Merge branch 'master' into gfs2

1 files changed, 801 insertions, 354 deletions

diff --git a/kernel/power/snapshot.c b/kernel/power/snapshot.c
index 75d4886e648e..1b84313cbab5 100644
--- a/kernel/power/snapshot.c
+++ b/kernel/power/snapshot.c
@@ -34,10 +34,12 @@
 
 #include "power.h"
 
-struct pbe *pagedir_nosave;
+/* List of PBEs used for creating and restoring the suspend image */
+struct pbe *restore_pblist;
+
 static unsigned int nr_copy_pages;
 static unsigned int nr_meta_pages;
-static unsigned long *buffer;
+static void *buffer;
 
 #ifdef CONFIG_HIGHMEM
 unsigned int count_highmem_pages(void)
@@ -156,240 +158,637 @@ static inline int save_highmem(void) {return 0;}
 static inline int restore_highmem(void) {return 0;}
 #endif
 
-static int pfn_is_nosave(unsigned long pfn)
+/**
+ *	@safe_needed - on resume, for storing the PBE list and the image,
+ *	we can only use memory pages that do not conflict with the pages
+ *	used before suspend.
+ *
+ *	The unsafe pages are marked with the PG_nosave_free flag
+ *	and we count them using unsafe_pages
+ */
+
+#define PG_ANY		0
+#define PG_SAFE		1
+#define PG_UNSAFE_CLEAR	1
+#define PG_UNSAFE_KEEP	0
+
+static unsigned int allocated_unsafe_pages;
+
+static void *alloc_image_page(gfp_t gfp_mask, int safe_needed)
 {
-	unsigned long nosave_begin_pfn = __pa(&__nosave_begin) >> PAGE_SHIFT;
-	unsigned long nosave_end_pfn = PAGE_ALIGN(__pa(&__nosave_end)) >> PAGE_SHIFT;
-	return (pfn >= nosave_begin_pfn) && (pfn < nosave_end_pfn);
+	void *res;
+
+	res = (void *)get_zeroed_page(gfp_mask);
+	if (safe_needed)
+		while (res && PageNosaveFree(virt_to_page(res))) {
+			/* The page is unsafe, mark it for swsusp_free() */
+			SetPageNosave(virt_to_page(res));
+			allocated_unsafe_pages++;
+			res = (void *)get_zeroed_page(gfp_mask);
+		}
+	if (res) {
+		SetPageNosave(virt_to_page(res));
+		SetPageNosaveFree(virt_to_page(res));
+	}
+	return res;
+}
+
+unsigned long get_safe_page(gfp_t gfp_mask)
+{
+	return (unsigned long)alloc_image_page(gfp_mask, PG_SAFE);
 }
 
 /**
- *	saveable - Determine whether a page should be cloned or not.
- *	@pfn:	The page
- *
- *	We save a page if it's Reserved, and not in the range of pages
- *	statically defined as 'unsaveable', or if it isn't reserved, and
- *	isn't part of a free chunk of pages.
+ *	free_image_page - free page represented by @addr, allocated with
+ *	alloc_image_page (page flags set by it must be cleared)
  */
 
-static int saveable(struct zone *zone, unsigned long *zone_pfn)
+static inline void free_image_page(void *addr, int clear_nosave_free)
 {
-	unsigned long pfn = *zone_pfn + zone->zone_start_pfn;
-	struct page *page;
+	ClearPageNosave(virt_to_page(addr));
+	if (clear_nosave_free)
+		ClearPageNosaveFree(virt_to_page(addr));
+	free_page((unsigned long)addr);
+}
 
-	if (!pfn_valid(pfn))
-		return 0;
+/* struct linked_page is used to build chains of pages */
 
-	page = pfn_to_page(pfn);
-	BUG_ON(PageReserved(page) && PageNosave(page));
-	if (PageNosave(page))
-		return 0;
-	if (PageReserved(page) && pfn_is_nosave(pfn))
-		return 0;
-	if (PageNosaveFree(page))
-		return 0;
+#define LINKED_PAGE_DATA_SIZE	(PAGE_SIZE - sizeof(void *))
 
-	return 1;
-}
+struct linked_page {
+	struct linked_page *next;
+	char data[LINKED_PAGE_DATA_SIZE];
+} __attribute__((packed));
 
-unsigned int count_data_pages(void)
+static inline void
+free_list_of_pages(struct linked_page *list, int clear_page_nosave)
 {
-	struct zone *zone;
-	unsigned long zone_pfn;
-	unsigned int n = 0;
+	while (list) {
+		struct linked_page *lp = list->next;
 
-	for_each_zone (zone) {
-		if (is_highmem(zone))
-			continue;
-		mark_free_pages(zone);
-		for (zone_pfn = 0; zone_pfn < zone->spanned_pages; ++zone_pfn)
-			n += saveable(zone, &zone_pfn);
+		free_image_page(list, clear_page_nosave);
+		list = lp;
 	}
-	return n;
 }
 
-static void copy_data_pages(struct pbe *pblist)
+/**
+  *	struct chain_allocator is used for allocating small objects out of
+  *	a linked list of pages called 'the chain'.
+  *
+  *	The chain grows each time when there is no room for a new object in
+  *	the current page.  The allocated objects cannot be freed individually.
+  *	It is only possible to free them all at once, by freeing the entire
+  *	chain.
+  *
+  *	NOTE: The chain allocator may be inefficient if the allocated objects
+  *	are not much smaller than PAGE_SIZE.
+  */
+
+struct chain_allocator {
+	struct linked_page *chain;	/* the chain */
+	unsigned int used_space;	/* total size of objects allocated out
+					 * of the current page
+					 */
+	gfp_t gfp_mask;		/* mask for allocating pages */
+	int safe_needed;	/* if set, only "safe" pages are allocated */
+};
+
+static void
+chain_init(struct chain_allocator *ca, gfp_t gfp_mask, int safe_needed)
 {
-	struct zone *zone;
-	unsigned long zone_pfn;
-	struct pbe *pbe, *p;
+	ca->chain = NULL;
+	ca->used_space = LINKED_PAGE_DATA_SIZE;
+	ca->gfp_mask = gfp_mask;
+	ca->safe_needed = safe_needed;
+}
 
