X-Git-Url: http://git.onelab.eu/?a=blobdiff_plain;f=kernel%2Fpower%2Fsnapshot.c;h=c024606221c4f9d6dd7ba35f3634f9f944fda23e;hb=refs%2Fheads%2Fvserver;hp=8d5a5986d6213919e991835d8a4fcc7ecaaf018d;hpb=76828883507a47dae78837ab5dec5a5b4513c667;p=linux-2.6.git

diff --git a/kernel/power/snapshot.c b/kernel/power/snapshot.c
index 8d5a5986d..c02460622 100644
--- a/kernel/power/snapshot.c
+++ b/kernel/power/snapshot.c
@@ -1,15 +1,16 @@
 /*
  * linux/kernel/power/snapshot.c
  *
- * This file provide system snapshot/restore functionality.
+ * This file provides system snapshot/restore functionality for swsusp.
  *
  * Copyright (C) 1998-2005 Pavel Machek <pavel@suse.cz>
+ * Copyright (C) 2006 Rafael J. Wysocki <rjw@sisk.pl>
  *
- * This file is released under the GPLv2, and is based on swsusp.c.
+ * This file is released under the GPLv2.
  *
  */
 
-
+#include <linux/version.h>
 #include <linux/module.h>
 #include <linux/mm.h>
 #include <linux/suspend.h>
@@ -33,469 +34,980 @@
 
 #include "power.h"
 
-struct pbe *pagedir_nosave;
-unsigned int nr_copy_pages;
+/* List of PBEs needed for restoring the pages that were allocated before
+ * the suspend and included in the suspend image, but have also been
+ * allocated by the "resume" kernel, so their contents cannot be written
+ * directly to their "original" page frames.
+ */
+struct pbe *restore_pblist;
 
-#ifdef CONFIG_HIGHMEM
-unsigned int count_highmem_pages(void)
+/* Pointer to an auxiliary buffer (1 page) */
+static void *buffer;
+
+/**
+ *	@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 have PageNosaveFree set
+ *	and we count them using unsafe_pages.
+ *
+ *	Each allocated image page is marked as PageNosave and PageNosaveFree
+ *	so that swsusp_free() can release it.
+ */
+
+#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 *get_image_page(gfp_t gfp_mask, int safe_needed)
 {
-	struct zone *zone;
-	unsigned long zone_pfn;
-	unsigned int n = 0;
+	void *res;
 
-	for_each_zone (zone)
-		if (is_highmem(zone)) {
-			mark_free_pages(zone);
-			for (zone_pfn = 0; zone_pfn < zone->spanned_pages; zone_pfn++) {
-				struct page *page;
-				unsigned long pfn = zone_pfn + zone->zone_start_pfn;
-				if (!pfn_valid(pfn))
-					continue;
-				page = pfn_to_page(pfn);
-				if (PageReserved(page))
-					continue;
-				if (PageNosaveFree(page))
-					continue;
-				n++;
-			}
+	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);
 		}
-	return n;
+	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)get_image_page(gfp_mask, PG_SAFE);
 }
 
-struct highmem_page {
-	char *data;
+static struct page *alloc_image_page(gfp_t gfp_mask)
+{
 	struct page *page;
-	struct highmem_page *next;
-};
 
-static struct highmem_page *highmem_copy;
+	page = alloc_page(gfp_mask);
+	if (page) {
+		SetPageNosave(page);
+		SetPageNosaveFree(page);
+	}
+	return page;
+}
+
+/**
+ *	free_image_page - free page represented by @addr, allocated with
+ *	get_image_page (page flags set by it must be cleared)
+ */
 
-static int save_highmem_zone(struct zone *zone)
+static inline void free_image_page(void *addr, int clear_nosave_free)
 {
-	unsigned long zone_pfn;
-	mark_free_pages(zone);
-	for (zone_pfn = 0; zone_pfn < zone->spanned_pages; ++zone_pfn) {
-		struct page *page;
-		struct highmem_page *save;
-		void *kaddr;
-		unsigned long pfn = zone_pfn + zone->zone_start_pfn;
+	struct page *page;
 
-		if (!(pfn%1000))
-			printk(".");
-		if (!pfn_valid(pfn))
-			continue;
-		page = pfn_to_page(pfn);
-		/*
-		 * This condition results from rvmalloc() sans vmalloc_32()
-		 * and architectural memory reservations. This should be
-		 * corrected eventually when the cases giving rise to this
-		 * are better understood.
-		 */
-		if (PageReserved(page))
-			continue;
-		BUG_ON(PageNosave(page));
-		if (PageNosaveFree(page))
-			continue;
-		save = kmalloc(sizeof(struct highmem_page), GFP_ATOMIC);
-		if (!save)
-			return -ENOMEM;
-		save->next = highmem_copy;
-		save->page = page;
-		save->data = (void *) get_zeroed_page(GFP_ATOMIC);
-		if (!save->data) {
-			kfree(save);
-			return -ENOMEM;
-		}
-		kaddr = kmap_atomic(page, KM_USER0);
-		memcpy(save->data, kaddr, PAGE_SIZE);
-		kunmap_atomic(kaddr, KM_USER0);
-		highmem_copy = save;
-	}
-	return 0;
+	BUG_ON(!virt_addr_valid(addr));
+
+	page = virt_to_page(addr);
+
+	ClearPageNosave(page);
+	if (clear_nosave_free)
+		ClearPageNosaveFree(page);
+
+	__free_page(page);
 }
 
-int save_highmem(void)
+/* struct linked_page is used to build chains of pages */
+
+#define LINKED_PAGE_DATA_SIZE	(PAGE_SIZE - sizeof(void *))
+
+struct linked_page {
+	struct linked_page *next;
+	char data[LINKED_PAGE_DATA_SIZE];
+} __attribute__((packed));
+
+static inline void
+free_list_of_pages(struct linked_page *list, int clear_page_nosave)
 {
-	struct zone *zone;
-	int res = 0;
+	while (list) {
+		struct linked_page *lp = list->next;
 
-	pr_debug("swsusp: Saving Highmem\n");
-	for_each_zone (zone) {
-		if (is_highmem(zone))
-			res = save_highmem_zone(zone);
-		if (res)
-			return res;
+		free_image_page(list, clear_page_nosave);
+		list = lp;
 	}
-	return 0;
 }
 
