Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/btrfs-unstable

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mason/btrfs-unstable:
  Btrfs: make sure reserve_metadata_bytes doesn't leak out strange errors
  Btrfs: use the commit_root for reading free_space_inode crcs
  Btrfs: reduce extent_state lock contention for metadata
  Btrfs: remove lockdep magic from btrfs_next_leaf
  Btrfs: make a lockdep class for each root
  Btrfs: switch the btrfs tree locks to reader/writer
  Btrfs: fix deadlock when throttling transactions
  Btrfs: stop using highmem for extent_buffers
  Btrfs: fix BUG_ON() caused by ENOSPC when relocating space
  Btrfs: tag pages for writeback in sync
  Btrfs: fix enospc problems with delalloc
  Btrfs: don't flush delalloc arbitrarily
  Btrfs: use find_or_create_page instead of grab_cache_page
  Btrfs: use a worker thread to do caching
  Btrfs: fix how we merge extent states and deal with cached states
  Btrfs: use the normal checksumming infrastructure for free space cache
  Btrfs: serialize flushers in reserve_metadata_bytes
  Btrfs: do transaction space reservation before joining the transaction
  Btrfs: try to only do one btrfs_search_slot in do_setxattr

1 files changed, 83 insertions, 85 deletions

diff --git a/fs/btrfs/extent_io.c b/fs/btrfs/extent_io.c
index 561262d35689..067b1747421b 100644
--- a/fs/btrfs/extent_io.c
+++ b/fs/btrfs/extent_io.c
@@ -281,11 +281,10 @@ static int merge_state(struct extent_io_tree *tree,
 		if (other->start == state->end + 1 &&
 		    other->state == state->state) {
 			merge_cb(tree, state, other);
-			other->start = state->start;
-			state->tree = NULL;
-			rb_erase(&state->rb_node, &tree->state);
-			free_extent_state(state);
-			state = NULL;
+			state->end = other->end;
+			other->tree = NULL;
+			rb_erase(&other->rb_node, &tree->state);
+			free_extent_state(other);
 		}
 	}
 
@@ -351,7 +350,6 @@ static int insert_state(struct extent_io_tree *tree,
 		       "%llu %llu\n", (unsigned long long)found->start,
 		       (unsigned long long)found->end,
 		       (unsigned long long)start, (unsigned long long)end);
-		free_extent_state(state);
 		return -EEXIST;
 	}
 	state->tree = tree;
@@ -500,7 +498,8 @@ again:
 			cached_state = NULL;
 		}
 
-		if (cached && cached->tree && cached->start == start) {
+		if (cached && cached->tree && cached->start <= start &&
+		    cached->end > start) {
 			if (clear)
 				atomic_dec(&cached->refs);
 			state = cached;
@@ -742,7 +741,8 @@ again:
 	spin_lock(&tree->lock);
 	if (cached_state && *cached_state) {
 		state = *cached_state;
-		if (state->start == start && state->tree) {
+		if (state->start <= start && state->end > start &&
+		    state->tree) {
 			node = &state->rb_node;
 			goto hit_next;
 		}
@@ -783,13 +783,13 @@ hit_next:
 		if (err)
 			goto out;
 
-		next_node = rb_next(node);
 		cache_state(state, cached_state);
 		merge_state(tree, state);
 		if (last_end == (u64)-1)
 			goto out;
 
 		start = last_end + 1;
+		next_node = rb_next(&state->rb_node);
 		if (next_node && start < end && prealloc && !need_resched()) {
 			state = rb_entry(next_node, struct extent_state,
 					 rb_node);
@@ -862,7 +862,6 @@ hit_next:
 		 * Avoid to free 'prealloc' if it can be merged with
 		 * the later extent.
 		 */
-		atomic_inc(&prealloc->refs);
 		err = insert_state(tree, prealloc, start, this_end,
 				   &bits);
 		BUG_ON(err == -EEXIST);
@@ -872,7 +871,6 @@ hit_next:
 			goto out;
 		}
 		cache_state(prealloc, cached_state);
-		free_extent_state(prealloc);
 		prealloc = NULL;
 		start = this_end + 1;
 		goto search_again;
@@ -1564,7 +1562,8 @@ int test_range_bit(struct extent_io_tree *tree, u64 start, u64 end,
 	int bitset = 0;
 