-	pbe = pblist;
-	for_each_zone (zone) {
-		if (is_highmem(zone))
-			continue;
-		mark_free_pages(zone);
-		/* This is necessary for swsusp_free() */
-		for_each_pb_page (p, pblist)
-			SetPageNosaveFree(virt_to_page(p));
-		for_each_pbe (p, pblist)
-			SetPageNosaveFree(virt_to_page(p->address));
-		for (zone_pfn = 0; zone_pfn < zone->spanned_pages; ++zone_pfn) {
-			if (saveable(zone, &zone_pfn)) {
-				struct page *page;
-				long *src, *dst;
-				int n;
-
-				page = pfn_to_page(zone_pfn + zone->zone_start_pfn);
-				BUG_ON(!pbe);
-				pbe->orig_address = (unsigned long)page_address(page);
-				/* copy_page and memcpy are not usable for copying task structs. */
-				dst = (long *)pbe->address;
-				src = (long *)pbe->orig_address;
-				for (n = PAGE_SIZE / sizeof(long); n; n--)
-					*dst++ = *src++;
-				pbe = pbe->next;
-			}
-		}
+static void *chain_alloc(struct chain_allocator *ca, unsigned int size)
+{
+	void *ret;
+
+	if (LINKED_PAGE_DATA_SIZE - ca->used_space < size) {
+		struct linked_page *lp;
+
+		lp = alloc_image_page(ca->gfp_mask, ca->safe_needed);
+		if (!lp)
+			return NULL;
+
+		lp->next = ca->chain;
+		ca->chain = lp;
+		ca->used_space = 0;
 	}
-	BUG_ON(pbe);
+	ret = ca->chain->data + ca->used_space;
+	ca->used_space += size;
+	return ret;
 }
 
+static void chain_free(struct chain_allocator *ca, int clear_page_nosave)
+{
+	free_list_of_pages(ca->chain, clear_page_nosave);
+	memset(ca, 0, sizeof(struct chain_allocator));
+}
 
 /**
- *	free_pagedir - free pages allocated with alloc_pagedir()
+ *	Data types related to memory bitmaps.
+ *
+ *	Memory bitmap is a structure consiting of many linked lists of
+ *	objects.  The main list's elements are of type struct zone_bitmap
+ *	and each of them corresonds to one zone.  For each zone bitmap
+ *	object there is a list of objects of type struct bm_block that
+ *	represent each blocks of bit chunks in which information is
+ *	stored.
+ *
+ *	struct memory_bitmap contains a pointer to the main list of zone
+ *	bitmap objects, a struct bm_position used for browsing the bitmap,
+ *	and a pointer to the list of pages used for allocating all of the
+ *	zone bitmap objects and bitmap block objects.
+ *
+ *	NOTE: It has to be possible to lay out the bitmap in memory
+ *	using only allocations of order 0.  Additionally, the bitmap is
+ *	designed to work with arbitrary number of zones (this is over the
+ *	top for now, but let's avoid making unnecessary assumptions ;-).
+ *
+ *	struct zone_bitmap contains a pointer to a list of bitmap block
+ *	objects and a pointer to the bitmap block object that has been
+ *	most recently used for setting bits.  Additionally, it contains the
+ *	pfns that correspond to the start and end of the represented zone.
+ *
+ *	struct bm_block contains a pointer to the memory page in which
+ *	information is stored (in the form of a block of bit chunks
+ *	of type unsigned long each).  It also contains the pfns that
+ *	correspond to the start and end of the represented memory area and
+ *	the number of bit chunks in the block.
+ *
+ *	NOTE: Memory bitmaps are used for two types of operations only:
+ *	"set a bit" and "find the next bit set".  Moreover, the searching
+ *	is always carried out after all of the "set a bit" operations
+ *	on given bitmap.
  */
 
-static void free_pagedir(struct pbe *pblist, int clear_nosave_free)
+#define BM_END_OF_MAP	(~0UL)
+
+#define BM_CHUNKS_PER_BLOCK	(PAGE_SIZE / sizeof(long))
+#define BM_BITS_PER_CHUNK	(sizeof(long) << 3)
+#define BM_BITS_PER_BLOCK	(PAGE_SIZE << 3)
+
+struct bm_block {
+	struct bm_block *next;		/* next element of the list */
+	unsigned long start_pfn;	/* pfn represented by the first bit */
+	unsigned long end_pfn;	/* pfn represented by the last bit plus 1 */
+	unsigned int size;	/* number of bit chunks */
+	unsigned long *data;	/* chunks of bits representing pages */
+};
+
+struct zone_bitmap {
+	struct zone_bitmap *next;	/* next element of the list */
+	unsigned long start_pfn;	/* minimal pfn in this zone */
+	unsigned long end_pfn;		/* maximal pfn in this zone plus 1 */
+	struct bm_block *bm_blocks;	/* list of bitmap blocks */
+	struct bm_block *cur_block;	/* recently used bitmap block */
+};
+
+/* strcut bm_position is used for browsing memory bitmaps */
+
+struct bm_position {
+	struct zone_bitmap *zone_bm;
+	struct bm_block *block;
+	int chunk;
+	int bit;
+};
+
+struct memory_bitmap {
+	struct zone_bitmap *zone_bm_list;	/* list of zone bitmaps */
+	struct linked_page *p_list;	/* list of pages used to store zone
+					 * bitmap objects and bitmap block
+					 * objects
+					 */
+	struct bm_position cur;	/* most recently used bit position */
+};
+
+/* Functions that operate on memory bitmaps */
+
+static inline void memory_bm_reset_chunk(struct memory_bitmap *bm)
 {
-	struct pbe *pbe;
+	bm->cur.chunk = 0;
+	bm->cur.bit = -1;
+}
 
-	while (pblist) {
-		pbe = (pblist + PB_PAGE_SKIP)->next;
-		ClearPageNosave(virt_to_page(pblist));
-		if (clear_nosave_free)
-			ClearPageNosaveFree(virt_to_page(pblist));
-		free_page((unsigned long)pblist);
-		pblist = pbe;
-	}
+static void memory_bm_position_reset(struct memory_bitmap *bm)
+{
+	struct zone_bitmap *zone_bm;
+
+	zone_bm = bm->zone_bm_list;
+	bm->cur.zone_bm = zone_bm;
+	bm->cur.block = zone_bm->bm_blocks;
+	memory_bm_reset_chunk(bm);
 }
 
+static void memory_bm_free(struct memory_bitmap *bm, int clear_nosave_free);
+
 /**
- *	fill_pb_page - Create a list of PBEs on a given memory page
+ *	create_bm_block_list - create a list of block bitmap objects
  */
 
-static inline void fill_pb_page(struct pbe *pbpage)
+static inline struct bm_block *
+create_bm_block_list(unsigned int nr_blocks, struct chain_allocator *ca)
 {
-	struct pbe *p;
+	struct bm_block *bblist = NULL;
+
+	while (nr_blocks-- > 0) {
+		struct bm_block *bb;
 
-	p = pbpage;
-	pbpage += PB_PAGE_SKIP;
-	do
-		p->next = p + 1;
-	while (++p < pbpage);
+		bb = chain_alloc(ca, sizeof(struct bm_block));
+		if (!bb)
+			return NULL;
+
+		bb->next = bblist;
+		bblist = bb;
+	}
+	return bblist;
 }
 