-int restore_highmem(void)
+/**
+  *	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)
+{
+	ca->chain = NULL;
+	ca->used_space = LINKED_PAGE_DATA_SIZE;
+	ca->gfp_mask = gfp_mask;
+	ca->safe_needed = safe_needed;
+}
+
+static void *chain_alloc(struct chain_allocator *ca, unsigned int size)
 {
-	printk("swsusp: Restoring Highmem\n");
-	while (highmem_copy) {
-		struct highmem_page *save = highmem_copy;
-		void *kaddr;
-		highmem_copy = save->next;
+	void *ret;
 
-		kaddr = kmap_atomic(save->page, KM_USER0);
-		memcpy(kaddr, save->data, PAGE_SIZE);
-		kunmap_atomic(kaddr, KM_USER0);
-		free_page((long) save->data);
-		kfree(save);
+	if (LINKED_PAGE_DATA_SIZE - ca->used_space < size) {
+		struct linked_page *lp;
+
+		lp = get_image_page(ca->gfp_mask, ca->safe_needed);
+		if (!lp)
+			return NULL;
+
+		lp->next = ca->chain;
+		ca->chain = lp;
+		ca->used_space = 0;
 	}
-	return 0;
+	ret = ca->chain->data + ca->used_space;
+	ca->used_space += size;
+	return ret;
 }
-#endif
 
-static int pfn_is_nosave(unsigned long pfn)
+static void chain_free(struct chain_allocator *ca, int clear_page_nosave)
 {
-	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);
+	free_list_of_pages(ca->chain, clear_page_nosave);
+	memset(ca, 0, sizeof(struct chain_allocator));
 }
 
 /**
- *	saveable - Determine whether a page should be cloned or not.
- *	@pfn:	The page
+ *	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.
  *
- *	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.
+ *	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 int saveable(struct zone *zone, unsigned long *zone_pfn)
-{
-	unsigned long pfn = *zone_pfn + zone->zone_start_pfn;
-	struct page *page;
+#define BM_END_OF_MAP	(~0UL)
 
-	if (!pfn_valid(pfn))
-		return 0;
+#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)
 
-	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;
+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 */
+};
 
-	return 1;
-}
+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 */
+};
 
-unsigned int count_data_pages(void)
-{
-	struct zone *zone;
-	unsigned long zone_pfn;
-	unsigned int n = 0;
+/* strcut bm_position is used for browsing memory bitmaps */
 
-	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);
-	}
-	return n;
+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)
+{
+	bm->cur.chunk = 0;
+	bm->cur.bit = -1;
 }
 
-static void copy_data_pages(struct pbe *pblist)
+static void memory_bm_position_reset(struct memory_bitmap *bm)
 {
-	struct zone *zone;
-	unsigned long zone_pfn;
-	struct pbe *pbe, *p;
+	struct zone_bitmap *zone_bm;
 
-	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;
-				page = pfn_to_page(zone_pfn + zone->zone_start_pfn);
-				BUG_ON(!pbe);
-				pbe->orig_address = (unsigned long)page_address(page);
-				/* copy_page is not usable for copying task structs. */
-				memcpy((void *)pbe->address, (void *)pbe->orig_address, PAGE_SIZE);
-				pbe = pbe->next;
-			}
-		}
-	}
-	BUG_ON(pbe);
+	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);
 
 /**
- *	free_pagedir - free pages allocated with alloc_pagedir()
+ *	create_bm_block_list - create a list of block bitmap objects
  */
 
-void free_pagedir(struct pbe *pblist)
+static inline struct bm_block *
+create_bm_block_list(unsigned int nr_blocks, struct chain_allocator *ca)
 {
-	struct pbe *pbe;
+	struct bm_block *bblist = NULL;
+
+	while (nr_blocks-- > 0) {
+		struct bm_block *bb;
+
+		bb = chain_alloc(ca, sizeof(struct bm_block));
+		if (!bb)
+			return NULL;
 
-	while (pblist) {
-		pbe = (pblist + PB_PAGE_SKIP)->next;
-		ClearPageNosave(virt_to_page(pblist));
-		ClearPageNosaveFree(virt_to_page(pblist));
-		free_page((unsigned long)pblist);
-		pblist = pbe;
+		bb->next = bblist;
+		bblist = bb;
 	}
+	return bblist;
 }
 
 /**
- *	fill_pb_page - Create a list of PBEs on a given memory page
+ *	create_zone_bm_list - create a list of zone bitmap objects
  */
 
-static inline void fill_pb_page(struct pbe *pbpage)
+static inline struct zone_bitmap *
+create_zone_bm_list(unsigned int nr_zones, struct chain_allocator *ca)
 {
-	struct pbe *p;
+	struct zone_bitmap *zbmlist = NULL;
 
-	p = pbpage;
-	pbpage += PB_PAGE_SKIP;
-	do
-		p->next = p + 1;
-	while (++p < pbpage);
+	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;
 }
 
 /**
- *	create_pbe_list - Create a list of PBEs on top of a given chain
- *	of memory pages allocated with alloc_pagedir()
- */
+  *	memory_bm_create - allocate memory for a memory bitmap
+  */
 
-static inline void create_pbe_list(struct pbe *pblist, unsigned int nr_pages)
+static int
+memory_bm_create(struct memory_bitmap *bm, gfp_t gfp_mask, int safe_needed)
 {
-	struct pbe *pbpage, *p;
-	unsigned int num = PBES_PER_PAGE;
+	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);
+
+	/* Compute the number of zones */
+	nr = 0;
+	for_each_zone(zone)
+		if (populated_zone(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;
+	}
 
-	for_each_pb_page (pbpage, pblist) {
-		if (num >= nr_pages)
-			break;
+	/* Initialize the zone bitmap objects */
+	for_each_zone(zone) {
+		unsigned long pfn;
+
+		if (!populated_zone(zone))
+			continue;
 
-		fill_pb_page(pbpage);
-		num += PBES_PER_PAGE;
+		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 = get_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;
 	}
-	if (pbpage) {
-		for (num -= PBES_PER_PAGE - 1, p = pbpage; num < nr_pages; p++, num++)
-			p->next = p + 1;
-		p->next = NULL;
+	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;
+}
+
+/**
+  *	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;
 }
 
