpage = alloc_page(GFP_HIGHUSER);
	if (!kpage)
		return err;

	down_read(&mm1->mmap_sem);
H
Hugh Dickins 已提交
844 845 846 847
	if (ksm_test_exit(mm1)) {
		up_read(&mm1->mmap_sem);
		goto out;
	}
I
Izik Eidus 已提交
848 849 850
	vma = find_vma(mm1, addr1);
	if (!vma || vma->vm_start > addr1) {
		up_read(&mm1->mmap_sem);
H
Hugh Dickins 已提交
851
		goto out;
I
Izik Eidus 已提交
852 853 854 855 856 857 858
	}

	copy_user_highpage(kpage, page1, addr1, vma);
	err = try_to_merge_one_page(vma, page1, kpage);
	up_read(&mm1->mmap_sem);

	if (!err) {
H
Hugh Dickins 已提交
859
		err = try_to_merge_with_ksm_page(mm2, addr2, page2, kpage);
I
Izik Eidus 已提交
860
		/*
H
Hugh Dickins 已提交
861 862
		 * If that fails, we have a ksm page with only one pte
		 * pointing to it: so break it.
I
Izik Eidus 已提交
863 864 865 866
		 */
		if (err)
			break_cow(mm1, addr1);
	}
H
Hugh Dickins 已提交
867
out:
I
Izik Eidus 已提交
868 869 870 871 872 873 874 875 876 877 878 879 880 881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898 899 900 901 902 903 904 905 906 907 908 909 910 911 912 913 914 915 916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931 932 933 934 935 936 937 938 939 940 941 942 943 944 945 946 947 948 949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967 968 969 970 971 972 973 974 975 976 977 978 979 980 981 982
	put_page(kpage);
	return err;
}

/*
 * stable_tree_search - search page inside the stable tree
 * @page: the page that we are searching identical pages to.
 * @page2: pointer into identical page that we are holding inside the stable
 *	   tree that we have found.
 * @rmap_item: the reverse mapping item
 *
 * This function checks if there is a page inside the stable tree
 * with identical content to the page that we are scanning right now.
 *
 * This function return rmap_item pointer to the identical item if found,
 * NULL otherwise.
 */
static struct rmap_item *stable_tree_search(struct page *page,
					    struct page **page2,
					    struct rmap_item *rmap_item)
{
	struct rb_node *node = root_stable_tree.rb_node;

	while (node) {
		struct rmap_item *tree_rmap_item, *next_rmap_item;
		int ret;

		tree_rmap_item = rb_entry(node, struct rmap_item, node);
		while (tree_rmap_item) {
			BUG_ON(!in_stable_tree(tree_rmap_item));
			cond_resched();
			page2[0] = get_ksm_page(tree_rmap_item);
			if (page2[0])
				break;
			next_rmap_item = tree_rmap_item->next;
			remove_rmap_item_from_tree(tree_rmap_item);
			tree_rmap_item = next_rmap_item;
		}
		if (!tree_rmap_item)
			return NULL;

		ret = memcmp_pages(page, page2[0]);

		if (ret < 0) {
			put_page(page2[0]);
			node = node->rb_left;
		} else if (ret > 0) {
			put_page(page2[0]);
			node = node->rb_right;
		} else {
			return tree_rmap_item;
		}
	}

	return NULL;
}

/*
 * stable_tree_insert - insert rmap_item pointing to new ksm page
 * into the stable tree.
 *
 * @page: the page that we are searching identical page to inside the stable
 *	  tree.
 * @rmap_item: pointer to the reverse mapping item.
 *
 * This function returns rmap_item if success, NULL otherwise.
 */
static struct rmap_item *stable_tree_insert(struct page *page,
					    struct rmap_item *rmap_item)
{
	struct rb_node **new = &root_stable_tree.rb_node;
	struct rb_node *parent = NULL;

	while (*new) {
		struct rmap_item *tree_rmap_item, *next_rmap_item;
		struct page *tree_page;
		int ret;

		tree_rmap_item = rb_entry(*new, struct rmap_item, node);
		while (tree_rmap_item) {
			BUG_ON(!in_stable_tree(tree_rmap_item));
			cond_resched();
			tree_page = get_ksm_page(tree_rmap_item);
			if (tree_page)
				break;
			next_rmap_item = tree_rmap_item->next;
			remove_rmap_item_from_tree(tree_rmap_item);
			tree_rmap_item = next_rmap_item;
		}
		if (!tree_rmap_item)
			return NULL;

