btrfs: implement swap file support

index e181b6b..e542ce4 100644
@@ -120,6 +120,7 @@ struct dio {
 	spinlock_t bio_lock;		/* protects BIO fields below */
 	int page_errors;		/* errno from get_user_pages() */
 	int is_async;			/* is IO async ? */
+	int should_dirty;		/* should we mark read pages dirty? */
 	bool defer_completion;		/* defer AIO completion to workqueue? */
 	int io_error;			/* IO error in completion path */
 	unsigned long refcount;		/* direct_io_worker() and bios */

@@ -392,7 +393,7 @@ static inline void dio_bio_submit(struct dio *dio, struct dio_submit *sdio)
 	dio->refcount++;
 	spin_unlock_irqrestore(&dio->bio_lock, flags);
 
-	if (dio->is_async && dio->rw == READ)
+	if (dio->is_async && dio->rw == READ && dio->should_dirty)
 		bio_set_pages_dirty(bio);
 
 	if (sdio->submit_io)

@@ -463,13 +464,13 @@ static int dio_bio_complete(struct dio *dio, struct bio *bio)
 	if (!uptodate)
 		dio->io_error = -EIO;
 
-	if (dio->is_async && dio->rw == READ) {
+	if (dio->is_async && dio->rw == READ && dio->should_dirty) {
 		bio_check_pages_dirty(bio);	/* transfers ownership */
 	} else {
 		bio_for_each_segment_all(bvec, bio, i) {
 			struct page *page = bvec->bv_page;
 
-			if (dio->rw == READ && !PageCompound(page))
+			if (dio->rw == READ && !PageCompound(page) && dio->should_dirty)
 				set_page_dirty_lock(page);
 			page_cache_release(page);
 		}

@@ -1177,6 +1178,7 @@ do_blockdev_direct_IO(int rw, struct kiocb *iocb, struct inode *inode,
 
 	dio->inode = inode;
 	dio->rw = rw;
+	dio->should_dirty = !(iter->type & ITER_BVEC);
 
 	/*
 	 * For AIO O_(D)SYNC writes we need to defer completions to a workqueue

index 10bf072..a67fa2c 100644
@@ -88,6 +88,7 @@ struct nfs_direct_req {
 	struct pnfs_ds_commit_info ds_cinfo;	/* Storage for cinfo */
 	struct work_struct	work;
 	int			flags;
+	int			should_dirty;	/* should we mark read pages dirty? */
 #define NFS_ODIRECT_DO_COMMIT		(1)	/* an unstable reply was received */
 #define NFS_ODIRECT_RESCHED_WRITES	(2)	/* write verification failed */
 	struct nfs_writeverf	verf;		/* unstable write verifier */

@@ -370,7 +371,8 @@ static void nfs_direct_read_completion(struct nfs_pgio_header *hdr)
 		struct nfs_page *req = nfs_list_entry(hdr->pages.next);
 		struct page *page = req->wb_page;
 
-		if (!PageCompound(page) && bytes < hdr->good_bytes)
+		if (!PageCompound(page) && bytes < hdr->good_bytes &&
+		    dreq->should_dirty)
 			set_page_dirty(page);
 		bytes += req->wb_bytes;
 		nfs_list_remove_request(req);

@@ -542,6 +544,7 @@ ssize_t nfs_file_direct_read(struct kiocb *iocb, struct iov_iter *iter,
 	dreq->inode = inode;
 	dreq->bytes_left = count;
 	dreq->ctx = get_nfs_open_context(nfs_file_open_context(iocb->ki_filp));
+	dreq->should_dirty = !(iter->type & ITER_BVEC);
 	l_ctx = nfs_get_lock_context(dreq->ctx);
 	if (IS_ERR(l_ctx)) {
 		result = PTR_ERR(l_ctx);

index 955db8b..10715e0 100644
@@ -283,8 +283,7 @@ int __swap_writepage(struct page *page, struct writeback_control *wbc,
 
 		set_page_writeback(page);
 		unlock_page(page);
-		ret = mapping->a_ops->direct_IO(ITER_BVEC | WRITE,
-						&kiocb, &from,
+		ret = mapping->a_ops->direct_IO(WRITE, &kiocb, &from,
 						kiocb.ki_pos);
 		if (ret == PAGE_SIZE) {
 			count_vm_event(PSWPOUT);