 /**
- *	create_pbe_list - Create a list of PBEs on top of a given chain
- *	of memory pages allocated with alloc_pagedir()
+ *	create_zone_bm_list - create a list of zone bitmap objects
  */
 
-static inline void create_pbe_list(struct pbe *pblist, unsigned int nr_pages)
+static inline struct zone_bitmap *
+create_zone_bm_list(unsigned int nr_zones, struct chain_allocator *ca)
 {
-	struct pbe *pbpage, *p;
-	unsigned int num = PBES_PER_PAGE;
+	struct zone_bitmap *zbmlist = NULL;
 
-	for_each_pb_page (pbpage, pblist) {
-		if (num >= nr_pages)
-			break;
+	while (nr_zones-- > 0) {
+		struct zone_bitmap *zbm;
+
+		zbm = chain_alloc(ca, sizeof(struct zone_bitmap));
+		if (!zbm)
+			return NULL;
+
+		zbm->next = zbmlist;
+		zbmlist = zbm;
+	}
+	return zbmlist;
+}
+
+/**
+  *	memory_bm_create - allocate memory for a memory bitmap
+  */
+
+static int
+memory_bm_create(struct memory_bitmap *bm, gfp_t gfp_mask, int safe_needed)
+{
+	struct chain_allocator ca;
+	struct zone *zone;
+	struct zone_bitmap *zone_bm;
+	struct bm_block *bb;
+	unsigned int nr;
+
+	chain_init(&ca, gfp_mask, safe_needed);
 
-		fill_pb_page(pbpage);
-		num += PBES_PER_PAGE;
+	/* Compute the number of zones */
+	nr = 0;
+	for_each_zone (zone)
+		if (populated_zone(zone) && !is_highmem(zone))
+			nr++;
+
+	/* Allocate the list of zones bitmap objects */
+	zone_bm = create_zone_bm_list(nr, &ca);
+	bm->zone_bm_list = zone_bm;
+	if (!zone_bm) {
+		chain_free(&ca, PG_UNSAFE_CLEAR);
+		return -ENOMEM;
 	}
-	if (pbpage) {
-		for (num -= PBES_PER_PAGE - 1, p = pbpage; num < nr_pages; p++, num++)
-			p->next = p + 1;
-		p->next = NULL;
+
+	/* Initialize the zone bitmap objects */
+	for_each_zone (zone) {
+		unsigned long pfn;
+
+		if (!populated_zone(zone) || is_highmem(zone))
+			continue;
+
+		zone_bm->start_pfn = zone->zone_start_pfn;
+		zone_bm->end_pfn = zone->zone_start_pfn + zone->spanned_pages;
+		/* Allocate the list of bitmap block objects */
+		nr = DIV_ROUND_UP(zone->spanned_pages, BM_BITS_PER_BLOCK);
+		bb = create_bm_block_list(nr, &ca);
+		zone_bm->bm_blocks = bb;
+		zone_bm->cur_block = bb;
+		if (!bb)
+			goto Free;
+
+		nr = zone->spanned_pages;
+		pfn = zone->zone_start_pfn;
+		/* Initialize the bitmap block objects */
+		while (bb) {
+			unsigned long *ptr;
+
+			ptr = alloc_image_page(gfp_mask, safe_needed);
+			bb->data = ptr;
+			if (!ptr)
+				goto Free;
+
+			bb->start_pfn = pfn;
+			if (nr >= BM_BITS_PER_BLOCK) {
+				pfn += BM_BITS_PER_BLOCK;
+				bb->size = BM_CHUNKS_PER_BLOCK;
+				nr -= BM_BITS_PER_BLOCK;
+			} else {
+				/* This is executed only once in the loop */
+				pfn += nr;
+				bb->size = DIV_ROUND_UP(nr, BM_BITS_PER_CHUNK);
+			}
+			bb->end_pfn = pfn;
+			bb = bb->next;
+		}
+		zone_bm = zone_bm->next;
 	}
+	bm->p_list = ca.chain;
+	memory_bm_position_reset(bm);
+	return 0;
+
+Free:
+	bm->p_list = ca.chain;
+	memory_bm_free(bm, PG_UNSAFE_CLEAR);
+	return -ENOMEM;
 }
 
-static unsigned int unsafe_pages;
+/**
+  *	memory_bm_free - free memory occupied by the memory bitmap @bm
+  */
+
+static void memory_bm_free(struct memory_bitmap *bm, int clear_nosave_free)
+{
+	struct zone_bitmap *zone_bm;
+
+	/* Free the list of bit blocks for each zone_bitmap object */
+	zone_bm = bm->zone_bm_list;
+	while (zone_bm) {
+		struct bm_block *bb;
+
+		bb = zone_bm->bm_blocks;
+		while (bb) {
+			if (bb->data)
+				free_image_page(bb->data, clear_nosave_free);
+			bb = bb->next;
+		}
+		zone_bm = zone_bm->next;
+	}
+	free_list_of_pages(bm->p_list, clear_nosave_free);
+	bm->zone_bm_list = NULL;
+}
 
 /**
- *	@safe_needed - on resume, for storing the PBE list and the image,
- *	we can only use memory pages that do not conflict with the pages
- *	used before suspend.
+ *	memory_bm_set_bit - set the bit in the bitmap @bm that corresponds
+ *	to given pfn.  The cur_zone_bm member of @bm and the cur_block member
+ *	of @bm->cur_zone_bm are updated.
  *
- *	The unsafe pages are marked with the PG_nosave_free flag
- *	and we count them using unsafe_pages
+ *	If the bit cannot be set, the function returns -EINVAL .
  */
 
-static inline void *alloc_image_page(gfp_t gfp_mask, int safe_needed)
+static int
+memory_bm_set_bit(struct memory_bitmap *bm, unsigned long pfn)
 {
-	void *res;
-
-	res = (void *)get_zeroed_page(gfp_mask);
-	if (safe_needed)
-		while (res && PageNosaveFree(virt_to_page(res))) {
-			/* The page is unsafe, mark it for swsusp_free() */
-			SetPageNosave(virt_to_page(res));
-			unsafe_pages++;
-			res = (void *)get_zeroed_page(gfp_mask);
+	struct zone_bitmap *zone_bm;
+	struct bm_block *bb;
+
+	/* Check if the pfn is from the current zone */
+	zone_bm = bm->cur.zone_bm;
+	if (pfn < zone_bm->start_pfn || pfn >= zone_bm->end_pfn) {
+		zone_bm = bm->zone_bm_list;
+		/* We don't assume that the zones are sorted by pfns */
+		while (pfn < zone_bm->start_pfn || pfn >= zone_bm->end_pfn) {
+			zone_bm = zone_bm->next;
+			if (unlikely(!zone_bm))
+				return -EINVAL;
 		}
-	if (res) {
-		SetPageNosave(virt_to_page(res));
-		SetPageNosaveFree(virt_to_page(res));
+		bm->cur.zone_bm = zone_bm;
 	}
-	return res;
+	/* Check if the pfn corresponds to the current bitmap block */
+	bb = zone_bm->cur_block;
+	if (pfn < bb->start_pfn)
+		bb = zone_bm->bm_blocks;
+
+	while (pfn >= bb->end_pfn) {
+		bb = bb->next;
+		if (unlikely(!bb))
+			return -EINVAL;
+	}
+	zone_bm->cur_block = bb;
+	pfn -= bb->start_pfn;
+	set_bit(pfn % BM_BITS_PER_CHUNK, bb->data + pfn / BM_BITS_PER_CHUNK);
+	return 0;
 }
 