		ret = memcmp_pages(page, tree_page);
		put_page(tree_page);

		parent = *new;
		if (ret < 0)
			new = &parent->rb_left;
		else if (ret > 0)
			new = &parent->rb_right;
		else {
			/*
			 * It is not a bug that stable_tree_search() didn't
			 * find this node: because at that time our page was
			 * not yet write-protected, so may have changed since.
			 */
			return NULL;
		}
	}

	rmap_item->address |= NODE_FLAG | STABLE_FLAG;
	rmap_item->next = NULL;
	rb_link_node(&rmap_item->node, parent, new);
	rb_insert_color(&rmap_item->node, &root_stable_tree);

H
Hugh Dickins 已提交
983
	ksm_pages_shared++;
I
Izik Eidus 已提交
984 985 986 987 988 989 990 991 992 993 994 995 996 997 998 999 1000 1001 1002 1003 1004 1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015
	return rmap_item;
}

/*
 * unstable_tree_search_insert - search and insert items into the unstable tree.
 *
 * @page: the page that we are going to search for identical page or to insert
 *	  into the unstable tree
 * @page2: pointer into identical page that was found inside the unstable tree
 * @rmap_item: the reverse mapping item of page
 *
 * This function searches for a page in the unstable tree identical to the
 * page currently being scanned; and if no identical page is found in the
 * tree, we insert rmap_item as a new object into the unstable tree.
 *
 * This function returns pointer to rmap_item found to be identical
 * to the currently scanned page, NULL otherwise.
 *
 * This function does both searching and inserting, because they share
 * the same walking algorithm in an rbtree.
 */
static struct rmap_item *unstable_tree_search_insert(struct page *page,
						struct page **page2,
						struct rmap_item *rmap_item)
{
	struct rb_node **new = &root_unstable_tree.rb_node;
	struct rb_node *parent = NULL;

	while (*new) {
		struct rmap_item *tree_rmap_item;
		int ret;

H
Hugh Dickins 已提交
1016
		cond_resched();
I
Izik Eidus 已提交
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
		tree_rmap_item = rb_entry(*new, struct rmap_item, node);
		page2[0] = get_mergeable_page(tree_rmap_item);
		if (!page2[0])
			return NULL;

		/*
		 * Don't substitute an unswappable ksm page
		 * just for one good swappable forked page.
		 */
		if (page == page2[0]) {
			put_page(page2[0]);
			return NULL;
		}

		ret = memcmp_pages(page, page2[0]);

		parent = *new;
		if (ret < 0) {
			put_page(page2[0]);
			new = &parent->rb_left;
		} else if (ret > 0) {
			put_page(page2[0]);
			new = &parent->rb_right;
		} else {
			return tree_rmap_item;
		}
	}

	rmap_item->address |= NODE_FLAG;
	rmap_item->address |= (ksm_scan.seqnr & SEQNR_MASK);
	rb_link_node(&rmap_item->node, parent, new);
	rb_insert_color(&rmap_item->node, &root_unstable_tree);

1050
	ksm_pages_unshared++;
I
Izik Eidus 已提交
1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069
	return NULL;
}

/*
 * stable_tree_append - add another rmap_item to the linked list of
 * rmap_items hanging off a given node of the stable tree, all sharing
 * the same ksm page.
 */
static void stable_tree_append(struct rmap_item *rmap_item,
			       struct rmap_item *tree_rmap_item)
{
	rmap_item->next = tree_rmap_item->next;
	rmap_item->prev = tree_rmap_item;

	if (tree_rmap_item->next)
		tree_rmap_item->next->prev = rmap_item;

	tree_rmap_item->next = rmap_item;
	rmap_item->address |= STABLE_FLAG;
H
Hugh Dickins 已提交
1070 1071

	ksm_pages_sharing++;
I
Izik Eidus 已提交
1072 1073 1074
}

/*
H
Hugh Dickins 已提交
1075 1076 1077 1078
 * cmp_and_merge_page - first see if page can be merged into the stable tree;
 * if not, compare checksum to previous and if it's the same, see if page can
 * be inserted into the unstable tree, or merged with a page already there and
 * both transferred to the stable tree.
I
Izik Eidus 已提交
1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089
 *
 * @page: the page that we are searching identical page to.
 * @rmap_item: the reverse mapping into the virtual address of this page
 */
static void cmp_and_merge_page(struct page *page, struct rmap_item *rmap_item)
{
	struct page *page2[1];
	struct rmap_item *tree_rmap_item;
	unsigned int checksum;
	int err;