@@ -348,12 +347,37 @@ int swap_readpage(struct page *page)
 	}
 
 	if (sis->flags & SWP_FILE) {
+		struct kiocb kiocb;
 		struct file *swap_file = sis->swap_file;
 		struct address_space *mapping = swap_file->f_mapping;
+		struct bio_vec bv = {
+			.bv_page = page,
+			.bv_len = PAGE_SIZE,
+			.bv_offset = 0,
+		};
+		struct iov_iter to = {
+			.type = ITER_BVEC | READ,
+			.count = PAGE_SIZE,
+			.iov_offset = 0,
+			.nr_segs = 1,
+		};
+		to.bvec = &bv;	/* older gcc versions are broken */
+
+		init_sync_kiocb(&kiocb, swap_file);
+		kiocb.ki_pos = page_file_offset(page);
+		kiocb.ki_nbytes = PAGE_SIZE;
 
-		ret = mapping->a_ops->readpage(swap_file, page);
-		if (!ret)
+		ret = mapping->a_ops->direct_IO(READ, &kiocb, &to,
+						kiocb.ki_pos);
+		if (ret == PAGE_SIZE) {
+			SetPageUptodate(page);
 			count_vm_event(PSWPIN);
+			ret = 0;
+		} else {
+			ClearPageUptodate(page);
+			SetPageError(page);
+		}
+		unlock_page(page);
 		return ret;
 	}
 

index fe69edd..38979b9 100644
@@ -1891,6 +1891,9 @@ struct btrfs_root {
 	int send_in_progress;
 	struct btrfs_subvolume_writers *subv_writers;
 	atomic_t will_be_snapshoted;
+
+	/* Number of active swapfiles */
+	atomic_t nr_swapfiles;
 };
 
 struct btrfs_ioctl_defrag_range_args {

index 1bf9f89..60094c4 100644
@@ -1265,6 +1265,7 @@ static void __setup_root(u32 nodesize, u32 sectorsize, u32 stripesize,
 	atomic_set(&root->orphan_inodes, 0);
 	atomic_set(&root->refs, 1);
 	atomic_set(&root->will_be_snapshoted, 0);
+	atomic_set(&root->nr_swapfiles, 0);
 	root->log_transid = 0;
 	root->log_transid_committed = -1;
 	root->last_log_commit = 0;

index 4399f0c..18fa95c 100644
@@ -291,9 +291,11 @@ static int btrfs_ioctl_setflags(struct file *file, void __user *arg)
 		} else {
 			ip->flags |= BTRFS_INODE_NODATACOW;
 		}
-	} else {
+	} else if (!IS_SWAPFILE(inode)) {
 		/*
-		 * Revert back under same assuptions as above
+		 * Revert back under same assumptions as above. swap_activate
+		 * checks that we don't swapon a copy-on-write file, but we also
+		 * make sure that it doesn't become copy-on-write here.
 		 */
 		if (S_ISREG(mode)) {
 			if (inode->i_size == 0)

@@ -316,7 +318,12 @@ static int btrfs_ioctl_setflags(struct file *file, void __user *arg)
 		ret = btrfs_set_prop(inode, "btrfs.compression", NULL, 0, 0);
 		if (ret && ret != -ENODATA)
 			goto out_drop;
-	} else if (flags & FS_COMPR_FL) {
+	} else if (flags & FS_COMPR_FL && !IS_SWAPFILE(inode)) {
+		/*
+		 * Like nodatacow, swap_activate checks that we don't swapon a
+		 * compressed file, so we shouldn't let it become compressed.
+		 */
+
 		const char *comp;
 
 		ip->flags |= BTRFS_INODE_COMPRESS;

@@ -330,7 +337,6 @@ static int btrfs_ioctl_setflags(struct file *file, void __user *arg)
 				     comp, strlen(comp), 0);
 		if (ret)
 			goto out_drop;
-
 	} else {
 		ret = btrfs_set_prop(inode, "btrfs.compression", NULL, 0, 0);
 		if (ret && ret != -ENODATA)

@@ -647,6 +653,12 @@ static int create_snapshot(struct btrfs_root *root, struct inode *dir,
 	if (!test_bit(BTRFS_ROOT_REF_COWS, &root->state))
 		return -EINVAL;
 
+	if (atomic_read(&root->nr_swapfiles)) {
+		btrfs_err(root->fs_info,
+			  "cannot create snapshot with active swapfile");
+		return -ETXTBSY;
+	}
+
 	atomic_inc(&root->will_be_snapshoted);
 	smp_mb__after_atomic();
 	btrfs_wait_nocow_write(root);

@@ -1292,6 +1304,12 @@ int btrfs_defrag_file(struct inode *inode, struct file *file,
 			compress_type = range->compress_type;
 	}
 