-unsigned long get_safe_page(gfp_t gfp_mask)
+/* Two auxiliary functions for memory_bm_next_pfn */
+
+/* Find the first set bit in the given chunk, if there is one */
+
+static inline int next_bit_in_chunk(int bit, unsigned long *chunk_p)
 {
-	return (unsigned long)alloc_image_page(gfp_mask, 1);
+	bit++;
+	while (bit < BM_BITS_PER_CHUNK) {
+		if (test_bit(bit, chunk_p))
+			return bit;
+
+		bit++;
+	}
+	return -1;
+}
+
+/* Find a chunk containing some bits set in given block of bits */
+
+static inline int next_chunk_in_block(int n, struct bm_block *bb)
+{
+	n++;
+	while (n < bb->size) {
+		if (bb->data[n])
+			return n;
+
+		n++;
+	}
+	return -1;
 }
 
 /**
- *	alloc_pagedir - Allocate the page directory.
- *
- *	First, determine exactly how many pages we need and
- *	allocate them.
+ *	memory_bm_next_pfn - find the pfn that corresponds to the next set bit
+ *	in the bitmap @bm.  If the pfn cannot be found, BM_END_OF_MAP is
+ *	returned.
  *
- *	We arrange the pages in a chain: each page is an array of PBES_PER_PAGE
- *	struct pbe elements (pbes) and the last element in the page points
- *	to the next page.
+ *	It is required to run memory_bm_position_reset() before the first call to
+ *	this function.
+ */
+
+static unsigned long memory_bm_next_pfn(struct memory_bitmap *bm)
+{
+	struct zone_bitmap *zone_bm;
+	struct bm_block *bb;
+	int chunk;
+	int bit;
+
+	do {
+		bb = bm->cur.block;
+		do {
+			chunk = bm->cur.chunk;
+			bit = bm->cur.bit;
+			do {
+				bit = next_bit_in_chunk(bit, bb->data + chunk);
+				if (bit >= 0)
+					goto Return_pfn;
+
+				chunk = next_chunk_in_block(chunk, bb);
+				bit = -1;
+			} while (chunk >= 0);
+			bb = bb->next;
+			bm->cur.block = bb;
+			memory_bm_reset_chunk(bm);
+		} while (bb);
+		zone_bm = bm->cur.zone_bm->next;
+		if (zone_bm) {
+			bm->cur.zone_bm = zone_bm;
+			bm->cur.block = zone_bm->bm_blocks;
+			memory_bm_reset_chunk(bm);
+		}
+	} while (zone_bm);
+	memory_bm_position_reset(bm);
+	return BM_END_OF_MAP;
+
+Return_pfn:
+	bm->cur.chunk = chunk;
+	bm->cur.bit = bit;
+	return bb->start_pfn + chunk * BM_BITS_PER_CHUNK + bit;
+}
+
+/**
+ *	snapshot_additional_pages - estimate the number of additional pages
+ *	be needed for setting up the suspend image data structures for given
+ *	zone (usually the returned value is greater than the exact number)
+ */
+
+unsigned int snapshot_additional_pages(struct zone *zone)
+{
+	unsigned int res;
+
+	res = DIV_ROUND_UP(zone->spanned_pages, BM_BITS_PER_BLOCK);
+	res += DIV_ROUND_UP(res * sizeof(struct bm_block), PAGE_SIZE);
+	return res;
+}
+
+/**
+ *	pfn_is_nosave - check if given pfn is in the 'nosave' section
+ */
+
+static inline int pfn_is_nosave(unsigned long pfn)
+{
+	unsigned long nosave_begin_pfn = __pa(&__nosave_begin) >> PAGE_SHIFT;
+	unsigned long nosave_end_pfn = PAGE_ALIGN(__pa(&__nosave_end)) >> PAGE_SHIFT;
+	return (pfn >= nosave_begin_pfn) && (pfn < nosave_end_pfn);
+}
+
+/**
+ *	saveable - Determine whether a page should be cloned or not.
+ *	@pfn:	The page
  *
- *	On each page we set up a list of struct_pbe elements.
+ *	We save a page if it isn't Nosave, and is not in the range of pages
+ *	statically defined as 'unsaveable', and it
+ *	isn't a part of a free chunk of pages.
  */
 
-static struct pbe *alloc_pagedir(unsigned int nr_pages, gfp_t gfp_mask,
-				 int safe_needed)
+static struct page *saveable_page(unsigned long pfn)
 {
-	unsigned int num;
-	struct pbe *pblist, *pbe;
+	struct page *page;
+
+	if (!pfn_valid(pfn))
+		return NULL;
 
-	if (!nr_pages)
+	page = pfn_to_page(pfn);
+
+	if (PageNosave(page))
+		return NULL;
+	if (PageReserved(page) && pfn_is_nosave(pfn))
 		return NULL;
+	if (PageNosaveFree(page))
+		return NULL;
+
+	return page;
+}
+
+unsigned int count_data_pages(void)
+{
+	struct zone *zone;
+	unsigned long pfn, max_zone_pfn;
+	unsigned int n = 0;
 