1090
	remove_rmap_item_from_tree(rmap_item);
I
Izik Eidus 已提交
1091 1092 1093 1094

	/* We first start with searching the page inside the stable tree */
	tree_rmap_item = stable_tree_search(page, page2, rmap_item);
	if (tree_rmap_item) {
H
Hugh Dickins 已提交
1095
		if (page == page2[0])			/* forked */
I
Izik Eidus 已提交
1096
			err = 0;
H
Hugh Dickins 已提交
1097
		else
I
Izik Eidus 已提交
1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115
			err = try_to_merge_with_ksm_page(rmap_item->mm,
							 rmap_item->address,
							 page, page2[0]);
		put_page(page2[0]);

		if (!err) {
			/*
			 * The page was successfully merged:
			 * add its rmap_item to the stable tree.
			 */
			stable_tree_append(rmap_item, tree_rmap_item);
		}
		return;
	}

	/*
	 * A ksm page might have got here by fork, but its other
	 * references have already been removed from the stable tree.
H
Hugh Dickins 已提交
1116 1117
	 * Or it might be left over from a break_ksm which failed
	 * when the mem_cgroup had reached its limit: try again now.
I
Izik Eidus 已提交
1118 1119 1120 1121 1122 1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145
	 */
	if (PageKsm(page))
		break_cow(rmap_item->mm, rmap_item->address);

	/*
	 * In case the hash value of the page was changed from the last time we
	 * have calculated it, this page to be changed frequely, therefore we
	 * don't want to insert it to the unstable tree, and we don't want to
	 * waste our time to search if there is something identical to it there.
	 */
	checksum = calc_checksum(page);
	if (rmap_item->oldchecksum != checksum) {
		rmap_item->oldchecksum = checksum;
		return;
	}

	tree_rmap_item = unstable_tree_search_insert(page, page2, rmap_item);
	if (tree_rmap_item) {
		err = try_to_merge_two_pages(rmap_item->mm,
					     rmap_item->address, page,
					     tree_rmap_item->mm,
					     tree_rmap_item->address, page2[0]);
		/*
		 * As soon as we merge this page, we want to remove the
		 * rmap_item of the page we have merged with from the unstable
		 * tree, and insert it instead as new node in the stable tree.
		 */
		if (!err) {
1146
			remove_rmap_item_from_tree(tree_rmap_item);
1147

I
Izik Eidus 已提交
1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169 1170 1171 1172 1173 1174
			/*
			 * If we fail to insert the page into the stable tree,
			 * we will have 2 virtual addresses that are pointing
			 * to a ksm page left outside the stable tree,
			 * in which case we need to break_cow on both.
			 */
			if (stable_tree_insert(page2[0], tree_rmap_item))
				stable_tree_append(rmap_item, tree_rmap_item);
			else {
				break_cow(tree_rmap_item->mm,
						tree_rmap_item->address);
				break_cow(rmap_item->mm, rmap_item->address);
			}
		}

		put_page(page2[0]);
	}
}

static struct rmap_item *get_next_rmap_item(struct mm_slot *mm_slot,
					    struct list_head *cur,
					    unsigned long addr)
{
	struct rmap_item *rmap_item;

	while (cur != &mm_slot->rmap_list) {
		rmap_item = list_entry(cur, struct rmap_item, link);
1175
		if ((rmap_item->address & PAGE_MASK) == addr)
I
Izik Eidus 已提交
1176 1177 1178 1179 1180 1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220
			return rmap_item;
		if (rmap_item->address > addr)
			break;
		cur = cur->next;
		remove_rmap_item_from_tree(rmap_item);
		list_del(&rmap_item->link);
		free_rmap_item(rmap_item);
	}

	rmap_item = alloc_rmap_item();
	if (rmap_item) {
		/* It has already been zeroed */
		rmap_item->mm = mm_slot->mm;
		rmap_item->address = addr;
		list_add_tail(&rmap_item->link, cur);
	}
	return rmap_item;
}

static struct rmap_item *scan_get_next_rmap_item(struct page **page)
{
	struct mm_struct *mm;
	struct mm_slot *slot;
	struct vm_area_struct *vma;
	struct rmap_item *rmap_item;