 	spin_lock(&tree->lock);
-	if (cached && cached->tree && cached->start == start)
+	if (cached && cached->tree && cached->start <= start &&
+	    cached->end > start)
 		node = &cached->rb_node;
 	else
 		node = tree_search(tree, start);
@@ -2432,6 +2431,7 @@ static int extent_write_cache_pages(struct extent_io_tree *tree,
 	pgoff_t index;
 	pgoff_t end;		/* Inclusive */
 	int scanned = 0;
+	int tag;
 
 	pagevec_init(&pvec, 0);
 	if (wbc->range_cyclic) {
@@ -2442,11 +2442,16 @@ static int extent_write_cache_pages(struct extent_io_tree *tree,
 		end = wbc->range_end >> PAGE_CACHE_SHIFT;
 		scanned = 1;
 	}
+	if (wbc->sync_mode == WB_SYNC_ALL)
+		tag = PAGECACHE_TAG_TOWRITE;
+	else
+		tag = PAGECACHE_TAG_DIRTY;
 retry:
+	if (wbc->sync_mode == WB_SYNC_ALL)
+		tag_pages_for_writeback(mapping, index, end);
 	while (!done && !nr_to_write_done && (index <= end) &&
-	       (nr_pages = pagevec_lookup_tag(&pvec, mapping, &index,
-			      PAGECACHE_TAG_DIRTY, min(end - index,
-				  (pgoff_t)PAGEVEC_SIZE-1) + 1))) {
+	       (nr_pages = pagevec_lookup_tag(&pvec, mapping, &index, tag,
+			min(end - index, (pgoff_t)PAGEVEC_SIZE-1) + 1))) {
 		unsigned i;
 
 		scanned = 1;
@@ -3020,8 +3025,15 @@ static struct extent_buffer *__alloc_extent_buffer(struct extent_io_tree *tree,
 		return NULL;
 	eb->start = start;
 	eb->len = len;
-	spin_lock_init(&eb->lock);
-	init_waitqueue_head(&eb->lock_wq);
+	rwlock_init(&eb->lock);
+	atomic_set(&eb->write_locks, 0);
+	atomic_set(&eb->read_locks, 0);
+	atomic_set(&eb->blocking_readers, 0);
+	atomic_set(&eb->blocking_writers, 0);
+	atomic_set(&eb->spinning_readers, 0);
+	atomic_set(&eb->spinning_writers, 0);
+	init_waitqueue_head(&eb->write_lock_wq);
+	init_waitqueue_head(&eb->read_lock_wq);
 
 #if LEAK_DEBUG
 	spin_lock_irqsave(&leak_lock, flags);
@@ -3117,7 +3129,7 @@ struct extent_buffer *alloc_extent_buffer(struct extent_io_tree *tree,
 		i = 0;
 	}
 	for (; i < num_pages; i++, index++) {
-		p = find_or_create_page(mapping, index, GFP_NOFS | __GFP_HIGHMEM);
+		p = find_or_create_page(mapping, index, GFP_NOFS);
 		if (!p) {
 			WARN_ON(1);
 			goto free_eb;
@@ -3264,6 +3276,22 @@ int set_extent_buffer_dirty(struct extent_io_tree *tree,
 	return was_dirty;
 }
 
+static int __eb_straddles_pages(u64 start, u64 len)
+{
+	if (len < PAGE_CACHE_SIZE)
+		return 1;
+	if (start & (PAGE_CACHE_SIZE - 1))
+		return 1;
+	if ((start + len) & (PAGE_CACHE_SIZE - 1))
+		return 1;
+	return 0;
+}
+
+static int eb_straddles_pages(struct extent_buffer *eb)
+{
+	return __eb_straddles_pages(eb->start, eb->len);
+}
+
 int clear_extent_buffer_uptodate(struct extent_io_tree *tree,
 				struct extent_buffer *eb,
 				struct extent_state **cached_state)
@@ -3275,8 +3303,10 @@ int clear_extent_buffer_uptodate(struct extent_io_tree *tree,
 	num_pages = num_extent_pages(eb->start, eb->len);
 	clear_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags);
 
-	clear_extent_uptodate(tree, eb->start, eb->start + eb->len - 1,
-			      cached_state, GFP_NOFS);
+	if (eb_straddles_pages(eb)) {
+		clear_extent_uptodate(tree, eb->start, eb->start + eb->len - 1,
+				      cached_state, GFP_NOFS);
+	}
 	for (i = 0; i < num_pages; i++) {
 		page = extent_buffer_page(eb, i);
 		if (page)
@@ -3294,8 +3324,10 @@ int set_extent_buffer_uptodate(struct extent_io_tree *tree,
 
 	num_pages = num_extent_pages(eb->start, eb->len);
 
