[Ext4 punch hole 4/4 v4] Ext4 Punch Hole Support: Enable Punch Hole

[Allison Henderson <achender@linux.vnet.ibm.com>]

This patch adds the  new "ext4_punch_hole" "ext4_ext_punch_hole" routines.

fallocate has been modified to call ext4_punch_hole when the punch hole
flag is passed.  At the moment, we only support punching holes in
extents, so this routine is pretty much a wrapper for the ext4_ext_punch_hole
routine.

The ext4_ext_punch_hole routine completes all outstanding writes
with the associated pages, and then releases them.  The unblock
aligned data is zeroed, and all blocks in between are identified
as mapped or already punched out. Mapped blocks are then punched
out using the "ext4_ext_remove_space" routine

Signed-off-by: Allison Henderson <achender@us.ibm.com>
---
:100644 100644 c722e31... fe92010... M	fs/ext4/ext4.h
:100644 100644 01a83a1... 0020044... M	fs/ext4/extents.c
:100644 100644 78c5bc4... a37a2f4... M	fs/ext4/inode.c
 fs/ext4/ext4.h    |    3 +
 fs/ext4/extents.c |  170 +++++++++++++++++++++++++++++++++++++++++++++++++++-
 fs/ext4/inode.c   |   27 +++++++++
 3 files changed, 196 insertions(+), 4 deletions(-)

diff --git a/fs/ext4/ext4.h b/fs/ext4/ext4.h
index c722e31..fe92010 100644
--- a/fs/ext4/ext4.h
+++ b/fs/ext4/ext4.h
@@ -1729,6 +1729,7 @@ extern int ext4_change_inode_journal_flag(struct inode *, int);
 extern int ext4_get_inode_loc(struct inode *, struct ext4_iloc *);
 extern int ext4_can_truncate(struct inode *inode);
 extern void ext4_truncate(struct inode *);
+extern int ext4_punch_hole(struct file *file, loff_t offset, loff_t length);
 extern int ext4_truncate_restart_trans(handle_t *, struct inode *, int nblocks);
 extern void ext4_set_inode_flags(struct inode *);
 extern void ext4_get_inode_flags(struct ext4_inode_info *);
@@ -2066,6 +2067,8 @@ extern int ext4_ext_index_trans_blocks(struct inode *inode, int nrblocks,
 extern int ext4_ext_map_blocks(handle_t *handle, struct inode *inode,
 			       struct ext4_map_blocks *map, int flags);
 extern void ext4_ext_truncate(struct inode *);
+extern int ext4_ext_punch_hole(struct file *file, loff_t offset,
+				loff_t length);
 extern void ext4_ext_init(struct super_block *);
 extern void ext4_ext_release(struct super_block *);
 extern long ext4_fallocate(struct file *file, int mode, loff_t offset,
diff --git a/fs/ext4/extents.c b/fs/ext4/extents.c
index 01a83a1..0020044 100644
--- a/fs/ext4/extents.c
+++ b/fs/ext4/extents.c
@@ -3963,10 +3963,6 @@ long ext4_fallocate(struct file *file, int mode, loff_t offset, loff_t len)
 	struct ext4_map_blocks map;
 	unsigned int credits, blkbits = inode->i_blkbits;
 
-	/* We only support the FALLOC_FL_KEEP_SIZE mode */
-	if (mode & ~FALLOC_FL_KEEP_SIZE)
-		return -EOPNOTSUPP;
-
 	/*
 	 * currently supporting (pre)allocate mode for extent-based
 	 * files _only_
@@ -3974,6 +3970,13 @@ long ext4_fallocate(struct file *file, int mode, loff_t offset, loff_t len)
 	if (!(ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)))
 		return -EOPNOTSUPP;
 
+	/* Return error if mode is not supported */
+	if (mode & ~(FALLOC_FL_KEEP_SIZE | FALLOC_FL_PUNCH_HOLE))
+		return -EOPNOTSUPP;
+
+	if (mode & FALLOC_FL_PUNCH_HOLE)
+		return ext4_punch_hole(file, offset, len);
+
 	map.m_lblk = offset >> blkbits;
 	/*
 	 * We can't just convert len to max_blocks because
@@ -4351,3 +4354,162 @@ int ext4_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 	return error;
 }
 