	if (list_empty(&ksm_mm_head.mm_list))
		return NULL;

	slot = ksm_scan.mm_slot;
	if (slot == &ksm_mm_head) {
		root_unstable_tree = RB_ROOT;

		spin_lock(&ksm_mmlist_lock);
		slot = list_entry(slot->mm_list.next, struct mm_slot, mm_list);
		ksm_scan.mm_slot = slot;
		spin_unlock(&ksm_mmlist_lock);
next_mm:
		ksm_scan.address = 0;
		ksm_scan.rmap_item = list_entry(&slot->rmap_list,
						struct rmap_item, link);
	}

	mm = slot->mm;
	down_read(&mm->mmap_sem);
H
Hugh Dickins 已提交
1221 1222 1223 1224 1225 1226
	if (ksm_test_exit(mm))
		vma = NULL;
	else
		vma = find_vma(mm, ksm_scan.address);

	for (; vma; vma = vma->vm_next) {
I
Izik Eidus 已提交
1227 1228 1229 1230 1231 1232 1233 1234
		if (!(vma->vm_flags & VM_MERGEABLE))
			continue;
		if (ksm_scan.address < vma->vm_start)
			ksm_scan.address = vma->vm_start;
		if (!vma->anon_vma)
			ksm_scan.address = vma->vm_end;

		while (ksm_scan.address < vma->vm_end) {
H
Hugh Dickins 已提交
1235 1236
			if (ksm_test_exit(mm))
				break;
I
Izik Eidus 已提交
1237 1238 1239 1240 1241 1242 1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258
			*page = follow_page(vma, ksm_scan.address, FOLL_GET);
			if (*page && PageAnon(*page)) {
				flush_anon_page(vma, *page, ksm_scan.address);
				flush_dcache_page(*page);
				rmap_item = get_next_rmap_item(slot,
					ksm_scan.rmap_item->link.next,
					ksm_scan.address);
				if (rmap_item) {
					ksm_scan.rmap_item = rmap_item;
					ksm_scan.address += PAGE_SIZE;
				} else
					put_page(*page);
				up_read(&mm->mmap_sem);
				return rmap_item;
			}
			if (*page)
				put_page(*page);
			ksm_scan.address += PAGE_SIZE;
			cond_resched();
		}
	}

H
Hugh Dickins 已提交
1259 1260 1261 1262 1263
	if (ksm_test_exit(mm)) {
		ksm_scan.address = 0;
		ksm_scan.rmap_item = list_entry(&slot->rmap_list,
						struct rmap_item, link);
	}
I
Izik Eidus 已提交
1264 1265 1266 1267 1268 1269 1270
	/*
	 * Nuke all the rmap_items that are above this current rmap:
	 * because there were no VM_MERGEABLE vmas with such addresses.
	 */
	remove_trailing_rmap_items(slot, ksm_scan.rmap_item->link.next);

	spin_lock(&ksm_mmlist_lock);
1271 1272 1273 1274 1275 1276 1277
	ksm_scan.mm_slot = list_entry(slot->mm_list.next,
						struct mm_slot, mm_list);
	if (ksm_scan.address == 0) {
		/*
		 * We've completed a full scan of all vmas, holding mmap_sem
		 * throughout, and found no VM_MERGEABLE: so do the same as
		 * __ksm_exit does to remove this mm from all our lists now.
H
Hugh Dickins 已提交
1278 1279 1280 1281
		 * This applies either when cleaning up after __ksm_exit
		 * (but beware: we can reach here even before __ksm_exit),
		 * or when all VM_MERGEABLE areas have been unmapped (and
		 * mmap_sem then protects against race with MADV_MERGEABLE).
1282 1283 1284
		 */
		hlist_del(&slot->link);
		list_del(&slot->mm_list);
H
Hugh Dickins 已提交
1285 1286
		spin_unlock(&ksm_mmlist_lock);

1287 1288
		free_mm_slot(slot);
		clear_bit(MMF_VM_MERGEABLE, &mm->flags);
H
Hugh Dickins 已提交
1289 1290 1291 1292 1293
		up_read(&mm->mmap_sem);
		mmdrop(mm);
	} else {
		spin_unlock(&ksm_mmlist_lock);
		up_read(&mm->mmap_sem);
1294
	}
I
Izik Eidus 已提交
1295 1296