-	set_extent_uptodate(tree, eb->start, eb->start + eb->len - 1,
-			    NULL, GFP_NOFS);
+	if (eb_straddles_pages(eb)) {
+		set_extent_uptodate(tree, eb->start, eb->start + eb->len - 1,
+				    NULL, GFP_NOFS);
+	}
 	for (i = 0; i < num_pages; i++) {
 		page = extent_buffer_page(eb, i);
 		if ((i == 0 && (eb->start & (PAGE_CACHE_SIZE - 1))) ||
@@ -3318,9 +3350,12 @@ int extent_range_uptodate(struct extent_io_tree *tree,
 	int uptodate;
 	unsigned long index;
 
-	ret = test_range_bit(tree, start, end, EXTENT_UPTODATE, 1, NULL);
-	if (ret)
-		return 1;
+	if (__eb_straddles_pages(start, end - start + 1)) {
+		ret = test_range_bit(tree, start, end,
+				     EXTENT_UPTODATE, 1, NULL);
+		if (ret)
+			return 1;
+	}
 	while (start <= end) {
 		index = start >> PAGE_CACHE_SHIFT;
 		page = find_get_page(tree->mapping, index);
@@ -3348,10 +3383,12 @@ int extent_buffer_uptodate(struct extent_io_tree *tree,
 	if (test_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags))
 		return 1;
 
-	ret = test_range_bit(tree, eb->start, eb->start + eb->len - 1,
-			   EXTENT_UPTODATE, 1, cached_state);
-	if (ret)
-		return ret;
+	if (eb_straddles_pages(eb)) {
+		ret = test_range_bit(tree, eb->start, eb->start + eb->len - 1,
+				   EXTENT_UPTODATE, 1, cached_state);
+		if (ret)
+			return ret;
+	}
 
 	num_pages = num_extent_pages(eb->start, eb->len);
 	for (i = 0; i < num_pages; i++) {
@@ -3384,9 +3421,11 @@ int read_extent_buffer_pages(struct extent_io_tree *tree,
 	if (test_bit(EXTENT_BUFFER_UPTODATE, &eb->bflags))
 		return 0;
 
-	if (test_range_bit(tree, eb->start, eb->start + eb->len - 1,
-			   EXTENT_UPTODATE, 1, NULL)) {
-		return 0;
+	if (eb_straddles_pages(eb)) {
+		if (test_range_bit(tree, eb->start, eb->start + eb->len - 1,
+				   EXTENT_UPTODATE, 1, NULL)) {
+			return 0;
+		}
 	}
 
 	if (start) {
@@ -3490,9 +3529,8 @@ void read_extent_buffer(struct extent_buffer *eb, void *dstv,
 		page = extent_buffer_page(eb, i);
 
 		cur = min(len, (PAGE_CACHE_SIZE - offset));
-		kaddr = kmap_atomic(page, KM_USER1);
+		kaddr = page_address(page);
 		memcpy(dst, kaddr + offset, cur);
-		kunmap_atomic(kaddr, KM_USER1);
 
 		dst += cur;
 		len -= cur;
@@ -3502,9 +3540,9 @@ void read_extent_buffer(struct extent_buffer *eb, void *dstv,
 }
 
 int map_private_extent_buffer(struct extent_buffer *eb, unsigned long start,
-			       unsigned long min_len, char **token, char **map,
+			       unsigned long min_len, char **map,
 			       unsigned long *map_start,
-			       unsigned long *map_len, int km)
+			       unsigned long *map_len)
 {
 	size_t offset = start & (PAGE_CACHE_SIZE - 1);
 	char *kaddr;
@@ -3534,42 +3572,12 @@ int map_private_extent_buffer(struct extent_buffer *eb, unsigned long start,
 	}
 
 	p = extent_buffer_page(eb, i);
-	kaddr = kmap_atomic(p, km);
-	*token = kaddr;
+	kaddr = page_address(p);
 	*map = kaddr + offset;
 	*map_len = PAGE_CACHE_SIZE - offset;
 	return 0;
 }
 