+/*
+ * ext4_ext_punch_hole
+ *
+ * Punches a hole of "length" bytes in a file starting
+ * at byte "offset"
+ *
+ * @inode:  The inode of the file to punch a hole in
+ * @offset: The starting byte offset of the hole
+ * @length: The length of the hole
+ *
+ * Returns the number of blocks removed or negative on err
+ */
+int ext4_ext_punch_hole(struct file *file, loff_t offset, loff_t length)
+{
+	struct inode *inode = file->f_path.dentry->d_inode;
+	struct super_block *sb = inode->i_sb;
+	struct ext4_extent newex;
+	ext4_lblk_t first_block, last_block, num_blocks, iblock, max_blocks;
+	struct address_space *mapping = inode->i_mapping;
+	struct ext4_map_blocks map;
+	handle_t *handle;
+	loff_t first_block_offset, last_block_offset, block_len;
+	loff_t first_page, last_page, first_page_offset, last_page_offset;
+	int err, credits, ret = 0;
+	int blocks_removed = 0;
+
+	first_block = (offset + sb->s_blocksize - 1)
+			>> EXT4_BLOCK_SIZE_BITS(sb);
+	last_block = (offset + length) >> EXT4_BLOCK_SIZE_BITS(sb);
+
+	first_block_offset = first_block
+			<< EXT4_BLOCK_SIZE_BITS(sb);
+	last_block_offset = last_block << EXT4_BLOCK_SIZE_BITS(sb);
+
+	first_page = (offset + PAGE_CACHE_SIZE - 1)
+			>> PAGE_CACHE_SHIFT;
+	last_page = (offset + length) >> PAGE_CACHE_SHIFT;
+
+	first_page_offset = first_page
+			<< PAGE_CACHE_SHIFT;
+	last_page_offset = last_page << PAGE_CACHE_SHIFT;
+
+	/*
+	 * Write out all dirty pages to avoid race conditions
+	 * Then release them.
+	 */
+	if (last_page_offset > first_page_offset) {
+		if (mapping->nrpages && mapping_tagged(mapping,
+			PAGECACHE_TAG_DIRTY)){
+			err = filemap_write_and_wait_range(mapping,
+				first_page_offset, last_page_offset-1);
+			if (err)
+				return err;
+		}
+		truncate_inode_pages_range(mapping, first_page_offset,
+			last_page_offset-1);
+	}
+
+	/* finish any pending end_io work */
+	ext4_flush_completed_IO(inode);
+
+	credits = ext4_writepage_trans_blocks(inode);
+	handle = ext4_journal_start(inode, credits);
+	if (IS_ERR(handle))
+		return PTR_ERR(handle);
+
+	if (ext4_orphan_add(handle, inode))
+		goto out;
+
+	/*
+	 * Now we need to zero out the un block aligned data.
+	 * If the file is smaller than a block, just
+	 * zero out the middle and return
+	 */
+	if (first_block > last_block)
+		ext4_block_zero_page_range(handle, mapping, offset, length);
+	else {
+		/* zero out the head of the hole before the first block */
+		block_len  = first_block_offset - offset;
+		if (block_len > 0)
+			ext4_block_zero_page_range(handle, mapping,
+				offset, block_len);
+
+		/* zero out the tail of the hole after the last block */
+		block_len = offset + length - last_block_offset;
+		if (block_len > 0) {
+			ext4_block_zero_page_range(handle, mapping,
+				last_block_offset, block_len);
+		}
+	}
+
+	/* If there are no blocks to remove, return now */
+	if (first_block >= last_block)
+		goto out;
+
+	down_write(&EXT4_I(inode)->i_data_sem);
+	ext4_ext_invalidate_cache(inode);
+	ext4_discard_preallocations(inode);
+
+	/*
+	 * Loop over all the blocks and identify blocks
+	 * that need to be punched out
+	 */
+	iblock = first_block;
+	while (iblock < last_block) {
+		max_blocks = last_block - iblock;
+		num_blocks = 1;
+		memset(&map, 0, sizeof(map));
+		map.m_lblk = iblock;
+		map.m_len = max_blocks;
+		ret = ext4_ext_map_blocks(handle, inode, &map, 0);
+
+		if (ret > 0) {
+			num_blocks = ret > max_blocks ? max_blocks : ret;
+
+			/*
+			 * If the blocks are mapped or preallocated,
+			 * release them
+			 */
+			if (map.m_flags & EXT4_MAP_MAPPED)
+				ext4_ext_convert_blocks_uninit(handle,
+					inode, iblock, num_blocks);
+
+			ext4_ext_remove_space(inode, iblock,
+				iblock + num_blocks);
+			blocks_removed += num_blocks;
+		} else if (ret == 0) {
+			/*
+			 * If map blocks could not find the block,
+			 * then it is in a hole.  If the hole was
+			 * not already cached, then map blocks should
+			 * put it in the cache
+			 */
+			memset(&newex, 0, sizeof(newex));
+			if ((ext4_ext_in_cache(inode, iblock, &newex)) &&
+				(!newex.ee_start_lo && !newex.ee_start_hi)) {
+				/* The hole is cached */
+				num_blocks = (newex.ee_block + newex.ee_len) -
+						iblock;
+			}
+		} else {
+			/* Block lookup error */
+			ext_debug("Block lookup failed");
+		}
+
+		iblock += num_blocks;
+	}
+
+	if (IS_SYNC(inode))
+		ext4_handle_sync(handle);
+
+	up_write(&EXT4_I(inode)->i_data_sem);
+out:
+	ext4_orphan_del(handle, inode);
+	inode->i_mtime = inode->i_ctime = ext4_current_time(inode);
+	ext4_mark_inode_dirty(handle, inode);
+	ext4_journal_stop(handle);
+	return blocks_removed;
+}
diff --git a/fs/ext4/inode.c b/fs/ext4/inode.c
index 78c5bc4..a37a2f4 100644
--- a/fs/ext4/inode.c
+++ b/fs/ext4/inode.c
@@ -4406,6 +4406,33 @@ int ext4_can_truncate(struct inode *inode)
 }
 
 /*
+ * ext4_punch_hole: punches a hole in a file by releaseing the blocks
+ * associated with the given offset and length
+ *
+ * @inode:  File inode
+ * @offset: The offset where the hole will begin
+ * @len:    The length of the hole
+ *
+ * Returns: 0 on sucess or negative on failure
+ */
+
+int ext4_punch_hole(struct file *file, loff_t offset, loff_t length)
+{
+	struct inode *inode = file->f_path.dentry->d_inode;
+	if (!S_ISREG(inode->i_mode))
+		return -ENOTSUPP;
+
+	if (!ext4_test_inode_flag(inode, EXT4_INODE_EXTENTS)) {
+		/* TODO: Add support for non extent hole punching */
+		return -ENOTSUPP;
+	}
+
+	ext4_ext_punch_hole(file, offset, length);
+
+	return 0;
+}
+
+/*
  * ext4_truncate()
  *
  * We block out ext4_get_block() block instantiations across the entire
-- 
1.7.1