	/* Repeat until we've completed scanning the whole list */
1297
	slot = ksm_scan.mm_slot;
I
Izik Eidus 已提交
1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320
	if (slot != &ksm_mm_head)
		goto next_mm;

	ksm_scan.seqnr++;
	return NULL;
}

/**
 * ksm_do_scan  - the ksm scanner main worker function.
 * @scan_npages - number of pages we want to scan before we return.
 */
static void ksm_do_scan(unsigned int scan_npages)
{
	struct rmap_item *rmap_item;
	struct page *page;

	while (scan_npages--) {
		cond_resched();
		rmap_item = scan_get_next_rmap_item(&page);
		if (!rmap_item)
			return;
		if (!PageKsm(page) || !in_stable_tree(rmap_item))
			cmp_and_merge_page(page, rmap_item);
H
Hugh Dickins 已提交
1321 1322 1323 1324 1325 1326 1327 1328
		else if (page_mapcount(page) == 1) {
			/*
			 * Replace now-unshared ksm page by ordinary page.
			 */
			break_cow(rmap_item->mm, rmap_item->address);
			remove_rmap_item_from_tree(rmap_item);
			rmap_item->oldchecksum = calc_checksum(page);
		}
I
Izik Eidus 已提交
1329 1330 1331 1332
		put_page(page);
	}
}

H
Hugh Dickins 已提交
1333 1334 1335 1336 1337
static int ksmd_should_run(void)
{
	return (ksm_run & KSM_RUN_MERGE) && !list_empty(&ksm_mm_head.mm_list);
}

I
Izik Eidus 已提交
1338 1339
static int ksm_scan_thread(void *nothing)
{
I
Izik Eidus 已提交
1340
	set_user_nice(current, 5);
I
Izik Eidus 已提交
1341 1342

	while (!kthread_should_stop()) {
H
Hugh Dickins 已提交
1343 1344
		mutex_lock(&ksm_thread_mutex);
		if (ksmd_should_run())
I
Izik Eidus 已提交
1345
			ksm_do_scan(ksm_thread_pages_to_scan);
H
Hugh Dickins 已提交
1346 1347 1348
		mutex_unlock(&ksm_thread_mutex);

		if (ksmd_should_run()) {
I
Izik Eidus 已提交
1349 1350 1351 1352
			schedule_timeout_interruptible(
				msecs_to_jiffies(ksm_thread_sleep_millisecs));
		} else {
			wait_event_interruptible(ksm_thread_wait,
H
Hugh Dickins 已提交
1353
				ksmd_should_run() || kthread_should_stop());
I
Izik Eidus 已提交
1354 1355 1356 1357 1358
		}
	}
	return 0;
}

H
Hugh Dickins 已提交
1359 1360 1361 1362
int ksm_madvise(struct vm_area_struct *vma, unsigned long start,
		unsigned long end, int advice, unsigned long *vm_flags)
{
	struct mm_struct *mm = vma->vm_mm;
H
Hugh Dickins 已提交
1363
	int err;
H
Hugh Dickins 已提交
1364 1365 1366 1367 1368 1369 1370 1371 1372 1373 1374 1375

	switch (advice) {
	case MADV_MERGEABLE:
		/*
		 * Be somewhat over-protective for now!
		 */
		if (*vm_flags & (VM_MERGEABLE | VM_SHARED  | VM_MAYSHARE   |
				 VM_PFNMAP    | VM_IO      | VM_DONTEXPAND |
				 VM_RESERVED  | VM_HUGETLB | VM_INSERTPAGE |
				 VM_MIXEDMAP  | VM_SAO))
			return 0;		/* just ignore the advice */

H
Hugh Dickins 已提交
1376 1377 1378 1379 1380
		if (!test_bit(MMF_VM_MERGEABLE, &mm->flags)) {
			err = __ksm_enter(mm);
			if (err)
				return err;
		}
H
Hugh Dickins 已提交
1381 1382 1383 1384 1385 1386 1387 1388

		*vm_flags |= VM_MERGEABLE;
		break;

	case MADV_UNMERGEABLE:
		if (!(*vm_flags & VM_MERGEABLE))
			return 0;		/* just ignore the advice */

H
Hugh Dickins 已提交
1389 1390 1391 1392 1393
		if (vma->anon_vma) {
			err = unmerge_ksm_pages(vma, start, end);
			if (err)
				return err;
		}
H
Hugh Dickins 已提交
1394 1395 1396 1397 1398 1399 1400 1401 1402 1403