 /**
- *	On resume it is necessary to trace and eventually free the unsafe
- *	pages that have been allocated, because they are needed for I/O
- *	(on x86-64 we likely will "eat" these pages once again while
- *	creating the temporary page translation tables)
+ *	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.
+ *
+ *	If the bit cannot be set, the function returns -EINVAL .
  */
 
-struct eaten_page {
-	struct eaten_page *next;
-	char padding[PAGE_SIZE - sizeof(void *)];
-};
+static int
+memory_bm_set_bit(struct memory_bitmap *bm, unsigned long pfn)
+{
+	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;
+		}
+		bm->cur.zone_bm = zone_bm;
+	}
+	/* 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;
+}
+
+/* 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)
+{
+	bit++;
+	while (bit < BM_BITS_PER_CHUNK) {
+		if (test_bit(bit, chunk_p))
+			return bit;
+
+		bit++;
+	}
+	return -1;
+}
 
-static struct eaten_page *eaten_pages = NULL;
+/* Find a chunk containing some bits set in given block of bits */
 
-void release_eaten_pages(void)
+static inline int next_chunk_in_block(int n, struct bm_block *bb)
 {
-	struct eaten_page *p, *q;
+	n++;
+	while (n < bb->size) {
+		if (bb->data[n])
+			return n;
 
-	p = eaten_pages;
-	while (p) {
-		q = p->next;
-		/* We don't want swsusp_free() to free this page again */
-		ClearPageNosave(virt_to_page(p));
-		free_page((unsigned long)p);
-		p = q;
+		n++;
 	}
-	eaten_pages = NULL;
+	return -1;
 }
 
 /**
- *	@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
- *	which had been used before suspend.
+ *	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.
  *
- *	The unsafe pages are marked with the PG_nosave_free flag
- *
- *	Allocated but unusable (ie eaten) memory pages should be marked
- *	so that swsusp_free() can release them
+ *	It is required to run memory_bm_position_reset() before the first call to
+ *	this function.
  */
 
-static inline void *alloc_image_page(gfp_t gfp_mask, int safe_needed)
+static unsigned long memory_bm_next_pfn(struct memory_bitmap *bm)
 {
-	void *res;
+	struct zone_bitmap *zone_bm;
+	struct bm_block *bb;
+	int chunk;
+	int bit;
 
-	if (safe_needed)
+	do {
+		bb = bm->cur.block;
 		do {
-			res = (void *)get_zeroed_page(gfp_mask);
-			if (res && PageNosaveFree(virt_to_page(res))) {
-				/* This is for swsusp_free() */
-				SetPageNosave(virt_to_page(res));
-				((struct eaten_page *)res)->next = eaten_pages;
-				eaten_pages = res;
-			}
-		} while (res && PageNosaveFree(virt_to_page(res)));
-	else
-		res = (void *)get_zeroed_page(gfp_mask);
-	if (res) {
-		SetPageNosave(virt_to_page(res));
-		SetPageNosaveFree(virt_to_page(res));
-	}
-	return res;
+			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;
 }
 
-unsigned long get_safe_page(gfp_t gfp_mask)
+/**
+ *	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)
 {
-	return (unsigned long)alloc_image_page(gfp_mask, 1);
+	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 2 * res;
 }
 
+#ifdef CONFIG_HIGHMEM
 /**
- *	alloc_pagedir - Allocate the page directory.
- *
- *	First, determine exactly how many pages we need and
- *	allocate them.
- *
- *	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.
- *
- *	On each page we set up a list of struct_pbe elements.
+ *	count_free_highmem_pages - compute the total number of free highmem
+ *	pages, system-wide.
  */
 
-struct pbe *alloc_pagedir(unsigned int nr_pages, gfp_t gfp_mask, int safe_needed)
+static unsigned int count_free_highmem_pages(void)
 {
-	unsigned int num;
-	struct pbe *pblist, *pbe;
+	struct zone *zone;
+	unsigned int cnt = 0;
 
-	if (!nr_pages)
-		return NULL;
+	for_each_zone(zone)
+		if (populated_zone(zone) && is_highmem(zone))
+			cnt += zone->free_pages;
 
-	pr_debug("alloc_pagedir(): nr_pages = %d\n", nr_pages);
-	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);
-	}
-	if (!pbe) { /* get_zeroed_page() failed */
-		free_pagedir(pblist);
-		pblist = NULL;
-        } else
-        	create_pbe_list(pblist, nr_pages);
-	return pblist;
+	return cnt;
 }
 
 /**
- * Free pages we allocated for suspend. Suspend pages are alocated
- * before atomic copy, so we need to free them after resume.
+ *	saveable_highmem_page - Determine whether a highmem page should be
+ *	included in the suspend image.
+ *
+ *	We should save the page if it isn't Nosave or NosaveFree, or Reserved,
+ *	and it isn't a part of a free chunk of pages.
  */
 
-void swsusp_free(void)
+static struct page *saveable_highmem_page(unsigned long pfn)
 {
-	struct zone *zone;
-	unsigned long zone_pfn;
+	struct page *page;
 
-	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);
-				if (PageNosave(page) && PageNosaveFree(page)) {
-					ClearPageNosave(page);
-					ClearPageNosaveFree(page);
-					free_page((long) page_address(page));
-				}
-			}
-	}
-}
+	if (!pfn_valid(pfn))
+		return NULL;
+
+	page = pfn_to_page(pfn);
+
+	BUG_ON(!PageHighMem(page));
+
+	if (PageNosave(page) || PageReserved(page) || PageNosaveFree(page))
+		return NULL;
 
+	return page;
+}
 
 /**
- *	enough_free_mem - Make sure we enough free memory to snapshot.
- *
- *	Returns TRUE or FALSE after checking the number of available
- *	free pages.
+ *	count_highmem_pages - compute the total number of saveable highmem
+ *	pages.
  */
 
-static int enough_free_mem(unsigned int nr_pages)
+unsigned int count_highmem_pages(void)
 {
 	struct zone *zone;
 	unsigned int n = 0;
 
-	for_each_zone (zone)
+	for_each_zone(zone) {
+		unsigned long pfn, max_zone_pfn;
+
 		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);
+			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_highmem_page(pfn))
+				n++;
+	}
+	return n;
 }
+#else
+static inline void *saveable_highmem_page(unsigned long pfn) { return NULL; }
+static inline unsigned int count_highmem_pages(void) { return 0; }
+#endif /* CONFIG_HIGHMEM */
 
