X-Git-Url: http://ftp.safe.ca/?a=blobdiff_plain;f=mm%2Fswapfile.c;h=fa5f10b9c28b0b6366f742c10c72b91f03374a71;hb=73c34b6accc8427584f5d7db4d5acb230ed8c912;hp=7e6304dfafab174884bad5820a0194f449906a32;hpb=85d9fc89fb0f0703df6444f260187c088a8d59ff;p=safe%2Fjmp%2Flinux-2.6

diff --git a/mm/swapfile.c b/mm/swapfile.c
index 7e6304d..fa5f10b 100644
--- a/mm/swapfile.c
+++ b/mm/swapfile.c
@@ -49,10 +49,62 @@ static const char Unused_offset[] = "Unused swap offset entry ";
 
 static struct swap_list_t swap_list = {-1, -1};
 
-static struct swap_info_struct swap_info[MAX_SWAPFILES];
+static struct swap_info_struct *swap_info[MAX_SWAPFILES];
 
 static DEFINE_MUTEX(swapon_mutex);
 
+/* For reference count accounting in swap_map */
+/* enum for swap_map[] handling. internal use only */
+enum {
+	SWAP_MAP = 0,	/* ops for reference from swap users */
+	SWAP_CACHE,	/* ops for reference from swap cache */
+};
+
+static inline int swap_count(unsigned short ent)
+{
+	return ent & SWAP_COUNT_MASK;
+}
+
+static inline bool swap_has_cache(unsigned short ent)
+{
+	return !!(ent & SWAP_HAS_CACHE);
+}
+
+static inline unsigned short encode_swapmap(int count, bool has_cache)
+{
+	unsigned short ret = count;
+
+	if (has_cache)
+		return SWAP_HAS_CACHE | ret;
+	return ret;
+}
+
+/* returns 1 if swap entry is freed */
+static int
+__try_to_reclaim_swap(struct swap_info_struct *si, unsigned long offset)
+{
+	swp_entry_t entry = swp_entry(si->type, offset);
+	struct page *page;
+	int ret = 0;
+
+	page = find_get_page(&swapper_space, entry.val);
+	if (!page)
+		return 0;
+	/*
+	 * This function is called from scan_swap_map() and it's called
+	 * by vmscan.c at reclaiming pages. So, we hold a lock on a page, here.
+	 * We have to use trylock for avoiding deadlock. This is a special
+	 * case and you should use try_to_free_swap() with explicit lock_page()
+	 * in usual operations.
+	 */
+	if (trylock_page(page)) {
+		ret = try_to_free_swap(page);
+		unlock_page(page);
+	}
+	page_cache_release(page);
+	return ret;
+}
+
 /*
  * We need this because the bdev->unplug_fn can sleep and we cannot
  * hold swap_lock while calling the unplug_fn. And swap_lock
@@ -67,7 +119,7 @@ void swap_unplug_io_fn(struct backing_dev_info *unused_bdi, struct page *page)
 	down_read(&swap_unplug_sem);
 	entry.val = page_private(page);
 	if (PageSwapCache(page)) {
-		struct block_device *bdev = swap_info[swp_type(entry)].bdev;
+		struct block_device *bdev = swap_info[swp_type(entry)]->bdev;
 		struct backing_dev_info *bdi;
 
 		/*
@@ -93,22 +145,28 @@ void swap_unplug_io_fn(struct backing_dev_info *unused_bdi, struct page *page)
 static int discard_swap(struct swap_info_struct *si)
 {
 	struct swap_extent *se;
+	sector_t start_block;
+	sector_t nr_blocks;
 	int err = 0;
 
-	list_for_each_entry(se, &si->extent_list, list) {
-		sector_t start_block = se->start_block << (PAGE_SHIFT - 9);
-		sector_t nr_blocks = (sector_t)se->nr_pages << (PAGE_SHIFT - 9);
+	/* Do not discard the swap header page! */
+	se = &si->first_swap_extent;
+	start_block = (se->start_block + 1) << (PAGE_SHIFT - 9);
+	nr_blocks = ((sector_t)se->nr_pages - 1) << (PAGE_SHIFT - 9);
+	if (nr_blocks) {
+		err = blkdev_issue_discard(si->bdev, start_block,
+				nr_blocks, GFP_KERNEL, DISCARD_FL_BARRIER);
+		if (err)
+			return err;
+		cond_resched();
+	}
 
-		if (se->start_page == 0) {
-			/* Do not discard the swap header page! */
-			start_block += 1 << (PAGE_SHIFT - 9);
-			nr_blocks -= 1 << (PAGE_SHIFT - 9);
-			if (!nr_blocks)
-				continue;
-		}
+	list_for_each_entry(se, &si->first_swap_extent.list, list) {
+		start_block = se->start_block << (PAGE_SHIFT - 9);
+		nr_blocks = (sector_t)se->nr_pages << (PAGE_SHIFT - 9);
 
 		err = blkdev_issue_discard(si->bdev, start_block,
-						nr_blocks, GFP_KERNEL);
+				nr_blocks, GFP_KERNEL, DISCARD_FL_BARRIER);
 		if (err)
 			break;
 
@@ -147,13 +205,11 @@ static void discard_swap_cluster(struct swap_info_struct *si,
 			start_block <<= PAGE_SHIFT - 9;
 			nr_blocks <<= PAGE_SHIFT - 9;
 			if (blkdev_issue_discard(si->bdev, start_block,
-							nr_blocks, GFP_NOIO))
+				    nr_blocks, GFP_NOIO, DISCARD_FL_BARRIER))
 				break;
 		}
 
 		lh = se->list.next;
-		if (lh == &si->extent_list)
-			lh = lh->next;
 		se = list_entry(lh, struct swap_extent, list);
 	}
 }
@@ -167,7 +223,8 @@ static int wait_for_discard(void *word)
 #define SWAPFILE_CLUSTER	256
 #define LATENCY_LIMIT		256
 