-	pblist = alloc_image_page(gfp_mask, safe_needed);
-	/* FIXME: rewrite this ugly loop */
-	for (pbe = pblist, num = PBES_PER_PAGE; pbe && num < nr_pages;
-        		pbe = pbe->next, num += PBES_PER_PAGE) {
-		pbe += PB_PAGE_SKIP;
-		pbe->next = alloc_image_page(gfp_mask, safe_needed);
+	for_each_zone (zone) {
+		if (is_highmem(zone))
+			continue;
+		mark_free_pages(zone);
+		max_zone_pfn = zone->zone_start_pfn + zone->spanned_pages;
+		for (pfn = zone->zone_start_pfn; pfn < max_zone_pfn; pfn++)
+			n += !!saveable_page(pfn);
 	}
-	if (!pbe) { /* get_zeroed_page() failed */
-		free_pagedir(pblist, 1);
-		pblist = NULL;
-        } else
-		create_pbe_list(pblist, nr_pages);
-	return pblist;
+	return n;
+}
+
+static inline void copy_data_page(long *dst, long *src)
+{
+	int n;
+
+	/* copy_page and memcpy are not usable for copying task structs. */
+	for (n = PAGE_SIZE / sizeof(long); n; n--)
+		*dst++ = *src++;
+}
+
+static void
+copy_data_pages(struct memory_bitmap *copy_bm, struct memory_bitmap *orig_bm)
+{
+	struct zone *zone;
+	unsigned long pfn;
+
+	for_each_zone (zone) {
+		unsigned long max_zone_pfn;
+
+		if (is_highmem(zone))
+			continue;
+
+		mark_free_pages(zone);
+		max_zone_pfn = zone->zone_start_pfn + zone->spanned_pages;
+		for (pfn = zone->zone_start_pfn; pfn < max_zone_pfn; pfn++)
+			if (saveable_page(pfn))
+				memory_bm_set_bit(orig_bm, pfn);
+	}
+	memory_bm_position_reset(orig_bm);
+	memory_bm_position_reset(copy_bm);
+	do {
+		pfn = memory_bm_next_pfn(orig_bm);
+		if (likely(pfn != BM_END_OF_MAP)) {
+			struct page *page;
+			void *src;
+
+			page = pfn_to_page(pfn);
+			src = page_address(page);
+			page = pfn_to_page(memory_bm_next_pfn(copy_bm));
+			copy_data_page(page_address(page), src);
+		}
+	} while (pfn != BM_END_OF_MAP);
 }
 
 /**
- * Free pages we allocated for suspend. Suspend pages are alocated
- * before atomic copy, so we need to free them after resume.
+ *	swsusp_free - free pages allocated for the suspend.
+ *
+ *	Suspend pages are alocated before the atomic copy is made, so we
+ *	need to release them after the resume.
  */
 
 void swsusp_free(void)
 {
 	struct zone *zone;
-	unsigned long zone_pfn;
+	unsigned long pfn, max_zone_pfn;
 
 	for_each_zone(zone) {
-		for (zone_pfn = 0; zone_pfn < zone->spanned_pages; ++zone_pfn)
-			if (pfn_valid(zone_pfn + zone->zone_start_pfn)) {
-				struct page *page;
-				page = pfn_to_page(zone_pfn + zone->zone_start_pfn);
+		max_zone_pfn = zone->zone_start_pfn + zone->spanned_pages;
+		for (pfn = zone->zone_start_pfn; pfn < max_zone_pfn; pfn++)
+			if (pfn_valid(pfn)) {
+				struct page *page = pfn_to_page(pfn);
+
 				if (PageNosave(page) && PageNosaveFree(page)) {
 					ClearPageNosave(page);
 					ClearPageNosaveFree(page);
@@ -399,7 +798,7 @@ void swsusp_free(void)
 	}
 	nr_copy_pages = 0;
 	nr_meta_pages = 0;
-	pagedir_nosave = NULL;
+	restore_pblist = NULL;
 	buffer = NULL;
 }
 
@@ -414,46 +813,57 @@ void swsusp_free(void)
 static int enough_free_mem(unsigned int nr_pages)
 {
 	struct zone *zone;
-	unsigned int n = 0;
+	unsigned int free = 0, meta = 0;
 
 	for_each_zone (zone)
-		if (!is_highmem(zone))
-			n += zone->free_pages;
-	pr_debug("swsusp: available memory: %u pages\n", n);
-	return n > (nr_pages + PAGES_FOR_IO +
-		(nr_pages + PBES_PER_PAGE - 1) / PBES_PER_PAGE);
-}
+		if (!is_highmem(zone)) {
+			free += zone->free_pages;
+			meta += snapshot_additional_pages(zone);
+		}
 
-static int alloc_data_pages(struct pbe *pblist, gfp_t gfp_mask, int safe_needed)
-{
-	struct pbe *p;
+	pr_debug("swsusp: pages needed: %u + %u + %u, available pages: %u\n",
+		nr_pages, PAGES_FOR_IO, meta, free);
 
-	for_each_pbe (p, pblist) {
-		p->address = (unsigned long)alloc_image_page(gfp_mask, safe_needed);
-		if (!p->address)
-			return -ENOMEM;
-	}
-	return 0;
+	return free > nr_pages + PAGES_FOR_IO + meta;
 }
 
-static struct pbe *swsusp_alloc(unsigned int nr_pages)
+static int
+swsusp_alloc(struct memory_bitmap *orig_bm, struct memory_bitmap *copy_bm,
+		unsigned int nr_pages)
 {
-	struct pbe *pblist;
+	int error;
 
-	if (!(pblist = alloc_pagedir(nr_pages, GFP_ATOMIC | __GFP_COLD, 0))) {
-		printk(KERN_ERR "suspend: Allocating pagedir failed.\n");
-		return NULL;
-	}
+	error = memory_bm_create(orig_bm, GFP_ATOMIC | __GFP_COLD, PG_ANY);
+	if (error)
+		goto Free;
 
-	if (alloc_data_pages(pblist, GFP_ATOMIC | __GFP_COLD, 0)) {
-		printk(KERN_ERR "suspend: Allocating image pages failed.\n");
-		swsusp_free();
-		return NULL;
+	error = memory_bm_create(copy_bm, GFP_ATOMIC | __GFP_COLD, PG_ANY);
+	if (error)
+		goto Free;
+
+	while (nr_pages-- > 0) {
+		struct page *page = alloc_page(GFP_ATOMIC | __GFP_COLD);
+		if (!page)
+			goto Free;
+
+		SetPageNosave(page);
+		SetPageNosaveFree(page);
+		memory_bm_set_bit(copy_bm, page_to_pfn(page));
 	}
+	return 0;
 
-	return pblist;
+Free:
+	swsusp_free();
+	return -ENOMEM;
 }
 
+/* Memory bitmap used for marking saveable pages */
+static struct memory_bitmap orig_bm;
+/* Memory bitmap used for marking allocated pages that will contain the copies
+ * of saveable pages
+ */
+static struct memory_bitmap copy_bm;
+
 asmlinkage int swsusp_save(void)
 {
 	unsigned int nr_pages;
@@ -464,25 +874,19 @@ asmlinkage int swsusp_save(void)
 	nr_pages = count_data_pages();
 	printk("swsusp: Need to copy %u pages\n", nr_pages);
 