-int alloc_data_pages(struct pbe *pblist, gfp_t gfp_mask, int safe_needed)
-{
-	struct pbe *p;
+/**
+ *	pfn_is_nosave - check if given pfn is in the 'nosave' section
+ */
 
-	for_each_pbe (p, pblist) {
-		p->address = (unsigned long)alloc_image_page(gfp_mask, safe_needed);
-		if (!p->address)
-			return -ENOMEM;
-	}
-	return 0;
+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);
 }
 
-static struct pbe *swsusp_alloc(unsigned int nr_pages)
+/**
+ *	saveable - Determine whether a non-highmem page should be included in
+ *	the suspend image.
+ *
+ *	We should save the 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 page *saveable_page(unsigned long pfn)
 {
-	struct pbe *pblist;
+	struct page *page;
 
-	if (!(pblist = alloc_pagedir(nr_pages, GFP_ATOMIC | __GFP_COLD, 0))) {
-		printk(KERN_ERR "suspend: Allocating pagedir failed.\n");
+	if (!pfn_valid(pfn))
 		return NULL;
-	}
 
-	if (alloc_data_pages(pblist, GFP_ATOMIC | __GFP_COLD, 0)) {
-		printk(KERN_ERR "suspend: Allocating image pages failed.\n");
-		swsusp_free();
+	page = pfn_to_page(pfn);
+
+	BUG_ON(PageHighMem(page));
+
+	if (PageNosave(page) || PageNosaveFree(page))
+		return NULL;
+
+	if (PageReserved(page) && pfn_is_nosave(pfn))
 		return NULL;
-	}
 
-	return pblist;
+	return page;
 }
 
-asmlinkage int swsusp_save(void)
+/**
+ *	count_data_pages - compute the total number of saveable non-highmem
+ *	pages.
+ */
+
+unsigned int count_data_pages(void)
 {
-	unsigned int nr_pages;
+	struct zone *zone;
+	unsigned long pfn, max_zone_pfn;
+	unsigned int n = 0;
 
-	pr_debug("swsusp: critical section: \n");
+	for_each_zone(zone) {
+		if (is_highmem(zone))
+			continue;
 
-	drain_local_pages();
-	nr_pages = count_data_pages();
-	printk("swsusp: Need to copy %u pages\n", nr_pages);
+		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))
+				n++;
+	}
+	return n;
+}
 
-	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());
+/* This is needed, because copy_page and memcpy are not usable for copying
+ * task structs.
+ */
+static inline void do_copy_page(long *dst, long *src)
+{
+	int n;
 
-	if (!enough_free_mem(nr_pages)) {
-		printk(KERN_ERR "swsusp: Not enough free memory\n");
-		return -ENOMEM;
-	}
+	for (n = PAGE_SIZE / sizeof(long); n; n--)
+		*dst++ = *src++;
+}
 