-static inline unsigned long scan_swap_map(struct swap_info_struct *si)
+static inline unsigned long scan_swap_map(struct swap_info_struct *si,
+					  int cache)
 {
 	unsigned long offset;
 	unsigned long scan_base;
@@ -273,6 +330,19 @@ checks:
 		goto no_page;
 	if (offset > si->highest_bit)
 		scan_base = offset = si->lowest_bit;
+
+	/* reuse swap entry of cache-only swap if not busy. */
+	if (vm_swap_full() && si->swap_map[offset] == SWAP_HAS_CACHE) {
+		int swap_was_freed;
+		spin_unlock(&swap_lock);
+		swap_was_freed = __try_to_reclaim_swap(si, offset);
+		spin_lock(&swap_lock);
+		/* entry was freed successfully, try to use this again */
+		if (swap_was_freed)
+			goto checks;
+		goto scan; /* check next one */
+	}
+
 	if (si->swap_map[offset])
 		goto scan;
 
@@ -285,7 +355,10 @@ checks:
 		si->lowest_bit = si->max;
 		si->highest_bit = 0;
 	}
-	si->swap_map[offset] = 1;
+	if (cache == SWAP_CACHE) /* at usual swap-out via vmscan.c */
+		si->swap_map[offset] = encode_swapmap(0, true);
+	else /* at suspend */
+		si->swap_map[offset] = encode_swapmap(1, false);
 	si->cluster_next = offset + 1;
 	si->flags -= SWP_SCANNING;
 
@@ -351,6 +424,10 @@ scan:
 			spin_lock(&swap_lock);
 			goto checks;
 		}
+		if (vm_swap_full() && si->swap_map[offset] == SWAP_HAS_CACHE) {
+			spin_lock(&swap_lock);
+			goto checks;
+		}
 		if (unlikely(--latency_ration < 0)) {
 			cond_resched();
 			latency_ration = LATENCY_LIMIT;
@@ -362,6 +439,10 @@ scan:
 			spin_lock(&swap_lock);
 			goto checks;
 		}
+		if (vm_swap_full() && si->swap_map[offset] == SWAP_HAS_CACHE) {
+			spin_lock(&swap_lock);
+			goto checks;
+		}
 		if (unlikely(--latency_ration < 0)) {
 			cond_resched();
 			latency_ration = LATENCY_LIMIT;
@@ -387,10 +468,10 @@ swp_entry_t get_swap_page(void)
 	nr_swap_pages--;
 
 	for (type = swap_list.next; type >= 0 && wrapped < 2; type = next) {
-		si = swap_info + type;
+		si = swap_info[type];
 		next = si->next;
 		if (next < 0 ||
-		    (!wrapped && si->prio != swap_info[next].prio)) {
+		    (!wrapped && si->prio != swap_info[next]->prio)) {
 			next = swap_list.head;
 			wrapped++;
 		}
@@ -401,7 +482,8 @@ swp_entry_t get_swap_page(void)
 			continue;
 
 		swap_list.next = next;
-		offset = scan_swap_map(si);
+		/* This is called for allocating swap entry for cache */
+		offset = scan_swap_map(si, SWAP_CACHE);
 		if (offset) {
 			spin_unlock(&swap_lock);
 			return swp_entry(type, offset);
@@ -415,16 +497,18 @@ noswap:
 	return (swp_entry_t) {0};
 }
 
+/* The only caller of this function is now susupend routine */
 swp_entry_t get_swap_page_of_type(int type)
 {
 	struct swap_info_struct *si;
 	pgoff_t offset;
 
 	spin_lock(&swap_lock);
-	si = swap_info + type;
-	if (si->flags & SWP_WRITEOK) {
+	si = swap_info[type];
+	if (si && (si->flags & SWP_WRITEOK)) {
 		nr_swap_pages--;
-		offset = scan_swap_map(si);
+		/* This is called for allocating swap entry, not cache */
+		offset = scan_swap_map(si, SWAP_MAP);
 		if (offset) {
 			spin_unlock(&swap_lock);
 			return swp_entry(type, offset);
@@ -435,9 +519,9 @@ swp_entry_t get_swap_page_of_type(int type)
 	return (swp_entry_t) {0};
 }
 
-static struct swap_info_struct * swap_info_get(swp_entry_t entry)
+static struct swap_info_struct *swap_info_get(swp_entry_t entry)
 {
-	struct swap_info_struct * p;
+	struct swap_info_struct *p;
 	unsigned long offset, type;
 
 	if (!entry.val)
@@ -445,7 +529,7 @@ static struct swap_info_struct * swap_info_get(swp_entry_t entry)
 	type = swp_type(entry);
 	if (type >= nr_swapfiles)
 		goto bad_nofile;
-	p = & swap_info[type];
+	p = swap_info[type];
 	if (!(p->flags & SWP_USED))
 		goto bad_device;
 	offset = swp_offset(entry);
@@ -471,26 +555,41 @@ out:
 	return NULL;
 }
 