-	pr_debug("swsusp: pages needed: %u + %lu + %u, free: %u\n",
-		 nr_pages,
-		 (nr_pages + PBES_PER_PAGE - 1) / PBES_PER_PAGE,
-		 PAGES_FOR_IO, nr_free_pages());
-
 	if (!enough_free_mem(nr_pages)) {
 		printk(KERN_ERR "swsusp: Not enough free memory\n");
 		return -ENOMEM;
 	}
 
-	pagedir_nosave = swsusp_alloc(nr_pages);
-	if (!pagedir_nosave)
+	if (swsusp_alloc(&orig_bm, &copy_bm, nr_pages))
 		return -ENOMEM;
 
 	/* During allocating of suspend pagedir, new cold pages may appear.
 	 * Kill them.
 	 */
 	drain_local_pages();
-	copy_data_pages(pagedir_nosave);
+	copy_data_pages(&copy_bm, &orig_bm);
 
 	/*
 	 * End of critical section. From now on, we can write to memory,
@@ -511,22 +915,20 @@ static void init_header(struct swsusp_info *info)
 }
 
 /**
- *	pack_orig_addresses - the .orig_address fields of the PBEs from the
- *	list starting at @pbe are stored in the array @buf[] (1 page)
+ *	pack_pfns - pfns corresponding to the set bits found in the bitmap @bm
+ *	are stored in the array @buf[] (1 page at a time)
  */
 
-static inline struct pbe *pack_orig_addresses(unsigned long *buf, struct pbe *pbe)
+static inline void
+pack_pfns(unsigned long *buf, struct memory_bitmap *bm)
 {
 	int j;
 
-	for (j = 0; j < PAGE_SIZE / sizeof(long) && pbe; j++) {
-		buf[j] = pbe->orig_address;
-		pbe = pbe->next;
+	for (j = 0; j < PAGE_SIZE / sizeof(long); j++) {
+		buf[j] = memory_bm_next_pfn(bm);
+		if (unlikely(buf[j] == BM_END_OF_MAP))
+			break;
 	}
-	if (!pbe)
-		for (; j < PAGE_SIZE / sizeof(long); j++)
-			buf[j] = 0;
-	return pbe;
 }
 
 /**
@@ -553,37 +955,39 @@ static inline struct pbe *pack_orig_addresses(unsigned long *buf, struct pbe *pb
 
 int snapshot_read_next(struct snapshot_handle *handle, size_t count)
 {
-	if (handle->page > nr_meta_pages + nr_copy_pages)
+	if (handle->cur > nr_meta_pages + nr_copy_pages)
 		return 0;
+
 	if (!buffer) {
 		/* This makes the buffer be freed by swsusp_free() */
-		buffer = alloc_image_page(GFP_ATOMIC, 0);
+		buffer = alloc_image_page(GFP_ATOMIC, PG_ANY);
 		if (!buffer)
 			return -ENOMEM;
 	}
 	if (!handle->offset) {
 		init_header((struct swsusp_info *)buffer);
 		handle->buffer = buffer;
-		handle->pbe = pagedir_nosave;
+		memory_bm_position_reset(&orig_bm);
+		memory_bm_position_reset(&copy_bm);
 	}
-	if (handle->prev < handle->page) {
-		if (handle->page <= nr_meta_pages) {
-			handle->pbe = pack_orig_addresses(buffer, handle->pbe);
-			if (!handle->pbe)
-				handle->pbe = pagedir_nosave;
+	if (handle->prev < handle->cur) {
+		if (handle->cur <= nr_meta_pages) {
+			memset(buffer, 0, PAGE_SIZE);
+			pack_pfns(buffer, &orig_bm);
 		} else {
-			handle->buffer = (void *)handle->pbe->address;
-			handle->pbe = handle->pbe->next;
+			unsigned long pfn = memory_bm_next_pfn(&copy_bm);
+
+			handle->buffer = page_address(pfn_to_page(pfn));
 		}
-		handle->prev = handle->page;
+		handle->prev = handle->cur;
 	}
-	handle->buf_offset = handle->page_offset;
-	if (handle->page_offset + count >= PAGE_SIZE) {
-		count = PAGE_SIZE - handle->page_offset;
-		handle->page_offset = 0;
-		handle->page++;
+	handle->buf_offset = handle->cur_offset;
+	if (handle->cur_offset + count >= PAGE_SIZE) {
+		count = PAGE_SIZE - handle->cur_offset;
+		handle->cur_offset = 0;
+		handle->cur++;
 	} else {
-		handle->page_offset += count;
+		handle->cur_offset += count;
 	}
 	handle->offset += count;
 	return count;
@@ -595,47 +999,50 @@ int snapshot_read_next(struct snapshot_handle *handle, size_t count)
  *	had been used before suspend
  */
 
-static int mark_unsafe_pages(struct pbe *pblist)
+static int mark_unsafe_pages(struct memory_bitmap *bm)
 {
 	struct zone *zone;
-	unsigned long zone_pfn;
-	struct pbe *p;
-
-	if (!pblist) /* a sanity check */
-		return -EINVAL;
+	unsigned long pfn, max_zone_pfn;
 
 	/* Clear page flags */
 	for_each_zone (zone) {
-		for (zone_pfn = 0; zone_pfn < zone->spanned_pages; ++zone_pfn)
-			if (pfn_valid(zone_pfn + zone->zone_start_pfn))
-				ClearPageNosaveFree(pfn_to_page(zone_pfn +
-					zone->zone_start_pfn));
+		max_zone_pfn = zone->zone_start_pfn + zone->spanned_pages;
+		for (pfn = zone->zone_start_pfn; pfn < max_zone_pfn; pfn++)
+			if (pfn_valid(pfn))
+				ClearPageNosaveFree(pfn_to_page(pfn));
 	}
 
-	/* Mark orig addresses */
-	for_each_pbe (p, pblist) {
-		if (virt_addr_valid(p->orig_address))
-			SetPageNosaveFree(virt_to_page(p->orig_address));
-		else
-			return -EFAULT;
-	}
+	/* Mark pages that correspond to the "original" pfns as "unsafe" */
+	memory_bm_position_reset(bm);
+	do {
+		pfn = memory_bm_next_pfn(bm);
+		if (likely(pfn != BM_END_OF_MAP)) {
+			if (likely(pfn_valid(pfn)))
+				SetPageNosaveFree(pfn_to_page(pfn));
+			else
+				return -EFAULT;
+		}
+	} while (pfn != BM_END_OF_MAP);
 