-	pagedir_nosave = swsusp_alloc(nr_pages);
-	if (!pagedir_nosave)
-		return -ENOMEM;
+#ifdef CONFIG_HIGHMEM
+static inline struct page *
+page_is_saveable(struct zone *zone, unsigned long pfn)
+{
+	return is_highmem(zone) ?
+			saveable_highmem_page(pfn) : saveable_page(pfn);
+}
+
+static inline void
+copy_data_page(unsigned long dst_pfn, unsigned long src_pfn)
+{
+	struct page *s_page, *d_page;
+	void *src, *dst;
+
+	s_page = pfn_to_page(src_pfn);
+	d_page = pfn_to_page(dst_pfn);
+	if (PageHighMem(s_page)) {
+		src = kmap_atomic(s_page, KM_USER0);
+		dst = kmap_atomic(d_page, KM_USER1);
+		do_copy_page(dst, src);
+		kunmap_atomic(src, KM_USER0);
+		kunmap_atomic(dst, KM_USER1);
+	} else {
+		src = page_address(s_page);
+		if (PageHighMem(d_page)) {
+			/* Page pointed to by src may contain some kernel
+			 * data modified by kmap_atomic()
+			 */
+			do_copy_page(buffer, src);
+			dst = kmap_atomic(pfn_to_page(dst_pfn), KM_USER0);
+			memcpy(dst, buffer, PAGE_SIZE);
+			kunmap_atomic(dst, KM_USER0);
+		} else {
+			dst = page_address(d_page);
+			do_copy_page(dst, src);
+		}
+	}
+}
+#else
+#define page_is_saveable(zone, pfn)	saveable_page(pfn)
+
+static inline void
+copy_data_page(unsigned long dst_pfn, unsigned long src_pfn)
+{
+	do_copy_page(page_address(pfn_to_page(dst_pfn)),
+			page_address(pfn_to_page(src_pfn)));
+}
+#endif /* CONFIG_HIGHMEM */
+
+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;
+
+		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 (page_is_saveable(zone, 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))
+			copy_data_page(memory_bm_next_pfn(copy_bm), pfn);
+	} while (pfn != BM_END_OF_MAP);
+}
+
+/* Total number of image pages */
+static unsigned int nr_copy_pages;
+/* Number of pages needed for saving the original pfns of the image pages */
+static unsigned int nr_meta_pages;
+
+/**
+ *	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 pfn, max_zone_pfn;
+
+	for_each_zone(zone) {
+		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);
+					__free_page(page);
+				}
+			}
+	}
+	nr_copy_pages = 0;
+	nr_meta_pages = 0;
+	restore_pblist = NULL;
+	buffer = NULL;
+}
+
+#ifdef CONFIG_HIGHMEM
+/**
+  *	count_pages_for_highmem - compute the number of non-highmem pages
+  *	that will be necessary for creating copies of highmem pages.
+  */
+
+static unsigned int count_pages_for_highmem(unsigned int nr_highmem)
+{
+	unsigned int free_highmem = count_free_highmem_pages();
+
+	if (free_highmem >= nr_highmem)
+		nr_highmem = 0;
+	else
+		nr_highmem -= free_highmem;
+
+	return nr_highmem;
+}
+#else
+static unsigned int
+count_pages_for_highmem(unsigned int nr_highmem) { return 0; }
+#endif /* CONFIG_HIGHMEM */
+
+/**
+ *	enough_free_mem - Make sure we have enough free memory for the
+ *	snapshot image.
+ */
+
+static int enough_free_mem(unsigned int nr_pages, unsigned int nr_highmem)
+{
+	struct zone *zone;
+	unsigned int free = 0, meta = 0;
+
+	for_each_zone(zone) {
+		meta += snapshot_additional_pages(zone);
+		if (!is_highmem(zone))
+			free += zone->free_pages;
+	}
+
+	nr_pages += count_pages_for_highmem(nr_highmem);
+	pr_debug("swsusp: Normal pages needed: %u + %u + %u, available pages: %u\n",
+		nr_pages, PAGES_FOR_IO, meta, free);
+
+	return free > nr_pages + PAGES_FOR_IO + meta;
+}
+
+#ifdef CONFIG_HIGHMEM
+/**
+ *	get_highmem_buffer - if there are some highmem pages in the suspend
+ *	image, we may need the buffer to copy them and/or load their data.
+ */
+
+static inline int get_highmem_buffer(int safe_needed)
+{
+	buffer = get_image_page(GFP_ATOMIC | __GFP_COLD, safe_needed);
+	return buffer ? 0 : -ENOMEM;
+}
+
+/**
+ *	alloc_highmem_image_pages - allocate some highmem pages for the image.
+ *	Try to allocate as many pages as needed, but if the number of free
+ *	highmem pages is lesser than that, allocate them all.
+ */
+
+static inline unsigned int
+alloc_highmem_image_pages(struct memory_bitmap *bm, unsigned int nr_highmem)
+{
+	unsigned int to_alloc = count_free_highmem_pages();
+
+	if (to_alloc > nr_highmem)
+		to_alloc = nr_highmem;
+
+	nr_highmem -= to_alloc;
+	while (to_alloc-- > 0) {
+		struct page *page;
+
+		page = alloc_image_page(__GFP_HIGHMEM);
+		memory_bm_set_bit(bm, page_to_pfn(page));
+	}
+	return nr_highmem;
+}
+#else
+static inline int get_highmem_buffer(int safe_needed) { return 0; }
+
+static inline unsigned int
+alloc_highmem_image_pages(struct memory_bitmap *bm, unsigned int n) { return 0; }
+#endif /* CONFIG_HIGHMEM */
+
+/**
+ *	swsusp_alloc - allocate memory for the suspend image
+ *
+ *	We first try to allocate as many highmem pages as there are
+ *	saveable highmem pages in the system.  If that fails, we allocate
+ *	non-highmem pages for the copies of the remaining highmem ones.
+ *
+ *	In this approach it is likely that the copies of highmem pages will
+ *	also be located in the high memory, because of the way in which
+ *	copy_data_pages() works.
+ */
+
+static int
+swsusp_alloc(struct memory_bitmap *orig_bm, struct memory_bitmap *copy_bm,
+		unsigned int nr_pages, unsigned int nr_highmem)
+{
+	int error;
+
+	error = memory_bm_create(orig_bm, GFP_ATOMIC | __GFP_COLD, PG_ANY);
+	if (error)
+		goto Free;
+
+	error = memory_bm_create(copy_bm, GFP_ATOMIC | __GFP_COLD, PG_ANY);
+	if (error)
+		goto Free;
+
+	if (nr_highmem > 0) {
+		error = get_highmem_buffer(PG_ANY);
+		if (error)
+			goto Free;
+
+		nr_pages += alloc_highmem_image_pages(copy_bm, nr_highmem);
+	}
+	while (nr_pages-- > 0) {
+		struct page *page = alloc_image_page(GFP_ATOMIC | __GFP_COLD);
+
+		if (!page)
+			goto Free;
+
+		memory_bm_set_bit(copy_bm, page_to_pfn(page));
+	}
+	return 0;
+
+ Free:
+	swsusp_free();
+	return -ENOMEM;
+}
+
+/* Memory bitmap used for marking saveable pages (during suspend) or the
+ * suspend image pages (during resume)
+ */
+static struct memory_bitmap orig_bm;
+/* Memory bitmap used on suspend for marking allocated pages that will contain
+ * the copies of saveable pages.  During resume it is initially used for
+ * marking the suspend image pages, but then its set bits are duplicated in
+ * @orig_bm and it is released.  Next, on systems with high memory, it may be
+ * used for marking "safe" highmem pages, but it has to be reinitialized for
+ * this purpose.
+ */
+static struct memory_bitmap copy_bm;
+
+asmlinkage int swsusp_save(void)
+{
+	unsigned int nr_pages, nr_highmem;
+
+	printk("swsusp: critical section: \n");
+
+	drain_local_pages();
+	nr_pages = count_data_pages();
+	nr_highmem = count_highmem_pages();
+	printk("swsusp: Need to copy %u pages\n", nr_pages + nr_highmem);
+
+	if (!enough_free_mem(nr_pages, nr_highmem)) {
+		printk(KERN_ERR "swsusp: Not enough free memory\n");
+		return -ENOMEM;
+	}
+
+	if (swsusp_alloc(&orig_bm, &copy_bm, nr_pages, nr_highmem)) {
+		printk(KERN_ERR "swsusp: Memory allocation failed\n");
+		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,
@@ -503,8 +1015,725 @@ asmlinkage int swsusp_save(void)
 	 * touch swap space! Except we must write out our image of course.
 	 */
 
+	nr_pages += nr_highmem;
 	nr_copy_pages = nr_pages;
+	nr_meta_pages = DIV_ROUND_UP(nr_pages * sizeof(long), PAGE_SIZE);
 