-static int swap_entry_free(struct swap_info_struct *p, swp_entry_t ent)
+static int swap_entry_free(struct swap_info_struct *p,
+			   swp_entry_t ent, int cache)
 {
 	unsigned long offset = swp_offset(ent);
-	int count = p->swap_map[offset];
-
-	if (count < SWAP_MAP_MAX) {
-		count--;
-		p->swap_map[offset] = count;
-		if (!count) {
-			if (offset < p->lowest_bit)
-				p->lowest_bit = offset;
-			if (offset > p->highest_bit)
-				p->highest_bit = offset;
-			if (p->prio > swap_info[swap_list.next].prio)
-				swap_list.next = p - swap_info;
-			nr_swap_pages++;
-			p->inuse_pages--;
-			mem_cgroup_uncharge_swap(ent);
+	int count = swap_count(p->swap_map[offset]);
+	bool has_cache;
+
+	has_cache = swap_has_cache(p->swap_map[offset]);
+
+	if (cache == SWAP_MAP) { /* dropping usage count of swap */
+		if (count < SWAP_MAP_MAX) {
+			count--;
+			p->swap_map[offset] = encode_swapmap(count, has_cache);
 		}
+	} else { /* dropping swap cache flag */
+		VM_BUG_ON(!has_cache);
+		p->swap_map[offset] = encode_swapmap(count, false);
+
+	}
+	/* return code. */
+	count = p->swap_map[offset];
+	/* free if no reference */
+	if (!count) {
+		if (offset < p->lowest_bit)
+			p->lowest_bit = offset;
+		if (offset > p->highest_bit)
+			p->highest_bit = offset;
+		if (swap_list.next >= 0 &&
+		    p->prio > swap_info[swap_list.next]->prio)
+			swap_list.next = p->type;
+		nr_swap_pages++;
+		p->inuse_pages--;
 	}
+	if (!swap_count(count))
+		mem_cgroup_uncharge_swap(ent);
 	return count;
 }
 
@@ -500,11 +599,34 @@ static int swap_entry_free(struct swap_info_struct *p, swp_entry_t ent)
  */
 void swap_free(swp_entry_t entry)
 {
-	struct swap_info_struct * p;
+	struct swap_info_struct *p;
 
 	p = swap_info_get(entry);
 	if (p) {
-		swap_entry_free(p, entry);
+		swap_entry_free(p, entry, SWAP_MAP);
+		spin_unlock(&swap_lock);
+	}
+}
+
+/*
+ * Called after dropping swapcache to decrease refcnt to swap entries.
+ */
+void swapcache_free(swp_entry_t entry, struct page *page)
+{
+	struct swap_info_struct *p;
+	int ret;
+
+	p = swap_info_get(entry);
+	if (p) {
+		ret = swap_entry_free(p, entry, SWAP_CACHE);
+		if (page) {
+			bool swapout;
+			if (ret)
+				swapout = true; /* the end of swap out */
+			else
+				swapout = false; /* no more swap users! */
+			mem_cgroup_uncharge_swapcache(page, entry, swapout);
+		}
 		spin_unlock(&swap_lock);
 	}
 }
@@ -521,8 +643,7 @@ static inline int page_swapcount(struct page *page)
 	entry.val = page_private(page);
 	p = swap_info_get(entry);
 	if (p) {
-		/* Subtract the 1 for the swap cache itself */
-		count = p->swap_map[swp_offset(entry)] - 1;
+		count = swap_count(p->swap_map[swp_offset(entry)]);
 		spin_unlock(&swap_lock);
 	}
 	return count;
@@ -579,12 +700,12 @@ int free_swap_and_cache(swp_entry_t entry)
 	struct swap_info_struct *p;
 	struct page *page = NULL;
 
-	if (is_migration_entry(entry))
+	if (non_swap_entry(entry))
 		return 1;
 
 	p = swap_info_get(entry);
 	if (p) {
-		if (swap_entry_free(p, entry) == 1) {
+		if (swap_entry_free(p, entry, SWAP_MAP) == SWAP_HAS_CACHE) {
 			page = find_get_page(&swapper_space, entry.val);
 			if (page && !trylock_page(page)) {
 				page_cache_release(page);
@@ -621,37 +742,35 @@ int free_swap_and_cache(swp_entry_t entry)
 int swap_type_of(dev_t device, sector_t offset, struct block_device **bdev_p)
 {
 	struct block_device *bdev = NULL;
-	int i;
+	int type;
 
 	if (device)
 		bdev = bdget(device);
 
 	spin_lock(&swap_lock);
-	for (i = 0; i < nr_swapfiles; i++) {
-		struct swap_info_struct *sis = swap_info + i;
+	for (type = 0; type < nr_swapfiles; type++) {
+		struct swap_info_struct *sis = swap_info[type];
 
 		if (!(sis->flags & SWP_WRITEOK))
 			continue;
 
 		if (!bdev) {
 			if (bdev_p)
-				*bdev_p = sis->bdev;
+				*bdev_p = bdgrab(sis->bdev);
 
 			spin_unlock(&swap_lock);
-			return i;
+			return type;
 		}
 		if (bdev == sis->bdev) {
-			struct swap_extent *se;
+			struct swap_extent *se = &sis->first_swap_extent;
 
-			se = list_entry(sis->extent_list.next,
-					struct swap_extent, list);
 			if (se->start_block == offset) {
 				if (bdev_p)
-					*bdev_p = sis->bdev;
+					*bdev_p = bdgrab(sis->bdev);
 
 				spin_unlock(&swap_lock);
 				bdput(bdev);
-				return i;
+				return type;
 			}
 		}
 	}
@@ -663,6 +782,21 @@ int swap_type_of(dev_t device, sector_t offset, struct block_device **bdev_p)
 }
 
 /*
+ * Get the (PAGE_SIZE) block corresponding to given offset on the swapdev
+ * corresponding to given index in swap_info (swap type).
+ */
+sector_t swapdev_block(int type, pgoff_t offset)
+{
+	struct block_device *bdev;
+
+	if ((unsigned int)type >= nr_swapfiles)
+		return 0;
+	if (!(swap_info[type]->flags & SWP_WRITEOK))
+		return 0;
+	return map_swap_page(swp_entry(type, offset), &bdev);
+}
+
+/*
  * Return either the total number of swap pages of given type, or the number
  * of free pages of that type (depending on @free)
  *
@@ -672,18 +806,20 @@ unsigned int count_swap_pages(int type, int free)
 {
 	unsigned int n = 0;
 