		*vm_flags &= ~VM_MERGEABLE;
		break;
	}

	return 0;
}

int __ksm_enter(struct mm_struct *mm)
{
H
Hugh Dickins 已提交
1404 1405 1406 1407
	struct mm_slot *mm_slot;
	int needs_wakeup;

	mm_slot = alloc_mm_slot();
I
Izik Eidus 已提交
1408 1409 1410
	if (!mm_slot)
		return -ENOMEM;

H
Hugh Dickins 已提交
1411 1412 1413
	/* Check ksm_run too?  Would need tighter locking */
	needs_wakeup = list_empty(&ksm_mm_head.mm_list);

I
Izik Eidus 已提交
1414 1415 1416 1417 1418 1419 1420 1421 1422 1423
	spin_lock(&ksm_mmlist_lock);
	insert_to_mm_slots_hash(mm, mm_slot);
	/*
	 * Insert just behind the scanning cursor, to let the area settle
	 * down a little; when fork is followed by immediate exec, we don't
	 * want ksmd to waste time setting up and tearing down an rmap_list.
	 */
	list_add_tail(&mm_slot->mm_list, &ksm_scan.mm_slot->mm_list);
	spin_unlock(&ksm_mmlist_lock);

H
Hugh Dickins 已提交
1424
	set_bit(MMF_VM_MERGEABLE, &mm->flags);
H
Hugh Dickins 已提交
1425
	atomic_inc(&mm->mm_count);
H
Hugh Dickins 已提交
1426 1427 1428 1429

	if (needs_wakeup)
		wake_up_interruptible(&ksm_thread_wait);

H
Hugh Dickins 已提交
1430 1431 1432
	return 0;
}

1433
void __ksm_exit(struct mm_struct *mm)
H
Hugh Dickins 已提交
1434
{
1435
	struct mm_slot *mm_slot;
H
Hugh Dickins 已提交
1436
	int easy_to_free = 0;
1437

I
Izik Eidus 已提交
1438
	/*
H
Hugh Dickins 已提交
1439 1440 1441 1442 1443 1444
	 * This process is exiting: if it's straightforward (as is the
	 * case when ksmd was never running), free mm_slot immediately.
	 * But if it's at the cursor or has rmap_items linked to it, use
	 * mmap_sem to synchronize with any break_cows before pagetables
	 * are freed, and leave the mm_slot on the list for ksmd to free.
	 * Beware: ksm may already have noticed it exiting and freed the slot.
I
Izik Eidus 已提交
1445
	 */
H
Hugh Dickins 已提交
1446

1447 1448
	spin_lock(&ksm_mmlist_lock);
	mm_slot = get_mm_slot(mm);
H
Hugh Dickins 已提交
1449 1450 1451 1452 1453 1454 1455 1456 1457
	if (mm_slot && ksm_scan.mm_slot != mm_slot) {
		if (list_empty(&mm_slot->rmap_list)) {
			hlist_del(&mm_slot->link);
			list_del(&mm_slot->mm_list);
			easy_to_free = 1;
		} else {
			list_move(&mm_slot->mm_list,
				  &ksm_scan.mm_slot->mm_list);
		}
1458 1459 1460
	}
	spin_unlock(&ksm_mmlist_lock);

H
Hugh Dickins 已提交
1461 1462 1463 1464 1465 1466 1467 1468
	if (easy_to_free) {
		free_mm_slot(mm_slot);
		clear_bit(MMF_VM_MERGEABLE, &mm->flags);
		mmdrop(mm);
	} else if (mm_slot) {
		down_write(&mm->mmap_sem);
		up_write(&mm->mmap_sem);
	}
I
Izik Eidus 已提交
1469 1470
}

H
Hugh Dickins 已提交
1471 1472 1473 1474 1475
#ifdef CONFIG_SYSFS
/*
 * This all compiles without CONFIG_SYSFS, but is a waste of space.
 */