 	printk("swsusp: critical section/: done (%d pages copied)\n", nr_pages);
+
+	return 0;
+}
+
+static void init_header(struct swsusp_info *info)
+{
+	memset(info, 0, sizeof(struct swsusp_info));
+	info->version_code = LINUX_VERSION_CODE;
+	info->num_physpages = num_physpages;
+	memcpy(&info->uts, init_utsname(), sizeof(struct new_utsname));
+	info->cpus = num_online_cpus();
+	info->image_pages = nr_copy_pages;
+	info->pages = nr_copy_pages + nr_meta_pages + 1;
+	info->size = info->pages;
+	info->size <<= PAGE_SHIFT;
+}
+
+/**
+ *	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 void
+pack_pfns(unsigned long *buf, struct memory_bitmap *bm)
+{
+	int j;
+
+	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;
+	}
+}
+
+/**
+ *	snapshot_read_next - used for reading the system memory snapshot.
+ *
+ *	On the first call to it @handle should point to a zeroed
+ *	snapshot_handle structure.  The structure gets updated and a pointer
+ *	to it should be passed to this function every next time.
+ *
+ *	The @count parameter should contain the number of bytes the caller
+ *	wants to read from the snapshot.  It must not be zero.
+ *
+ *	On success the function returns a positive number.  Then, the caller
+ *	is allowed to read up to the returned number of bytes from the memory
+ *	location computed by the data_of() macro.  The number returned
+ *	may be smaller than @count, but this only happens if the read would
+ *	cross a page boundary otherwise.
+ *
+ *	The function returns 0 to indicate the end of data stream condition,
+ *	and a negative number is returned on error.  In such cases the
+ *	structure pointed to by @handle is not updated and should not be used
+ *	any more.
+ */
+
+int snapshot_read_next(struct snapshot_handle *handle, size_t count)
+{
+	if (handle->cur > nr_meta_pages + nr_copy_pages)
+		return 0;
+
+	if (!buffer) {
+		/* This makes the buffer be freed by swsusp_free() */
+		buffer = get_image_page(GFP_ATOMIC, PG_ANY);
+		if (!buffer)
+			return -ENOMEM;
+	}
+	if (!handle->offset) {
+		init_header((struct swsusp_info *)buffer);
+		handle->buffer = buffer;
+		memory_bm_position_reset(&orig_bm);
+		memory_bm_position_reset(&copy_bm);
+	}
+	if (handle->prev < handle->cur) {
+		if (handle->cur <= nr_meta_pages) {
+			memset(buffer, 0, PAGE_SIZE);
+			pack_pfns(buffer, &orig_bm);
+		} else {
+			struct page *page;
+
+			page = pfn_to_page(memory_bm_next_pfn(&copy_bm));
+			if (PageHighMem(page)) {
+				/* Highmem pages are copied to the buffer,
+				 * because we can't return with a kmapped
+				 * highmem page (we may not be called again).
+				 */
+				void *kaddr;
+
+				kaddr = kmap_atomic(page, KM_USER0);
+				memcpy(buffer, kaddr, PAGE_SIZE);
+				kunmap_atomic(kaddr, KM_USER0);
+				handle->buffer = buffer;
+			} else {
+				handle->buffer = page_address(page);
+			}
+		}
+		handle->prev = handle->cur;
+	}
+	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->cur_offset += count;
+	}
+	handle->offset += count;
+	return count;
+}
+
+/**
+ *	mark_unsafe_pages - mark the pages that cannot be used for storing
+ *	the image during resume, because they conflict with the pages that
+ *	had been used before suspend
+ */
+
+static int mark_unsafe_pages(struct memory_bitmap *bm)
+{
+	struct zone *zone;
+	unsigned long pfn, max_zone_pfn;
+
+	/* Clear page flags */
+	for_each_zone(zone) {
+		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 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);
+
+	allocated_unsafe_pages = 0;
+
+	return 0;
+}
+
+static void
+duplicate_memory_bitmap(struct memory_bitmap *dst, struct memory_bitmap *src)
+{
+	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 inline int check_header(struct swsusp_info *info)
+{
+	char *reason = NULL;
+
+	if (info->version_code != LINUX_VERSION_CODE)
+		reason = "kernel version";
+	if (info->num_physpages != num_physpages)
+		reason = "memory size";
+	if (strcmp(info->uts.sysname,init_utsname()->sysname))
+		reason = "system type";
+	if (strcmp(info->uts.release,init_utsname()->release))
+		reason = "kernel release";
+	if (strcmp(info->uts.version,init_utsname()->version))
+		reason = "version";
+	if (strcmp(info->uts.machine,init_utsname()->machine))
+		reason = "machine";
+	if (reason) {
+		printk(KERN_ERR "swsusp: Resume mismatch: %s\n", reason);
+		return -EPERM;
+	}
+	return 0;
+}
+
+/**
+ *	load header - check the image header and copy data from it
+ */
+
+static int
+load_header(struct swsusp_info *info)
+{
+	int error;
+
+	restore_pblist = NULL;
+	error = check_header(info);
+	if (!error) {
+		nr_copy_pages = info->image_pages;
+		nr_meta_pages = info->pages - info->image_pages - 1;
+	}
+	return error;
+}
+
+/**
+ *	unpack_orig_pfns - for each element of @buf[] (1 page at a time) set
+ *	the corresponding bit in the memory bitmap @bm
+ */
+
+static inline void
+unpack_orig_pfns(unsigned long *buf, struct memory_bitmap *bm)
+{
+	int j;
+
+	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]);
+	}
+}
+
+/* List of "safe" pages that may be used to store data loaded from the suspend
+ * image
+ */
+static struct linked_page *safe_pages_list;
+
+#ifdef CONFIG_HIGHMEM
+/* struct highmem_pbe is used for creating the list of highmem pages that
+ * should be restored atomically during the resume from disk, because the page
+ * frames they have occupied before the suspend are in use.
+ */
+struct highmem_pbe {
+	struct page *copy_page;	/* data is here now */
+	struct page *orig_page;	/* data was here before the suspend */
+	struct highmem_pbe *next;
+};
+
+/* List of highmem PBEs needed for restoring the highmem pages that were
+ * allocated before the suspend and included in the suspend image, but have
+ * also been allocated by the "resume" kernel, so their contents cannot be