-	if (type < nr_swapfiles) {
-		spin_lock(&swap_lock);
-		if (swap_info[type].flags & SWP_WRITEOK) {
-			n = swap_info[type].pages;
+	spin_lock(&swap_lock);
+	if ((unsigned int)type < nr_swapfiles) {
+		struct swap_info_struct *sis = swap_info[type];
+
+		if (sis->flags & SWP_WRITEOK) {
+			n = sis->pages;
 			if (free)
-				n -= swap_info[type].inuse_pages;
+				n -= sis->inuse_pages;
 		}
-		spin_unlock(&swap_lock);
 	}
+	spin_unlock(&swap_lock);
 	return n;
 }
-#endif
+#endif /* CONFIG_HIBERNATION */
 
 /*
  * No need to decide whether this PTE shares the swap entry with others,
@@ -891,7 +1027,7 @@ static unsigned int find_next_to_unuse(struct swap_info_struct *si,
 			i = 1;
 		}
 		count = si->swap_map[i];
-		if (count && count != SWAP_MAP_BAD)
+		if (count && swap_count(count) != SWAP_MAP_BAD)
 			break;
 	}
 	return i;
@@ -904,7 +1040,7 @@ static unsigned int find_next_to_unuse(struct swap_info_struct *si,
  */
 static int try_to_unuse(unsigned int type)
 {
-	struct swap_info_struct * si = &swap_info[type];
+	struct swap_info_struct *si = swap_info[type];
 	struct mm_struct *start_mm;
 	unsigned short *swap_map;
 	unsigned short swcount;
@@ -995,13 +1131,13 @@ static int try_to_unuse(unsigned int type)
 		 */
 		shmem = 0;
 		swcount = *swap_map;
-		if (swcount > 1) {
+		if (swap_count(swcount)) {
 			if (start_mm == &init_mm)
 				shmem = shmem_unuse(entry, page);
 			else
 				retval = unuse_mm(start_mm, entry, page);
 		}
-		if (*swap_map > 1) {
+		if (swap_count(*swap_map)) {
 			int set_start_mm = (*swap_map >= swcount);
 			struct list_head *p = &start_mm->mmlist;
 			struct mm_struct *new_start_mm = start_mm;
@@ -1011,7 +1147,7 @@ static int try_to_unuse(unsigned int type)
 			atomic_inc(&new_start_mm->mm_users);
 			atomic_inc(&prev_mm->mm_users);
 			spin_lock(&mmlist_lock);
-			while (*swap_map > 1 && !retval && !shmem &&
+			while (swap_count(*swap_map) && !retval && !shmem &&
 					(p = p->next) != &start_mm->mmlist) {
 				mm = list_entry(p, struct mm_struct, mmlist);
 				if (!atomic_inc_not_zero(&mm->mm_users))
@@ -1023,13 +1159,14 @@ static int try_to_unuse(unsigned int type)
 				cond_resched();
 
 				swcount = *swap_map;
-				if (swcount <= 1)
+				if (!swap_count(swcount)) /* any usage ? */
 					;
 				else if (mm == &init_mm) {
 					set_start_mm = 1;
 					shmem = shmem_unuse(entry, page);
 				} else
 					retval = unuse_mm(mm, entry, page);
+
 				if (set_start_mm && *swap_map < swcount) {
 					mmput(new_start_mm);
 					atomic_inc(&mm->mm_users);
@@ -1057,21 +1194,25 @@ static int try_to_unuse(unsigned int type)
 		}
 
 		/*
-		 * How could swap count reach 0x7fff when the maximum
-		 * pid is 0x7fff, and there's no way to repeat a swap
-		 * page within an mm (except in shmem, where it's the
-		 * shared object which takes the reference count)?
-		 * We believe SWAP_MAP_MAX cannot occur in Linux 2.4.
-		 *
+		 * How could swap count reach 0x7ffe ?
+		 * There's no way to repeat a swap page within an mm
+		 * (except in shmem, where it's the shared object which takes
+		 * the reference count)?
+		 * We believe SWAP_MAP_MAX cannot occur.(if occur, unsigned
+		 * short is too small....)
 		 * If that's wrong, then we should worry more about
 		 * exit_mmap() and do_munmap() cases described above:
 		 * we might be resetting SWAP_MAP_MAX too early here.
 		 * We know "Undead"s can happen, they're okay, so don't
 		 * report them; but do report if we reset SWAP_MAP_MAX.
 		 */
-		if (*swap_map == SWAP_MAP_MAX) {
+		/* We might release the lock_page() in unuse_mm(). */
+		if (!PageSwapCache(page) || page_private(page) != entry.val)
+			goto retry;
+
+		if (swap_count(*swap_map) == SWAP_MAP_MAX) {
 			spin_lock(&swap_lock);
-			*swap_map = 1;
+			*swap_map = encode_swapmap(0, true);
 			spin_unlock(&swap_lock);
 			reset_overflow = 1;
 		}
@@ -1089,7 +1230,8 @@ static int try_to_unuse(unsigned int type)
 		 * pages would be incorrect if swap supported "shared
 		 * private" pages, but they are handled by tmpfs files.
 		 */
-		if ((*swap_map > 1) && PageDirty(page) && PageSwapCache(page)) {
+		if (swap_count(*swap_map) &&
+		     PageDirty(page) && PageSwapCache(page)) {
 			struct writeback_control wbc = {
 				.sync_mode = WB_SYNC_NONE,
 			};
@@ -1116,6 +1258,7 @@ static int try_to_unuse(unsigned int type)
 		 * mark page dirty so shrink_page_list will preserve it.
 		 */
 		SetPageDirty(page);
+retry:
 		unlock_page(page);
 		page_cache_release(page);
 