+	mutex_lock(&inode->i_mutex);
+	ret = IS_SWAPFILE(inode) ? -ETXTBSY : 0;
+	mutex_unlock(&inode->i_mutex);
+	if (ret)
+		return ret;
+
 	if (extent_thresh == 0)
 		extent_thresh = 256 * 1024;
 

@@ -2927,6 +2945,11 @@ static int btrfs_extent_same(struct inode *src, u64 loff, u64 len,
 
 	btrfs_double_lock(src, loff, dst, dst_loff, len);
 
+	if (IS_SWAPFILE(src) || IS_SWAPFILE(dst)) {
+		ret = -ETXTBSY;
+		goto out_unlock;
+	}
+
 	ret = extent_same_check_offsets(src, loff, len);
 	if (ret)
 		goto out_unlock;

@@ -3644,6 +3667,10 @@ static noinline long btrfs_ioctl_clone(struct file *file, unsigned long srcfd,
 		mutex_lock(&src->i_mutex);
 	}
 
+	ret = -ETXTBSY;
+	if (IS_SWAPFILE(src) || IS_SWAPFILE(inode))
+		goto out_unlock;
+
 	/* determine range to clone */
 	ret = -EINVAL;
 	if (off + len > src->i_size || off + len < off)

index bf3f424..36166d0 100644
@@ -4244,6 +4244,7 @@ int try_release_extent_mapping(struct extent_map_tree *map,
 				break;
 			}
 			if (test_bit(EXTENT_FLAG_PINNED, &em->flags) ||
+			    test_bit(EXTENT_FLAG_SWAPFILE, &em->flags) ||
 			    em->start != start) {
 				write_unlock(&map->lock);
 				free_extent_map(em);

index b2991fd..93b9548 100644
@@ -16,6 +16,7 @@
 #define EXTENT_FLAG_LOGGING 4 /* Logging this extent */
 #define EXTENT_FLAG_FILLING 5 /* Filling in a preallocated extent */
 #define EXTENT_FLAG_FS_MAPPING 6 /* filesystem extent mapping type */
+#define EXTENT_FLAG_SWAPFILE 7 /* this extent maps a swap file */
 
 struct extent_map {
 	struct rb_node rb_node;

index d23362f..7c2dfb2 100644
@@ -6353,6 +6353,7 @@ again:
 		else
 			goto out;
 	}
+	WARN_ON_ONCE(IS_SWAPFILE(inode));
 	em = alloc_extent_map();
 	if (!em) {
 		err = -ENOMEM;

index 1faecea..5c5f9de 100644
@@ -164,7 +164,8 @@ DEFINE_EVENT(btrfs__inode, btrfs_inode_evict,
 		{ (1 << EXTENT_FLAG_PREALLOC), 		"PREALLOC" 	},\
 		{ (1 << EXTENT_FLAG_LOGGING),	 	"LOGGING" 	},\
 		{ (1 << EXTENT_FLAG_FILLING),	 	"FILLING" 	},\
-		{ (1 << EXTENT_FLAG_FS_MAPPING),	"FS_MAPPING"	})
+		{ (1 << EXTENT_FLAG_FS_MAPPING),	"FS_MAPPING"	},\
+		{ (1 << EXTENT_FLAG_SWAPFILE),		"SWAPFILE"	})
 
 TRACE_EVENT_CONDITION(btrfs_get_extent,
 

index 7c2dfb2..76b58d7 100644
@@ -7191,6 +7191,7 @@ static int btrfs_get_blocks_direct(struct inode *inode, sector_t iblock,
 	 * this will cow the extent, reset the len in case we changed
 	 * it above
 	 */
+	WARN_ON_ONCE(IS_SWAPFILE(inode));
 	len = bh_result->b_size;
 	free_extent_map(em);
 	em = btrfs_new_extent_direct(inode, start, len);

@@ -9443,6 +9444,134 @@ out_inode:
 