-int map_extent_buffer(struct extent_buffer *eb, unsigned long start,
-		      unsigned long min_len,
-		      char **token, char **map,
-		      unsigned long *map_start,
-		      unsigned long *map_len, int km)
-{
-	int err;
-	int save = 0;
-	if (eb->map_token) {
-		unmap_extent_buffer(eb, eb->map_token, km);
-		eb->map_token = NULL;
-		save = 1;
-	}
-	err = map_private_extent_buffer(eb, start, min_len, token, map,
-				       map_start, map_len, km);
-	if (!err && save) {
-		eb->map_token = *token;
-		eb->kaddr = *map;
-		eb->map_start = *map_start;
-		eb->map_len = *map_len;
-	}
-	return err;
-}
-
-void unmap_extent_buffer(struct extent_buffer *eb, char *token, int km)
-{
-	kunmap_atomic(token, km);
-}
-
 int memcmp_extent_buffer(struct extent_buffer *eb, const void *ptrv,
 			  unsigned long start,
 			  unsigned long len)
@@ -3593,9 +3601,8 @@ int memcmp_extent_buffer(struct extent_buffer *eb, const void *ptrv,
 
 		cur = min(len, (PAGE_CACHE_SIZE - offset));
 
-		kaddr = kmap_atomic(page, KM_USER0);
+		kaddr = page_address(page);
 		ret = memcmp(ptr, kaddr + offset, cur);
-		kunmap_atomic(kaddr, KM_USER0);
 		if (ret)
 			break;
 
@@ -3628,9 +3635,8 @@ void write_extent_buffer(struct extent_buffer *eb, const void *srcv,
 		WARN_ON(!PageUptodate(page));
 
 		cur = min(len, PAGE_CACHE_SIZE - offset);
-		kaddr = kmap_atomic(page, KM_USER1);
+		kaddr = page_address(page);
 		memcpy(kaddr + offset, src, cur);
-		kunmap_atomic(kaddr, KM_USER1);
 
 		src += cur;
 		len -= cur;
@@ -3659,9 +3665,8 @@ void memset_extent_buffer(struct extent_buffer *eb, char c,
 		WARN_ON(!PageUptodate(page));
 
 		cur = min(len, PAGE_CACHE_SIZE - offset);
-		kaddr = kmap_atomic(page, KM_USER0);
+		kaddr = page_address(page);
 		memset(kaddr + offset, c, cur);
-		kunmap_atomic(kaddr, KM_USER0);
 
 		len -= cur;
 		offset = 0;
@@ -3692,9 +3697,8 @@ void copy_extent_buffer(struct extent_buffer *dst, struct extent_buffer *src,
 
 		cur = min(len, (unsigned long)(PAGE_CACHE_SIZE - offset));
 
-		kaddr = kmap_atomic(page, KM_USER0);
+		kaddr = page_address(page);
 		read_extent_buffer(src, kaddr + offset, src_offset, cur);
-		kunmap_atomic(kaddr, KM_USER0);
 
 		src_offset += cur;
 		len -= cur;
@@ -3707,20 +3711,17 @@ static void move_pages(struct page *dst_page, struct page *src_page,
 		       unsigned long dst_off, unsigned long src_off,
 		       unsigned long len)
 {
-	char *dst_kaddr = kmap_atomic(dst_page, KM_USER0);
+	char *dst_kaddr = page_address(dst_page);
 	if (dst_page == src_page) {
 		memmove(dst_kaddr + dst_off, dst_kaddr + src_off, len);
 	} else {
-		char *src_kaddr = kmap_atomic(src_page, KM_USER1);
+		char *src_kaddr = page_address(src_page);
 		char *p = dst_kaddr + dst_off + len;
 		char *s = src_kaddr + src_off + len;
 
 		while (len--)
 			*--p = *--s;
-
-		kunmap_atomic(src_kaddr, KM_USER1);
 	}
-	kunmap_atomic(dst_kaddr, KM_USER0);
 }
 
 static inline bool areas_overlap(unsigned long src, unsigned long dst, unsigned long len)
@@ -3733,20 +3734,17 @@ static void copy_pages(struct page *dst_page, struct page *src_page,
 		       unsigned long dst_off, unsigned long src_off,
 		       unsigned long len)
 {
-	char *dst_kaddr = kmap_atomic(dst_page, KM_USER0);
+	char *dst_kaddr = page_address(dst_page);
 	char *src_kaddr;
 
 	if (dst_page != src_page) {
-		src_kaddr = kmap_atomic(src_page, KM_USER1);
+		src_kaddr = page_address(src_page);
 	} else {
 		src_kaddr = dst_kaddr;
 		BUG_ON(areas_overlap(src_off, dst_off, len));
 	}
 
 	memcpy(dst_kaddr + dst_off, src_kaddr + src_off, len);
-	kunmap_atomic(dst_kaddr, KM_USER0);
-	if (dst_page != src_page)
-		kunmap_atomic(src_kaddr, KM_USER1);
 }
 
 void memcpy_extent_buffer(struct extent_buffer *dst, unsigned long dst_offset,