@@ -1143,10 +1286,10 @@ static int try_to_unuse(unsigned int type)
 static void drain_mmlist(void)
 {
 	struct list_head *p, *next;
-	unsigned int i;
+	unsigned int type;
 
-	for (i = 0; i < nr_swapfiles; i++)
-		if (swap_info[i].inuse_pages)
+	for (type = 0; type < nr_swapfiles; type++)
+		if (swap_info[type]->inuse_pages)
 			return;
 	spin_lock(&mmlist_lock);
 	list_for_each_safe(p, next, &init_mm.mmlist)
@@ -1156,12 +1299,22 @@ static void drain_mmlist(void)
 
 /*
  * Use this swapdev's extent info to locate the (PAGE_SIZE) block which
- * corresponds to page offset `offset'.
+ * corresponds to page offset `offset'.  Note that the type of this function
+ * is sector_t, but it returns page offset into the bdev, not sector offset.
  */
-sector_t map_swap_page(struct swap_info_struct *sis, pgoff_t offset)
+sector_t map_swap_page(swp_entry_t entry, struct block_device **bdev)
 {
-	struct swap_extent *se = sis->curr_swap_extent;
-	struct swap_extent *start_se = se;
+	struct swap_info_struct *sis;
+	struct swap_extent *start_se;
+	struct swap_extent *se;
+	pgoff_t offset;
+
+	sis = swap_info[swp_type(entry)];
+	*bdev = sis->bdev;
+
+	offset = swp_offset(entry);
+	start_se = sis->curr_swap_extent;
+	se = start_se;
 
 	for ( ; ; ) {
 		struct list_head *lh;
@@ -1171,40 +1324,21 @@ sector_t map_swap_page(struct swap_info_struct *sis, pgoff_t offset)
 			return se->start_block + (offset - se->start_page);
 		}
 		lh = se->list.next;
-		if (lh == &sis->extent_list)
-			lh = lh->next;
 		se = list_entry(lh, struct swap_extent, list);
 		sis->curr_swap_extent = se;
 		BUG_ON(se == start_se);		/* It *must* be present */
 	}
 }
 
-#ifdef CONFIG_HIBERNATION
-/*
- * Get the (PAGE_SIZE) block corresponding to given offset on the swapdev
- * corresponding to given index in swap_info (swap type).
- */
-sector_t swapdev_block(int swap_type, pgoff_t offset)
-{
-	struct swap_info_struct *sis;
-
-	if (swap_type >= nr_swapfiles)
-		return 0;
-
-	sis = swap_info + swap_type;
-	return (sis->flags & SWP_WRITEOK) ? map_swap_page(sis, offset) : 0;
-}
-#endif /* CONFIG_HIBERNATION */
-
 /*
  * Free all of a swapdev's extent information
  */
 static void destroy_swap_extents(struct swap_info_struct *sis)
 {
-	while (!list_empty(&sis->extent_list)) {
+	while (!list_empty(&sis->first_swap_extent.list)) {
 		struct swap_extent *se;
 
-		se = list_entry(sis->extent_list.next,
+		se = list_entry(sis->first_swap_extent.list.next,
 				struct swap_extent, list);
 		list_del(&se->list);
 		kfree(se);
@@ -1225,8 +1359,15 @@ add_swap_extent(struct swap_info_struct *sis, unsigned long start_page,
 	struct swap_extent *new_se;
 	struct list_head *lh;
 
-	lh = sis->extent_list.prev;	/* The highest page extent */
-	if (lh != &sis->extent_list) {
+	if (start_page == 0) {
+		se = &sis->first_swap_extent;
+		sis->curr_swap_extent = se;
+		se->start_page = 0;
+		se->nr_pages = nr_pages;
+		se->start_block = start_block;
+		return 1;
+	} else {
+		lh = sis->first_swap_extent.list.prev;	/* Highest extent */
 		se = list_entry(lh, struct swap_extent, list);
 		BUG_ON(se->start_page + se->nr_pages != start_page);
 		if (se->start_block + se->nr_pages == start_block) {
@@ -1246,7 +1387,7 @@ add_swap_extent(struct swap_info_struct *sis, unsigned long start_page,
 	new_se->nr_pages = nr_pages;
 	new_se->start_block = start_block;
 
-	list_add_tail(&new_se->list, &sis->extent_list);
+	list_add_tail(&new_se->list, &sis->first_swap_extent.list);
 	return 1;
 }
 
@@ -1298,7 +1439,7 @@ static int setup_swap_extents(struct swap_info_struct *sis, sector_t *span)
 	if (S_ISBLK(inode->i_mode)) {
 		ret = add_swap_extent(sis, 0, sis->max, 0);
 		*span = sis->pages;
-		goto done;
+		goto out;
 	}
 
 	blkbits = inode->i_blkbits;
@@ -1369,25 +1510,22 @@ reprobe:
 	sis->max = page_no;
 	sis->pages = page_no - 1;
 	sis->highest_bit = page_no - 1;
-done:
-	sis->curr_swap_extent = list_entry(sis->extent_list.prev,
-					struct swap_extent, list);
-	goto out;
+out:
+	return ret;
 bad_bmap:
 	printk(KERN_ERR "swapon: swapfile has holes\n");
 	ret = -EINVAL;
-out:
-	return ret;
+	goto out;
 }
 
 SYSCALL_DEFINE1(swapoff, const char __user *, specialfile)
 {
-	struct swap_info_struct * p = NULL;
+	struct swap_info_struct *p = NULL;
 	unsigned short *swap_map;
 	struct file *swap_file, *victim;
 	struct address_space *mapping;
 	struct inode *inode;
-	char * pathname;
+	char *pathname;
 	int i, type, prev;
 	int err;
 