-	unsafe_pages = 0;
+	allocated_unsafe_pages = 0;
 
 	return 0;
 }
 
-static void copy_page_backup_list(struct pbe *dst, struct pbe *src)
+static void
+duplicate_memory_bitmap(struct memory_bitmap *dst, struct memory_bitmap *src)
 {
-	/* We assume both lists contain the same number of elements */
-	while (src) {
-		dst->orig_address = src->orig_address;
-		dst = dst->next;
-		src = src->next;
+	unsigned long pfn;
+
+	memory_bm_position_reset(src);
+	pfn = memory_bm_next_pfn(src);
+	while (pfn != BM_END_OF_MAP) {
+		memory_bm_set_bit(dst, pfn);
+		pfn = memory_bm_next_pfn(src);
 	}
 }
 
-static int check_header(struct swsusp_info *info)
+static inline int check_header(struct swsusp_info *info)
 {
 	char *reason = NULL;
 
@@ -662,19 +1069,14 @@ static int check_header(struct swsusp_info *info)
  *	load header - check the image header and copy data from it
  */
 
-static int load_header(struct snapshot_handle *handle,
-                              struct swsusp_info *info)
+static int
+load_header(struct swsusp_info *info)
 {
 	int error;
-	struct pbe *pblist;
 
+	restore_pblist = NULL;
 	error = check_header(info);
 	if (!error) {
-		pblist = alloc_pagedir(info->image_pages, GFP_ATOMIC, 0);
-		if (!pblist)
-			return -ENOMEM;
-		pagedir_nosave = pblist;
-		handle->pbe = pblist;
 		nr_copy_pages = info->image_pages;
 		nr_meta_pages = info->pages - info->image_pages - 1;
 	}
@@ -682,113 +1084,137 @@ static int load_header(struct snapshot_handle *handle,
 }
 
 /**
- *	unpack_orig_addresses - copy the elements of @buf[] (1 page) to
- *	the PBEs in the list starting at @pbe
+ *	unpack_orig_pfns - for each element of @buf[] (1 page at a time) set
+ *	the corresponding bit in the memory bitmap @bm
  */
 
-static inline struct pbe *unpack_orig_addresses(unsigned long *buf,
-                                                struct pbe *pbe)
+static inline void
+unpack_orig_pfns(unsigned long *buf, struct memory_bitmap *bm)
 {
 	int j;
 
-	for (j = 0; j < PAGE_SIZE / sizeof(long) && pbe; j++) {
-		pbe->orig_address = buf[j];
-		pbe = pbe->next;
+	for (j = 0; j < PAGE_SIZE / sizeof(long); j++) {
+		if (unlikely(buf[j] == BM_END_OF_MAP))
+			break;
+
+		memory_bm_set_bit(bm, buf[j]);
 	}
-	return pbe;
 }
 
 /**
- *	prepare_image - use metadata contained in the PBE list
- *	pointed to by pagedir_nosave to mark the pages that will
- *	be overwritten in the process of restoring the system
- *	memory state from the image ("unsafe" pages) and allocate
- *	memory for the image
+ *	prepare_image - use the memory bitmap @bm to mark the pages that will
+ *	be overwritten in the process of restoring the system memory state
+ *	from the suspend image ("unsafe" pages) and allocate memory for the
+ *	image.
  *
- *	The idea is to allocate the PBE list first and then
- *	allocate as many pages as it's needed for the image data,
- *	but not to assign these pages to the PBEs initially.
- *	Instead, we just mark them as allocated and create a list
- *	of "safe" which will be used later
+ *	The idea is to allocate a new memory bitmap first and then allocate
+ *	as many pages as needed for the image data, but not to assign these
+ *	pages to specific tasks initially.  Instead, we just mark them as
+ *	allocated and create a list of "safe" pages that will be used later.
  */
 
-struct safe_page {
-	struct safe_page *next;
-	char padding[PAGE_SIZE - sizeof(void *)];
-};
+#define PBES_PER_LINKED_PAGE	(LINKED_PAGE_DATA_SIZE / sizeof(struct pbe))
 
-static struct safe_page *safe_pages;
+static struct linked_page *safe_pages_list;
 
-static int prepare_image(struct snapshot_handle *handle)
+static int
+prepare_image(struct memory_bitmap *new_bm, struct memory_bitmap *bm)
 {
-	int error = 0;
-	unsigned int nr_pages = nr_copy_pages;
-	struct pbe *p, *pblist = NULL;
+	unsigned int nr_pages;
+	struct linked_page *sp_list, *lp;
+	int error;
 
-	p = pagedir_nosave;
-	error = mark_unsafe_pages(p);
-	if (!error) {
-		pblist = alloc_pagedir(nr_pages, GFP_ATOMIC, 1);
-		if (pblist)
-			copy_page_backup_list(pblist, p);
-		free_pagedir(p, 0);
-		if (!pblist)
+	error = mark_unsafe_pages(bm);
+	if (error)
+		goto Free;
+
+	error = memory_bm_create(new_bm, GFP_ATOMIC, PG_SAFE);
+	if (error)
+		goto Free;
+
+	duplicate_memory_bitmap(new_bm, bm);
+	memory_bm_free(bm, PG_UNSAFE_KEEP);
+	/* Reserve some safe pages for potential later use.
+	 *
+	 * NOTE: This way we make sure there will be enough safe pages for the
+	 * chain_alloc() in get_buffer().  It is a bit wasteful, but
+	 * nr_copy_pages cannot be greater than 50% of the memory anyway.
+	 */
+	sp_list = NULL;
+	/* nr_copy_pages cannot be lesser than allocated_unsafe_pages */
+	nr_pages = nr_copy_pages - allocated_unsafe_pages;
+	nr_pages = DIV_ROUND_UP(nr_pages, PBES_PER_LINKED_PAGE);
+	while (nr_pages > 0) {
+		lp = alloc_image_page(GFP_ATOMIC, PG_SAFE);
+		if (!lp) {
 			error = -ENOMEM;
+			goto Free;
+		}
+		lp->next = sp_list;
+		sp_list = lp;
+		nr_pages--;
 	}
-	safe_pages = NULL;
-	if (!error && nr_pages > unsafe_pages) {
-		nr_pages -= unsafe_pages;
-		while (nr_pages--) {
-			struct safe_page *ptr;
-
-			ptr = (struct safe_page *)get_zeroed_page(GFP_ATOMIC);
-			if (!ptr) {
-				error = -ENOMEM;
-				break;
-			}
-			if (!PageNosaveFree(virt_to_page(ptr))) {
-				/* The page is "safe", add it to the list */
-				ptr->next = safe_pages;
-				safe_pages = ptr;
-			}
-			/* Mark the page as allocated */
-			SetPageNosave(virt_to_page(ptr));
-			SetPageNosaveFree(virt_to_page(ptr));
+	/* Preallocate memory for the image */
+	safe_pages_list = NULL;
+	nr_pages = nr_copy_pages - allocated_unsafe_pages;
+	while (nr_pages > 0) {
+		lp = (struct linked_page *)get_zeroed_page(GFP_ATOMIC);
+		if (!lp) {
+			error = -ENOMEM;
+			goto Free;
+		}
+		if (!PageNosaveFree(virt_to_page(lp))) {
+			/* The page is "safe", add it to the list */
+			lp->next = safe_pages_list;
+			safe_pages_list = lp;
 		}
+		/* Mark the page as allocated */
+		SetPageNosave(virt_to_page(lp));
+		SetPageNosaveFree(virt_to_page(lp));
+		nr_pages--;
 	}
-	if (!error) {
-		pagedir_nosave = pblist;
-	} else {
-		handle->pbe = NULL;
-		swsusp_free();
+	/* Free the reserved safe pages so that chain_alloc() can use them */
+	while (sp_list) {
+		lp = sp_list->next;
+		free_image_page(sp_list, PG_UNSAFE_CLEAR);
+		sp_list = lp;
 	}
+	return 0;
+
+Free:
+	swsusp_free();
 	return error;
 }
 