 }
 
+static void __clear_swapfile_extents(struct inode *inode)
+{
+	u64 isize = inode->i_size;
+	struct extent_map *em;
+	u64 start, len;
+
+	start = 0;
+	while (start < isize) {
+		len = isize - start;
+		em = btrfs_get_extent(inode, NULL, 0, start, len, 0);
+		if (IS_ERR(em))
+			return;
+
+		clear_bit(EXTENT_FLAG_SWAPFILE, &em->flags);
+
+		start = extent_map_end(em);
+		free_extent_map(em);
+	}
+}
+
+static int btrfs_swap_activate(struct swap_info_struct *sis, struct file *file,
+			       sector_t *span)
+{
+	struct inode *inode = file_inode(file);
+	struct btrfs_fs_info *fs_info = BTRFS_I(inode)->root->fs_info;
+	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+	int ret = 0;
+	u64 isize = inode->i_size;
+	struct extent_state *cached_state = NULL;
+	struct extent_map *em;
+	u64 start, len;
+
+	if (BTRFS_I(inode)->flags & BTRFS_INODE_COMPRESS) {
+		/* Can't do direct I/O on a compressed file. */
+		btrfs_err(fs_info, "swapfile is compressed");
+		return -EINVAL;
+	}
+	if (!(BTRFS_I(inode)->flags & BTRFS_INODE_NODATACOW)) {
+		/*
+		 * Going through the copy-on-write path while swapping pages
+		 * in/out and doing a bunch of allocations could stress the
+		 * memory management code that got us there in the first place,
+		 * and that's sure to be a bad time.
+		 */
+		btrfs_err(fs_info, "swapfile is copy-on-write");
+		return -EINVAL;
+	}
+
+	lock_extent_bits(io_tree, 0, isize - 1, 0, &cached_state);
+
+	/*
+	 * All of the extents must be allocated and support direct I/O. Inline
+	 * extents and compressed extents fall back to buffered I/O, so those
+	 * are no good. Additionally, all of the extents must be safe for nocow.
+	 */
+	atomic_inc(&BTRFS_I(inode)->root->nr_swapfiles);
+	start = 0;
+	while (start < isize) {
+		len = isize - start;
+		em = btrfs_get_extent(inode, NULL, 0, start, len, 0);
+		if (IS_ERR(em)) {
+			ret = PTR_ERR(em);
+			goto out;
+		}
+
+		if (test_bit(EXTENT_FLAG_VACANCY, &em->flags) ||
+		    em->block_start == EXTENT_MAP_HOLE) {
+			btrfs_err(fs_info, "swapfile has holes");
+			ret = -EINVAL;
+			goto out;
+		}
+		if (em->block_start == EXTENT_MAP_INLINE) {
+			/*
+			 * It's unlikely we'll ever actually find ourselves
+			 * here, as a file small enough to fit inline won't be
+			 * big enough to store more than the swap header, but in
+			 * case something changes in the future, let's catch it
+			 * here rather than later.
+			 */
+			btrfs_err(fs_info, "swapfile is inline");
+			ret = -EINVAL;
+			goto out;
+		}
+		if (test_bit(EXTENT_FLAG_COMPRESSED, &em->flags)) {
+			btrfs_err(fs_info, "swapfile is compresed");
+			ret = -EINVAL;
+			goto out;
+		}
+		ret = can_nocow_extent(inode, start, &len, NULL, NULL, NULL);
+		if (ret < 0) {
+			goto out;
+		} else if (ret == 1) {
+			ret = 0;
+		} else {
+			btrfs_err(fs_info, "swapfile has extent requiring COW (%llu-%llu)",
+				  start, start + len - 1);
+			ret = -EINVAL;
+			goto out;
+		}
+
+		set_bit(EXTENT_FLAG_SWAPFILE, &em->flags);
+
+		start = extent_map_end(em);
+		free_extent_map(em);
+	}
+
+out:
+	if (ret) {
+		__clear_swapfile_extents(inode);
+		atomic_dec(&BTRFS_I(inode)->root->nr_swapfiles);
+	}
+	unlock_extent_cached(io_tree, 0, isize - 1, &cached_state, GFP_NOFS);
+	return ret;
+}
+
+static void btrfs_swap_deactivate(struct file *file)
+{
+	struct inode *inode = file_inode(file);
+	struct extent_io_tree *io_tree = &BTRFS_I(inode)->io_tree;
+	struct extent_state *cached_state = NULL;
+	u64 isize = inode->i_size;
+
+	lock_extent_bits(io_tree, 0, isize - 1, 0, &cached_state);
+	__clear_swapfile_extents(inode);
+	unlock_extent_cached(io_tree, 0, isize - 1, &cached_state, GFP_NOFS);
+	atomic_dec(&BTRFS_I(inode)->root->nr_swapfiles);
+}
+
 static const struct inode_operations btrfs_dir_inode_operations = {
 	.getattr	= btrfs_getattr,
 	.lookup		= btrfs_lookup,

@@ -9520,6 +9649,8 @@ static const struct address_space_operations btrfs_aops = {
 	.releasepage	= btrfs_releasepage,
 	.set_page_dirty	= btrfs_set_page_dirty,
 	.error_remove_page = generic_error_remove_page,
+	.swap_activate	= btrfs_swap_activate,
+	.swap_deactivate = btrfs_swap_deactivate,
 };
 
 static const struct address_space_operations btrfs_symlink_aops = {