@@ -1408,8 +1546,8 @@ SYSCALL_DEFINE1(swapoff, const char __user *, specialfile)
 	mapping = victim->f_mapping;
 	prev = -1;
 	spin_lock(&swap_lock);
-	for (type = swap_list.head; type >= 0; type = swap_info[type].next) {
-		p = swap_info + type;
+	for (type = swap_list.head; type >= 0; type = swap_info[type]->next) {
+		p = swap_info[type];
 		if (p->flags & SWP_WRITEOK) {
 			if (p->swap_file->f_mapping == mapping)
 				break;
@@ -1428,18 +1566,17 @@ SYSCALL_DEFINE1(swapoff, const char __user *, specialfile)
 		spin_unlock(&swap_lock);
 		goto out_dput;
 	}
-	if (prev < 0) {
+	if (prev < 0)
 		swap_list.head = p->next;
-	} else {
-		swap_info[prev].next = p->next;
-	}
+	else
+		swap_info[prev]->next = p->next;
 	if (type == swap_list.next) {
 		/* just pick something that's safe... */
 		swap_list.next = swap_list.head;
 	}
 	if (p->prio < 0) {
-		for (i = p->next; i >= 0; i = swap_info[i].next)
-			swap_info[i].prio = p->prio--;
+		for (i = p->next; i >= 0; i = swap_info[i]->next)
+			swap_info[i]->prio = p->prio--;
 		least_priority++;
 	}
 	nr_swap_pages -= p->pages;
@@ -1447,9 +1584,9 @@ SYSCALL_DEFINE1(swapoff, const char __user *, specialfile)
 	p->flags &= ~SWP_WRITEOK;
 	spin_unlock(&swap_lock);
 
-	current->flags |= PF_SWAPOFF;
+	current->flags |= PF_OOM_ORIGIN;
 	err = try_to_unuse(type);
-	current->flags &= ~PF_SWAPOFF;
+	current->flags &= ~PF_OOM_ORIGIN;
 
 	if (err) {
 		/* re-insert swap space back into swap_list */
@@ -1457,16 +1594,16 @@ SYSCALL_DEFINE1(swapoff, const char __user *, specialfile)
 		if (p->prio < 0)
 			p->prio = --least_priority;
 		prev = -1;
-		for (i = swap_list.head; i >= 0; i = swap_info[i].next) {
-			if (p->prio >= swap_info[i].prio)
+		for (i = swap_list.head; i >= 0; i = swap_info[i]->next) {
+			if (p->prio >= swap_info[i]->prio)
 				break;
 			prev = i;
 		}
 		p->next = i;
 		if (prev < 0)
-			swap_list.head = swap_list.next = p - swap_info;
+			swap_list.head = swap_list.next = type;
 		else
-			swap_info[prev].next = p - swap_info;
+			swap_info[prev]->next = type;
 		nr_swap_pages += p->pages;
 		total_swap_pages += p->pages;
 		p->flags |= SWP_WRITEOK;
@@ -1526,8 +1663,8 @@ out:
 /* iterator */
 static void *swap_start(struct seq_file *swap, loff_t *pos)
 {
-	struct swap_info_struct *ptr = swap_info;
-	int i;
+	struct swap_info_struct *si;
+	int type;
 	loff_t l = *pos;
 
 	mutex_lock(&swapon_mutex);
@@ -1535,11 +1672,13 @@ static void *swap_start(struct seq_file *swap, loff_t *pos)
 	if (!l)
 		return SEQ_START_TOKEN;
 
-	for (i = 0; i < nr_swapfiles; i++, ptr++) {
-		if (!(ptr->flags & SWP_USED) || !ptr->swap_map)
+	for (type = 0; type < nr_swapfiles; type++) {
+		smp_rmb();	/* read nr_swapfiles before swap_info[type] */
+		si = swap_info[type];
+		if (!(si->flags & SWP_USED) || !si->swap_map)
 			continue;
 		if (!--l)
-			return ptr;
+			return si;
 	}
 
 	return NULL;
@@ -1547,21 +1686,21 @@ static void *swap_start(struct seq_file *swap, loff_t *pos)
 
 static void *swap_next(struct seq_file *swap, void *v, loff_t *pos)
 {
-	struct swap_info_struct *ptr;
-	struct swap_info_struct *endptr = swap_info + nr_swapfiles;
+	struct swap_info_struct *si = v;
+	int type;
 
 	if (v == SEQ_START_TOKEN)
-		ptr = swap_info;
-	else {
-		ptr = v;
-		ptr++;
-	}
+		type = 0;
+	else
+		type = si->type + 1;
 
-	for (; ptr < endptr; ptr++) {
-		if (!(ptr->flags & SWP_USED) || !ptr->swap_map)
+	for (; type < nr_swapfiles; type++) {
+		smp_rmb();	/* read nr_swapfiles before swap_info[type] */
+		si = swap_info[type];
+		if (!(si->flags & SWP_USED) || !si->swap_map)
 			continue;
 		++*pos;
-		return ptr;
+		return si;
 	}
 
 	return NULL;
@@ -1574,24 +1713,24 @@ static void swap_stop(struct seq_file *swap, void *v)
 
 static int swap_show(struct seq_file *swap, void *v)
 {
-	struct swap_info_struct *ptr = v;
+	struct swap_info_struct *si = v;
 	struct file *file;
 	int len;
 
-	if (ptr == SEQ_START_TOKEN) {
+	if (si == SEQ_START_TOKEN) {
 		seq_puts(swap,"Filename\t\t\t\tType\t\tSize\tUsed\tPriority\n");
 		return 0;
 	}
 