-static void *get_buffer(struct snapshot_handle *handle)
+/**
+ *	get_buffer - compute the address that snapshot_write_next() should
+ *	set for its caller to write to.
+ */
+
+static void *get_buffer(struct memory_bitmap *bm, struct chain_allocator *ca)
 {
-	struct pbe *pbe = handle->pbe, *last = handle->last_pbe;
-	struct page *page = virt_to_page(pbe->orig_address);
+	struct pbe *pbe;
+	struct page *page = pfn_to_page(memory_bm_next_pfn(bm));
 
-	if (PageNosave(page) && PageNosaveFree(page)) {
-		/*
-		 * We have allocated the "original" page frame and we can
-		 * use it directly to store the read page
+	if (PageNosave(page) && PageNosaveFree(page))
+		/* We have allocated the "original" page frame and we can
+		 * use it directly to store the loaded page.
 		 */
-		pbe->address = 0;
-		if (last && last->next)
-			last->next = NULL;
-		return (void *)pbe->orig_address;
-	}
-	/*
-	 * The "original" page frame has not been allocated and we have to
-	 * use a "safe" page frame to store the read page
+		return page_address(page);
+
+	/* The "original" page frame has not been allocated and we have to
+	 * use a "safe" page frame to store the loaded page.
 	 */
-	pbe->address = (unsigned long)safe_pages;
-	safe_pages = safe_pages->next;
-	if (last)
-		last->next = pbe;
-	handle->last_pbe = pbe;
+	pbe = chain_alloc(ca, sizeof(struct pbe));
+	if (!pbe) {
+		swsusp_free();
+		return NULL;
+	}
+	pbe->orig_address = (unsigned long)page_address(page);
+	pbe->address = (unsigned long)safe_pages_list;
+	safe_pages_list = safe_pages_list->next;
+	pbe->next = restore_pblist;
+	restore_pblist = pbe;
 	return (void *)pbe->address;
 }
 
@@ -816,46 +1242,60 @@ static void *get_buffer(struct snapshot_handle *handle)
 
 int snapshot_write_next(struct snapshot_handle *handle, size_t count)
 {
+	static struct chain_allocator ca;
 	int error = 0;
 
-	if (handle->prev && handle->page > nr_meta_pages + nr_copy_pages)
+	/* Check if we have already loaded the entire image */
+	if (handle->prev && handle->cur > nr_meta_pages + nr_copy_pages)
 		return 0;
+
 	if (!buffer) {
 		/* This makes the buffer be freed by swsusp_free() */
-		buffer = alloc_image_page(GFP_ATOMIC, 0);
+		buffer = alloc_image_page(GFP_ATOMIC, PG_ANY);
 		if (!buffer)
 			return -ENOMEM;
 	}
 	if (!handle->offset)
 		handle->buffer = buffer;
-	if (handle->prev < handle->page) {
-		if (!handle->prev) {
-			error = load_header(handle, (struct swsusp_info *)buffer);
+	handle->sync_read = 1;
+	if (handle->prev < handle->cur) {
+		if (handle->prev == 0) {
+			error = load_header(buffer);
 			if (error)
 				return error;
+
+			error = memory_bm_create(&copy_bm, GFP_ATOMIC, PG_ANY);
+			if (error)
+				return error;
+
 		} else if (handle->prev <= nr_meta_pages) {
-			handle->pbe = unpack_orig_addresses(buffer, handle->pbe);
-			if (!handle->pbe) {
-				error = prepare_image(handle);
+			unpack_orig_pfns(buffer, &copy_bm);
+			if (handle->prev == nr_meta_pages) {
+				error = prepare_image(&orig_bm, &copy_bm);
 				if (error)
 					return error;
-				handle->pbe = pagedir_nosave;
-				handle->last_pbe = NULL;
-				handle->buffer = get_buffer(handle);
+
+				chain_init(&ca, GFP_ATOMIC, PG_SAFE);
+				memory_bm_position_reset(&orig_bm);
+				restore_pblist = NULL;
+				handle->buffer = get_buffer(&orig_bm, &ca);
+				handle->sync_read = 0;
+				if (!handle->buffer)
+					return -ENOMEM;
 			}
 		} else {
-			handle->pbe = handle->pbe->next;
-			handle->buffer = get_buffer(handle);
+			handle->buffer = get_buffer(&orig_bm, &ca);
+			handle->sync_read = 0;
 		}
-		handle->prev = handle->page;
+		handle->prev = handle->cur;
 	}
-	handle->buf_offset = handle->page_offset;
-	if (handle->page_offset + count >= PAGE_SIZE) {
-		count = PAGE_SIZE - handle->page_offset;
-		handle->page_offset = 0;
-		handle->page++;
+	handle->buf_offset = handle->cur_offset;
+	if (handle->cur_offset + count >= PAGE_SIZE) {
+		count = PAGE_SIZE - handle->cur_offset;
+		handle->cur_offset = 0;
+		handle->cur++;
 	} else {
-		handle->page_offset += count;
+		handle->cur_offset += count;
 	}
 	handle->offset += count;
 	return count;
@@ -863,6 +1303,13 @@ int snapshot_write_next(struct snapshot_handle *handle, size_t count)
 
 int snapshot_image_loaded(struct snapshot_handle *handle)
 {
-	return !(!handle->pbe || handle->pbe->next || !nr_copy_pages ||
-		handle->page <= nr_meta_pages + nr_copy_pages);
+	return !(!nr_copy_pages ||
+			handle->cur <= nr_meta_pages + nr_copy_pages);
+}
+
+void snapshot_free_unused_memory(struct snapshot_handle *handle)
+{
+	/* Free only if we have loaded the image entirely */
+	if (handle->prev && handle->cur > nr_meta_pages + nr_copy_pages)
+		memory_bm_free(&orig_bm, PG_UNSAFE_CLEAR);
 }