+ * written directly to their "original" page frames.
+ */
+static struct highmem_pbe *highmem_pblist;
+
+/**
+ *	count_highmem_image_pages - compute the number of highmem pages in the
+ *	suspend image.  The bits in the memory bitmap @bm that correspond to the
+ *	image pages are assumed to be set.
+ */
+
+static unsigned int count_highmem_image_pages(struct memory_bitmap *bm)
+{
+	unsigned long pfn;
+	unsigned int cnt = 0;
+
+	memory_bm_position_reset(bm);
+	pfn = memory_bm_next_pfn(bm);
+	while (pfn != BM_END_OF_MAP) {
+		if (PageHighMem(pfn_to_page(pfn)))
+			cnt++;
+
+		pfn = memory_bm_next_pfn(bm);
+	}
+	return cnt;
+}
+
+/**
+ *	prepare_highmem_image - try to allocate as many highmem pages as
+ *	there are highmem image pages (@nr_highmem_p points to the variable
+ *	containing the number of highmem image pages).  The pages that are
+ *	"safe" (ie. will not be overwritten when the suspend image is
+ *	restored) have the corresponding bits set in @bm (it must be
+ *	unitialized).
+ *
+ *	NOTE: This function should not be called if there are no highmem
+ *	image pages.
+ */
+
+static unsigned int safe_highmem_pages;
+
+static struct memory_bitmap *safe_highmem_bm;
+
+static int
+prepare_highmem_image(struct memory_bitmap *bm, unsigned int *nr_highmem_p)
+{
+	unsigned int to_alloc;
+
+	if (memory_bm_create(bm, GFP_ATOMIC, PG_SAFE))
+		return -ENOMEM;
+
+	if (get_highmem_buffer(PG_SAFE))
+		return -ENOMEM;
+
+	to_alloc = count_free_highmem_pages();
+	if (to_alloc > *nr_highmem_p)
+		to_alloc = *nr_highmem_p;
+	else
+		*nr_highmem_p = to_alloc;
+
+	safe_highmem_pages = 0;
+	while (to_alloc-- > 0) {
+		struct page *page;
+
+		page = alloc_page(__GFP_HIGHMEM);
+		if (!PageNosaveFree(page)) {
+			/* The page is "safe", set its bit the bitmap */
+			memory_bm_set_bit(bm, page_to_pfn(page));
+			safe_highmem_pages++;
+		}
+		/* Mark the page as allocated */
+		SetPageNosave(page);
+		SetPageNosaveFree(page);
+	}
+	memory_bm_position_reset(bm);
+	safe_highmem_bm = bm;
+	return 0;
+}
+
+/**
+ *	get_highmem_page_buffer - for given highmem image page find the buffer
+ *	that suspend_write_next() should set for its caller to write to.
+ *
+ *	If the page is to be saved to its "original" page frame or a copy of
+ *	the page is to be made in the highmem, @buffer is returned.  Otherwise,
+ *	the copy of the page is to be made in normal memory, so the address of
+ *	the copy is returned.
+ *
+ *	If @buffer is returned, the caller of suspend_write_next() will write
+ *	the page's contents to @buffer, so they will have to be copied to the
+ *	right location on the next call to suspend_write_next() and it is done
+ *	with the help of copy_last_highmem_page().  For this purpose, if
+ *	@buffer is returned, @last_highmem page is set to the page to which
+ *	the data will have to be copied from @buffer.
+ */
+
+static struct page *last_highmem_page;
+
+static void *
+get_highmem_page_buffer(struct page *page, struct chain_allocator *ca)
+{
+	struct highmem_pbe *pbe;
+	void *kaddr;
+
+	if (PageNosave(page) && PageNosaveFree(page)) {
+		/* We have allocated the "original" page frame and we can
+		 * use it directly to store the loaded page.
+		 */
+		last_highmem_page = page;
+		return buffer;
+	}
+	/* The "original" page frame has not been allocated and we have to
+	 * use a "safe" page frame to store the loaded page.
+	 */
+	pbe = chain_alloc(ca, sizeof(struct highmem_pbe));
+	if (!pbe) {
+		swsusp_free();
+		return NULL;
+	}
+	pbe->orig_page = page;
+	if (safe_highmem_pages > 0) {
+		struct page *tmp;
+
+		/* Copy of the page will be stored in high memory */
+		kaddr = buffer;
+		tmp = pfn_to_page(memory_bm_next_pfn(safe_highmem_bm));
+		safe_highmem_pages--;
+		last_highmem_page = tmp;
+		pbe->copy_page = tmp;
+	} else {
+		/* Copy of the page will be stored in normal memory */
+		kaddr = safe_pages_list;
+		safe_pages_list = safe_pages_list->next;
+		pbe->copy_page = virt_to_page(kaddr);
+	}
+	pbe->next = highmem_pblist;
+	highmem_pblist = pbe;
+	return kaddr;
+}
+
+/**
+ *	copy_last_highmem_page - copy the contents of a highmem image from
+ *	@buffer, where the caller of snapshot_write_next() has place them,
+ *	to the right location represented by @last_highmem_page .
+ */
+
+static void copy_last_highmem_page(void)
+{
+	if (last_highmem_page) {
+		void *dst;
+
+		dst = kmap_atomic(last_highmem_page, KM_USER0);
+		memcpy(dst, buffer, PAGE_SIZE);
+		kunmap_atomic(dst, KM_USER0);
+		last_highmem_page = NULL;
+	}
+}
+
+static inline int last_highmem_page_copied(void)
+{
+	return !last_highmem_page;
+}
+
+static inline void free_highmem_data(void)
+{
+	if (safe_highmem_bm)
+		memory_bm_free(safe_highmem_bm, PG_UNSAFE_CLEAR);
+
+	if (buffer)
+		free_image_page(buffer, PG_UNSAFE_CLEAR);
+}
+#else
+static inline int get_safe_write_buffer(void) { return 0; }
+
+static unsigned int
+count_highmem_image_pages(struct memory_bitmap *bm) { return 0; }
+
+static inline int
+prepare_highmem_image(struct memory_bitmap *bm, unsigned int *nr_highmem_p)
+{
+	return 0;
+}
+
+static inline void *
+get_highmem_page_buffer(struct page *page, struct chain_allocator *ca)
+{
+	return NULL;
+}
+
+static inline void copy_last_highmem_page(void) {}
+static inline int last_highmem_page_copied(void) { return 1; }
+static inline void free_highmem_data(void) {}
+#endif /* CONFIG_HIGHMEM */
+
+/**
+ *	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 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 lists of "safe" pages that will be used
+ *	later.  On systems with high memory a list of "safe" highmem pages is
+ *	also created.
+ */
+
+#define PBES_PER_LINKED_PAGE	(LINKED_PAGE_DATA_SIZE / sizeof(struct pbe))
+
+static int