-	file = ptr->swap_file;
+	file = si->swap_file;
 	len = seq_path(swap, &file->f_path, " \t\n\\");
 	seq_printf(swap, "%*s%s\t%u\t%u\t%d\n",
 			len < 40 ? 40 - len : 1, " ",
 			S_ISBLK(file->f_path.dentry->d_inode->i_mode) ?
 				"partition" : "file\t",
-			ptr->pages << (PAGE_SHIFT - 10),
-			ptr->inuse_pages << (PAGE_SHIFT - 10),
-			ptr->prio);
+			si->pages << (PAGE_SHIFT - 10),
+			si->inuse_pages << (PAGE_SHIFT - 10),
+			si->prio);
 	return 0;
 }
 
@@ -1638,7 +1777,7 @@ late_initcall(max_swapfiles_check);
  */
 SYSCALL_DEFINE2(swapon, const char __user *, specialfile, int, swap_flags)
 {
-	struct swap_info_struct * p;
+	struct swap_info_struct *p;
 	char *name = NULL;
 	struct block_device *bdev = NULL;
 	struct file *swap_file = NULL;
@@ -1659,23 +1798,45 @@ SYSCALL_DEFINE2(swapon, const char __user *, specialfile, int, swap_flags)
 
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
+
+	p = kzalloc(sizeof(*p), GFP_KERNEL);
+	if (!p)
+		return -ENOMEM;
+
 	spin_lock(&swap_lock);
-	p = swap_info;
-	for (type = 0 ; type < nr_swapfiles ; type++,p++)
-		if (!(p->flags & SWP_USED))
+	for (type = 0; type < nr_swapfiles; type++) {
+		if (!(swap_info[type]->flags & SWP_USED))
 			break;
+	}
 	error = -EPERM;
 	if (type >= MAX_SWAPFILES) {
 		spin_unlock(&swap_lock);
+		kfree(p);
 		goto out;
 	}
-	if (type >= nr_swapfiles)
-		nr_swapfiles = type+1;
-	memset(p, 0, sizeof(*p));
-	INIT_LIST_HEAD(&p->extent_list);
+	if (type >= nr_swapfiles) {
+		p->type = type;
+		swap_info[type] = p;
+		/*
+		 * Write swap_info[type] before nr_swapfiles, in case a
+		 * racing procfs swap_start() or swap_next() is reading them.
+		 * (We never shrink nr_swapfiles, we never free this entry.)
+		 */
+		smp_wmb();
+		nr_swapfiles++;
+	} else {
+		kfree(p);
+		p = swap_info[type];
+		/*
+		 * Do not memset this entry: a racing procfs swap_next()
+		 * would be relying on p->type to remain valid.
+		 */
+	}
+	INIT_LIST_HEAD(&p->first_swap_extent.list);
 	p->flags = SWP_USED;
 	p->next = -1;
 	spin_unlock(&swap_lock);
+
 	name = getname(specialfile);
 	error = PTR_ERR(name);
 	if (IS_ERR(name)) {
@@ -1695,7 +1856,7 @@ SYSCALL_DEFINE2(swapon, const char __user *, specialfile, int, swap_flags)
 
 	error = -EBUSY;
 	for (i = 0; i < nr_swapfiles; i++) {
-		struct swap_info_struct *q = &swap_info[i];
+		struct swap_info_struct *q = swap_info[i];
 
 		if (i == type || !q->swap_file)
 			continue;
@@ -1770,6 +1931,7 @@ SYSCALL_DEFINE2(swapon, const char __user *, specialfile, int, swap_flags)
 
 	p->lowest_bit  = 1;
 	p->cluster_next = 1;
+	p->cluster_nr = 0;
 
 	/*
 	 * Find out how many pages are allowed for a single swap
@@ -1846,12 +2008,14 @@ SYSCALL_DEFINE2(swapon, const char __user *, specialfile, int, swap_flags)
 		goto bad_swap;
 	}
 
-	if (blk_queue_nonrot(bdev_get_queue(p->bdev))) {
-		p->flags |= SWP_SOLIDSTATE;
-		p->cluster_next = 1 + (random32() % p->highest_bit);
+	if (p->bdev) {
+		if (blk_queue_nonrot(bdev_get_queue(p->bdev))) {
+			p->flags |= SWP_SOLIDSTATE;
+			p->cluster_next = 1 + (random32() % p->highest_bit);
+		}
+		if (discard_swap(p) == 0)
+			p->flags |= SWP_DISCARDABLE;
 	}
-	if (discard_swap(p) == 0)
-		p->flags |= SWP_DISCARDABLE;
 
 	mutex_lock(&swapon_mutex);
 	spin_lock(&swap_lock);
@@ -1874,18 +2038,16 @@ SYSCALL_DEFINE2(swapon, const char __user *, specialfile, int, swap_flags)
 
 	/* insert swap space into swap_list: */
 	prev = -1;
-	for (i = swap_list.head; i >= 0; i = swap_info[i].next) {
-		if (p->prio >= swap_info[i].prio) {
+	for (i = swap_list.head; i >= 0; i = swap_info[i]->next) {
+		if (p->prio >= swap_info[i]->prio)
 			break;
-		}
 		prev = i;
 	}
 	p->next = i;
-	if (prev < 0) {
-		swap_list.head = swap_list.next = p - swap_info;
-	} else {
-		swap_info[prev].next = p - swap_info;
-	}
+	if (prev < 0)
+		swap_list.head = swap_list.next = type;
+	else
+		swap_info[prev]->next = type;
 	spin_unlock(&swap_lock);
 	mutex_unlock(&swapon_mutex);
 	error = 0;
@@ -1922,15 +2084,15 @@ out:
 
 void si_swapinfo(struct sysinfo *val)
 {
-	unsigned int i;
+	unsigned int type;
 	unsigned long nr_to_be_unused = 0;
 