I
Izik Eidus 已提交
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 1532 1533 1534 1535 1536 1537 1538 1539 1540 1541 1542 1543 1544 1545 1546 1547 1548
#define KSM_ATTR_RO(_name) \
	static struct kobj_attribute _name##_attr = __ATTR_RO(_name)
#define KSM_ATTR(_name) \
	static struct kobj_attribute _name##_attr = \
		__ATTR(_name, 0644, _name##_show, _name##_store)

static ssize_t sleep_millisecs_show(struct kobject *kobj,
				    struct kobj_attribute *attr, char *buf)
{
	return sprintf(buf, "%u\n", ksm_thread_sleep_millisecs);
}

static ssize_t sleep_millisecs_store(struct kobject *kobj,
				     struct kobj_attribute *attr,
				     const char *buf, size_t count)
{
	unsigned long msecs;
	int err;

	err = strict_strtoul(buf, 10, &msecs);
	if (err || msecs > UINT_MAX)
		return -EINVAL;

	ksm_thread_sleep_millisecs = msecs;

	return count;
}
KSM_ATTR(sleep_millisecs);

static ssize_t pages_to_scan_show(struct kobject *kobj,
				  struct kobj_attribute *attr, char *buf)
{
	return sprintf(buf, "%u\n", ksm_thread_pages_to_scan);
}

static ssize_t pages_to_scan_store(struct kobject *kobj,
				   struct kobj_attribute *attr,
				   const char *buf, size_t count)
{
	int err;
	unsigned long nr_pages;

	err = strict_strtoul(buf, 10, &nr_pages);
	if (err || nr_pages > UINT_MAX)
		return -EINVAL;

	ksm_thread_pages_to_scan = nr_pages;

	return count;
}
KSM_ATTR(pages_to_scan);

static ssize_t run_show(struct kobject *kobj, struct kobj_attribute *attr,
			char *buf)
{
	return sprintf(buf, "%u\n", ksm_run);
}

static ssize_t run_store(struct kobject *kobj, struct kobj_attribute *attr,
			 const char *buf, size_t count)
{
	int err;
	unsigned long flags;

	err = strict_strtoul(buf, 10, &flags);
	if (err || flags > UINT_MAX)
		return -EINVAL;
	if (flags > KSM_RUN_UNMERGE)
		return -EINVAL;

	/*
	 * KSM_RUN_MERGE sets ksmd running, and 0 stops it running.
	 * KSM_RUN_UNMERGE stops it running and unmerges all rmap_items,
1549
	 * breaking COW to free the unswappable pages_shared (but leaves
I
Izik Eidus 已提交
1550 1551 1552 1553 1554 1555
	 * mm_slots on the list for when ksmd may be set running again).
	 */

	mutex_lock(&ksm_thread_mutex);
	if (ksm_run != flags) {
		ksm_run = flags;
H
Hugh Dickins 已提交
1556
		if (flags & KSM_RUN_UNMERGE) {
H
Hugh Dickins 已提交
1557
			current->flags |= PF_OOM_ORIGIN;
H
Hugh Dickins 已提交
1558
			err = unmerge_and_remove_all_rmap_items();
H
Hugh Dickins 已提交
1559
			current->flags &= ~PF_OOM_ORIGIN;
H
Hugh Dickins 已提交
1560 1561 1562 1563 1564
			if (err) {
				ksm_run = KSM_RUN_STOP;
				count = err;
			}
		}
I
Izik Eidus 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
	}
	mutex_unlock(&ksm_thread_mutex);

	if (flags & KSM_RUN_MERGE)
		wake_up_interruptible(&ksm_thread_wait);

	return count;
}
KSM_ATTR(run);

static ssize_t max_kernel_pages_store(struct kobject *kobj,
				      struct kobj_attribute *attr,
				      const char *buf, size_t count)
{
	int err;
	unsigned long nr_pages;

	err = strict_strtoul(buf, 10, &nr_pages);
	if (err)
		return -EINVAL;

	ksm_max_kernel_pages = nr_pages;

	return count;
}

static ssize_t max_kernel_pages_show(struct kobject *kobj,
				     struct kobj_attribute *attr, char *buf)
{
	return sprintf(buf, "%lu\n", ksm_max_kernel_pages);
}
KSM_ATTR(max_kernel_pages);

1598 1599 1600 1601 1602 1603 1604 1605 1606 1607
static ssize_t pages_shared_show(struct kobject *kobj,
				 struct kobj_attribute *attr, char *buf)
{
	return sprintf(buf, "%lu\n", ksm_pages_shared);
}
KSM_ATTR_RO(pages_shared);

static ssize_t pages_sharing_show(struct kobject *kobj,
				  struct kobj_attribute *attr, char *buf)
{
H
Hugh Dickins 已提交
1608
	return sprintf(buf, "%lu\n", ksm_pages_sharing);
1609 1610 1611
}
KSM_ATTR_RO(pages_sharing);