+prepare_image(struct memory_bitmap *new_bm, struct memory_bitmap *bm)
+{
+	unsigned int nr_pages, nr_highmem;
+	struct linked_page *sp_list, *lp;
+	int error;
+
+	/* If there is no highmem, the buffer will not be necessary */
+	free_image_page(buffer, PG_UNSAFE_CLEAR);
+	buffer = NULL;
+
+	nr_highmem = count_highmem_image_pages(bm);
+	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);
+	if (nr_highmem > 0) {
+		error = prepare_highmem_image(bm, &nr_highmem);
+		if (error)
+			goto Free;
+	}
+	/* 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 - nr_highmem - allocated_unsafe_pages;
+	nr_pages = DIV_ROUND_UP(nr_pages, PBES_PER_LINKED_PAGE);
+	while (nr_pages > 0) {
+		lp = get_image_page(GFP_ATOMIC, PG_SAFE);
+		if (!lp) {
+			error = -ENOMEM;
+			goto Free;
+		}
+		lp->next = sp_list;
+		sp_list = lp;
+		nr_pages--;
+	}
+	/* Preallocate memory for the image */
+	safe_pages_list = NULL;
+	nr_pages = nr_copy_pages - nr_highmem - 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--;
+	}
+	/* 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;
+}
+
+/**
+ *	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;
+	struct page *page = pfn_to_page(memory_bm_next_pfn(bm));
+
+	if (PageHighMem(page))
+		return get_highmem_page_buffer(page, ca);
+
+	if (PageNosave(page) && PageNosaveFree(page))
+		/* We have allocated the "original" page frame and we can
+		 * use it directly to store the loaded 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 = chain_alloc(ca, sizeof(struct pbe));
+	if (!pbe) {
+		swsusp_free();
+		return NULL;
+	}
+	pbe->orig_address = page_address(page);
+	pbe->address = safe_pages_list;
+	safe_pages_list = safe_pages_list->next;
+	pbe->next = restore_pblist;
+	restore_pblist = pbe;
+	return pbe->address;
+}
+
+/**
+ *	snapshot_write_next - used for writing the system memory snapshot.
+ *
+ *	On the first call to it @handle should point to a zeroed
+ *	snapshot_handle structure.  The structure gets updated and a pointer
+ *	to it should be passed to this function every next time.
+ *
+ *	The @count parameter should contain the number of bytes the caller
+ *	wants to write to the image.  It must not be zero.
+ *
+ *	On success the function returns a positive number.  Then, the caller
+ *	is allowed to write up to the returned number of bytes to the memory
+ *	location computed by the data_of() macro.  The number returned
+ *	may be smaller than @count, but this only happens if the write would
+ *	cross a page boundary otherwise.
+ *
+ *	The function returns 0 to indicate the "end of file" condition,
+ *	and a negative number is returned on error.  In such cases the
+ *	structure pointed to by @handle is not updated and should not be used
+ *	any more.
+ */
+
+int snapshot_write_next(struct snapshot_handle *handle, size_t count)
+{
+	static struct chain_allocator ca;
+	int error = 0;
+
+	/* Check if we have already loaded the entire image */
+	if (handle->prev && handle->cur > nr_meta_pages + nr_copy_pages)
+		return 0;
+
+	if (handle->offset == 0) {
+		if (!buffer)
+			/* This makes the buffer be freed by swsusp_free() */
+			buffer = get_image_page(GFP_ATOMIC, PG_ANY);
+
+		if (!buffer)
+			return -ENOMEM;
+
+		handle->buffer = 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) {
+			unpack_orig_pfns(buffer, &copy_bm);
+			if (handle->prev == nr_meta_pages) {
+				error = prepare_image(&orig_bm, &copy_bm);
+				if (error)
+					return error;
+
+				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 {
+			copy_last_highmem_page();
+			handle->buffer = get_buffer(&orig_bm, &ca);
+			if (handle->buffer != buffer)
+				handle->sync_read = 0;
+		}
+		handle->prev = handle->cur;
+	}
+	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->cur_offset += count;
+	}
+	handle->offset += count;
+	return count;
+}
+
+/**
+ *	snapshot_write_finalize - must be called after the last call to
+ *	snapshot_write_next() in case the last page in the image happens
+ *	to be a highmem page and its contents should be stored in the
+ *	highmem.  Additionally, it releases the memory that will not be
+ *	used any more.
+ */
+
+void snapshot_write_finalize(struct snapshot_handle *handle)
+{
+	copy_last_highmem_page();
+	/* 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);
+		free_highmem_data();
+	}
+}
+
+int snapshot_image_loaded(struct snapshot_handle *handle)
+{
+	return !(!nr_copy_pages || !last_highmem_page_copied() ||
+			handle->cur <= nr_meta_pages + nr_copy_pages);
+}
+
+#ifdef CONFIG_HIGHMEM
+/* Assumes that @buf is ready and points to a "safe" page */
+static inline void
+swap_two_pages_data(struct page *p1, struct page *p2, void *buf)
+{
+	void *kaddr1, *kaddr2;
+
+	kaddr1 = kmap_atomic(p1, KM_USER0);
+	kaddr2 = kmap_atomic(p2, KM_USER1);
+	memcpy(buf, kaddr1, PAGE_SIZE);
+	memcpy(kaddr1, kaddr2, PAGE_SIZE);
+	memcpy(kaddr2, buf, PAGE_SIZE);
+	kunmap_atomic(kaddr1, KM_USER0);
+	kunmap_atomic(kaddr2, KM_USER1);
+}
+
+/**
+ *	restore_highmem - for each highmem page that was allocated before
+ *	the suspend and included in the suspend image, and also has been
+ *	allocated by the "resume" kernel swap its current (ie. "before
+ *	resume") contents with the previous (ie. "before suspend") one.
+ *
+ *	If the resume eventually fails, we can call this function once
+ *	again and restore the "before resume" highmem state.
+ */
+
+int restore_highmem(void)
+{
+	struct highmem_pbe *pbe = highmem_pblist;
+	void *buf;
+
+	if (!pbe)
+		return 0;
+
+	buf = get_image_page(GFP_ATOMIC, PG_SAFE);
+	if (!buf)
+		return -ENOMEM;
+
+	while (pbe) {
+		swap_two_pages_data(pbe->copy_page, pbe->orig_page, buf);
+		pbe = pbe->next;
+	}
+	free_image_page(buf, PG_UNSAFE_CLEAR);
 	return 0;
 }
+#endif /* CONFIG_HIGHMEM */