 	spin_lock(&swap_lock);
-	for (i = 0; i < nr_swapfiles; i++) {
-		if (!(swap_info[i].flags & SWP_USED) ||
-		     (swap_info[i].flags & SWP_WRITEOK))
-			continue;
-		nr_to_be_unused += swap_info[i].inuse_pages;
+	for (type = 0; type < nr_swapfiles; type++) {
+		struct swap_info_struct *si = swap_info[type];
+
+		if ((si->flags & SWP_USED) && !(si->flags & SWP_WRITEOK))
+			nr_to_be_unused += si->inuse_pages;
 	}
 	val->freeswap = nr_swap_pages + nr_to_be_unused;
 	val->totalswap = total_swap_pages + nr_to_be_unused;
@@ -1942,34 +2104,65 @@ void si_swapinfo(struct sysinfo *val)
  *
  * Note: if swap_map[] reaches SWAP_MAP_MAX the entries are treated as
  * "permanent", but will be reclaimed by the next swapoff.
+ * Returns error code in following case.
+ * - success -> 0
+ * - swp_entry is invalid -> EINVAL
+ * - swp_entry is migration entry -> EINVAL
+ * - swap-cache reference is requested but there is already one. -> EEXIST
+ * - swap-cache reference is requested but the entry is not used. -> ENOENT
  */
-int swap_duplicate(swp_entry_t entry)
+static int __swap_duplicate(swp_entry_t entry, bool cache)
 {
-	struct swap_info_struct * p;
+	struct swap_info_struct *p;
 	unsigned long offset, type;
-	int result = 0;
+	int result = -EINVAL;
+	int count;
+	bool has_cache;
 
-	if (is_migration_entry(entry))
-		return 1;
+	if (non_swap_entry(entry))
+		return -EINVAL;
 
 	type = swp_type(entry);
 	if (type >= nr_swapfiles)
 		goto bad_file;
-	p = type + swap_info;
+	p = swap_info[type];
 	offset = swp_offset(entry);
 
 	spin_lock(&swap_lock);
-	if (offset < p->max && p->swap_map[offset]) {
-		if (p->swap_map[offset] < SWAP_MAP_MAX - 1) {
-			p->swap_map[offset]++;
-			result = 1;
-		} else if (p->swap_map[offset] <= SWAP_MAP_MAX) {
+
+	if (unlikely(offset >= p->max))
+		goto unlock_out;
+
+	count = swap_count(p->swap_map[offset]);
+	has_cache = swap_has_cache(p->swap_map[offset]);
+
+	if (cache == SWAP_CACHE) { /* called for swapcache/swapin-readahead */
+
+		/* set SWAP_HAS_CACHE if there is no cache and entry is used */
+		if (!has_cache && count) {
+			p->swap_map[offset] = encode_swapmap(count, true);
+			result = 0;
+		} else if (has_cache) /* someone added cache */
+			result = -EEXIST;
+		else if (!count) /* no users */
+			result = -ENOENT;
+
+	} else if (count || has_cache) {
+		if (count < SWAP_MAP_MAX - 1) {
+			p->swap_map[offset] = encode_swapmap(count + 1,
+							     has_cache);
+			result = 0;
+		} else if (count <= SWAP_MAP_MAX) {
 			if (swap_overflow++ < 5)
-				printk(KERN_WARNING "swap_dup: swap entry overflow\n");
-			p->swap_map[offset] = SWAP_MAP_MAX;
-			result = 1;
+				printk(KERN_WARNING
+				       "swap_dup: swap entry overflow\n");
+			p->swap_map[offset] = encode_swapmap(SWAP_MAP_MAX,
+							      has_cache);
+			result = 0;
 		}
-	}
+	} else
+		result = -ENOENT; /* unused swap entry */
+unlock_out:
 	spin_unlock(&swap_lock);
 out:
 	return result;
@@ -1978,11 +2171,25 @@ bad_file:
 	printk(KERN_ERR "swap_dup: %s%08lx\n", Bad_file, entry.val);
 	goto out;
 }
+/*
+ * increase reference count of swap entry by 1.
+ */
+void swap_duplicate(swp_entry_t entry)
+{
+	__swap_duplicate(entry, SWAP_MAP);
+}
 
-struct swap_info_struct *
-get_swap_info_struct(unsigned type)
+/*
+ * @entry: swap entry for which we allocate swap cache.
+ *
+ * Called when allocating swap cache for existing swap entry,
+ * This can return error codes. Returns 0 at success.
+ * -EBUSY means there is a swap cache.
+ * Note: return code is different from swap_duplicate().
+ */
+int swapcache_prepare(swp_entry_t entry)
 {
-	return &swap_info[type];
+	return __swap_duplicate(entry, SWAP_CACHE);
 }
 
 /*
@@ -2000,7 +2207,7 @@ int valid_swaphandles(swp_entry_t entry, unsigned long *offset)
 	if (!our_page_cluster)	/* no readahead */
 		return 0;
 
-	si = &swap_info[swp_type(entry)];
+	si = swap_info[swp_type(entry)];
 	target = swp_offset(entry);
 	base = (target >> our_page_cluster) << our_page_cluster;
 	end = base + (1 << our_page_cluster);
@@ -2016,7 +2223,7 @@ int valid_swaphandles(swp_entry_t entry, unsigned long *offset)
 		/* Don't read in free or bad pages */
 		if (!si->swap_map[toff])
 			break;
-		if (si->swap_map[toff] == SWAP_MAP_BAD)
+		if (swap_count(si->swap_map[toff]) == SWAP_MAP_BAD)
 			break;
 	}
 	/* Count contiguous allocated slots below our target */
@@ -2024,7 +2231,7 @@ int valid_swaphandles(swp_entry_t entry, unsigned long *offset)
 		/* Don't read in free or bad pages */
 		if (!si->swap_map[toff])
 			break;
-		if (si->swap_map[toff] == SWAP_MAP_BAD)
+		if (swap_count(si->swap_map[toff]) == SWAP_MAP_BAD)
 			break;
 	}
 	spin_unlock(&swap_lock);