1612 1613 1614 1615 1616 1617 1618 1619 1620 1621 1622 1623 1624 1625 1626 1627 1628 1629 1630 1631 1632 1633 1634 1635 1636 1637 1638 1639 1640 1641 1642
static ssize_t pages_unshared_show(struct kobject *kobj,
				   struct kobj_attribute *attr, char *buf)
{
	return sprintf(buf, "%lu\n", ksm_pages_unshared);
}
KSM_ATTR_RO(pages_unshared);

static ssize_t pages_volatile_show(struct kobject *kobj,
				   struct kobj_attribute *attr, char *buf)
{
	long ksm_pages_volatile;

	ksm_pages_volatile = ksm_rmap_items - ksm_pages_shared
				- ksm_pages_sharing - ksm_pages_unshared;
	/*
	 * It was not worth any locking to calculate that statistic,
	 * but it might therefore sometimes be negative: conceal that.
	 */
	if (ksm_pages_volatile < 0)
		ksm_pages_volatile = 0;
	return sprintf(buf, "%ld\n", ksm_pages_volatile);
}
KSM_ATTR_RO(pages_volatile);

static ssize_t full_scans_show(struct kobject *kobj,
			       struct kobj_attribute *attr, char *buf)
{
	return sprintf(buf, "%lu\n", ksm_scan.seqnr);
}
KSM_ATTR_RO(full_scans);

I
Izik Eidus 已提交
1643 1644 1645 1646 1647
static struct attribute *ksm_attrs[] = {
	&sleep_millisecs_attr.attr,
	&pages_to_scan_attr.attr,
	&run_attr.attr,
	&max_kernel_pages_attr.attr,
1648 1649
	&pages_shared_attr.attr,
	&pages_sharing_attr.attr,
1650 1651 1652
	&pages_unshared_attr.attr,
	&pages_volatile_attr.attr,
	&full_scans_attr.attr,
I
Izik Eidus 已提交
1653 1654 1655 1656 1657 1658 1659
	NULL,
};

static struct attribute_group ksm_attr_group = {
	.attrs = ksm_attrs,
	.name = "ksm",
};
H
Hugh Dickins 已提交
1660
#endif /* CONFIG_SYSFS */
I
Izik Eidus 已提交
1661 1662 1663 1664 1665 1666

static int __init ksm_init(void)
{
	struct task_struct *ksm_thread;
	int err;

H
Hugh Dickins 已提交
1667
	ksm_max_kernel_pages = totalram_pages / 4;
1668

I
Izik Eidus 已提交
1669 1670 1671 1672 1673 1674 1675 1676 1677 1678 1679 1680 1681 1682 1683
	err = ksm_slab_init();
	if (err)
		goto out;

	err = mm_slots_hash_init();
	if (err)
		goto out_free1;

	ksm_thread = kthread_run(ksm_scan_thread, NULL, "ksmd");
	if (IS_ERR(ksm_thread)) {
		printk(KERN_ERR "ksm: creating kthread failed\n");
		err = PTR_ERR(ksm_thread);
		goto out_free2;
	}

H
Hugh Dickins 已提交
1684
#ifdef CONFIG_SYSFS
I
Izik Eidus 已提交
1685 1686 1687
	err = sysfs_create_group(mm_kobj, &ksm_attr_group);
	if (err) {
		printk(KERN_ERR "ksm: register sysfs failed\n");
H
Hugh Dickins 已提交
1688 1689
		kthread_stop(ksm_thread);
		goto out_free2;
I
Izik Eidus 已提交
1690
	}
H
Hugh Dickins 已提交
1691 1692 1693
#else
	ksm_run = KSM_RUN_MERGE;	/* no way for user to start it */

H
Hugh Dickins 已提交
1694
#endif /* CONFIG_SYSFS */
I
Izik Eidus 已提交
1695 1696 1697 1698 1699 1700 1701 1702 1703

	return 0;

out_free2:
	mm_slots_hash_free();
out_free1:
	ksm_slab_free();
out:
	return err;
H
Hugh Dickins 已提交
1704
}
I
Izik Eidus 已提交
1705
module_init(ksm_init)