ocfs2: Read from an unwritten extent returns zeros

Return an optional extent flags field from our lookup functions and wire up
callers to treat unwritten regions as holes for the purpose of returning
zeros to the user.

Signed-off-by: Mark Fasheh <mark.fasheh@oracle.com>

10 files changed, 45 insertions, 23 deletions

diff --git a/fs/ocfs2/alloc.c b/fs/ocfs2/alloc.c
index 0eab0d32828..412a2888a3e 100644
--- a/fs/ocfs2/alloc.c
+++ b/fs/ocfs2/alloc.c
@@ -3487,6 +3487,7 @@ static int ocfs2_grab_eof_pages(struct inode *inode, loff_t isize, struct page *
 {
 	int i, numpages = 0, ret = 0;
 	unsigned int csize = OCFS2_SB(inode->i_sb)->s_clustersize;
+	unsigned int ext_flags;
 	struct super_block *sb = inode->i_sb;
 	struct address_space *mapping = inode->i_mapping;
 	unsigned long index;
@@ -3499,7 +3500,7 @@ static int ocfs2_grab_eof_pages(struct inode *inode, loff_t isize, struct page *
 		goto out;
 
 	ret = ocfs2_extent_map_get_blocks(inode, isize >> sb->s_blocksize_bits,
-					  phys, NULL);
+					  phys, NULL, &ext_flags);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
@@ -3509,6 +3510,11 @@ static int ocfs2_grab_eof_pages(struct inode *inode, loff_t isize, struct page *
 	if (*phys == 0)
 		goto out;
 
+	/* Tail is marked as unwritten, we can count on write to zero
+	 * in that case. */
+	if (ext_flags & OCFS2_EXT_UNWRITTEN)
+		goto out;
+
 	next_cluster_bytes = ocfs2_align_bytes_to_clusters(inode->i_sb, isize);
 	index = isize >> PAGE_CACHE_SHIFT;
 	do {
@@ -3579,9 +3585,6 @@ int ocfs2_zero_tail_for_truncate(struct inode *inode, handle_t *handle,
 		goto out;
 	}
 
-	/*
-	 * Truncate on an i_size boundary - nothing more to do.
-	 */
 	if (numpages == 0)
 		goto out;
 
diff --git a/fs/ocfs2/aops.c b/fs/ocfs2/aops.c
index 014f4f52809..eb67c902b00 100644
--- a/fs/ocfs2/aops.c
+++ b/fs/ocfs2/aops.c
@@ -137,6 +137,7 @@ static int ocfs2_get_block(struct inode *inode, sector_t iblock,
 			   struct buffer_head *bh_result, int create)
 {
 	int err = 0;
+	unsigned int ext_flags;
 	u64 p_blkno, past_eof;
 	struct ocfs2_super *osb = OCFS2_SB(inode->i_sb);
 
@@ -153,7 +154,8 @@ static int ocfs2_get_block(struct inode *inode, sector_t iblock,
 		goto bail;
 	}
 
-	err = ocfs2_extent_map_get_blocks(inode, iblock, &p_blkno, NULL);
+	err = ocfs2_extent_map_get_blocks(inode, iblock, &p_blkno, NULL,
+					  &ext_flags);
 	if (err) {
 		mlog(ML_ERROR, "Error %d from get_blocks(0x%p, %llu, 1, "
 		     "%llu, NULL)\n", err, inode, (unsigned long long)iblock,
@@ -171,7 +173,8 @@ static int ocfs2_get_block(struct inode *inode, sector_t iblock,
 			"ino %lu, iblock %llu\n", inode->i_ino,
 			(unsigned long long)iblock);
 
-	if (p_blkno)
+	/* Treat the unwritten extent as a hole for zeroing purposes. */
+	if (p_blkno && !(ext_flags & OCFS2_EXT_UNWRITTEN))
 		map_bh(bh_result, inode->i_sb, p_blkno);
 
 	if (!ocfs2_sparse_alloc(osb)) {
@@ -396,7 +399,7 @@ static sector_t ocfs2_bmap(struct address_space *mapping, sector_t block)
 		down_read(&OCFS2_I(inode)->ip_alloc_sem);
 	}
 
-	err = ocfs2_extent_map_get_blocks(inode, block, &p_blkno, NULL);
+	err = ocfs2_extent_map_get_blocks(inode, block, &p_blkno, NULL, NULL);
 
 	if (!INODE_JOURNAL(inode)) {
 		up_read(&OCFS2_I(inode)->ip_alloc_sem);
@@ -438,6 +441,7 @@ static int ocfs2_direct_IO_get_blocks(struct inode *inode, sector_t iblock,
 	int ret;
 	u64 p_blkno, inode_blocks;
 	int contig_blocks;
+	unsigned int ext_flags;
 	unsigned char blocksize_bits = inode->i_sb->s_blocksize_bits;
 	unsigned long max_blocks = bh_result->b_size >> inode->i_blkbits;
 
@@ -458,7 +462,7 @@ static int ocfs2_direct_IO_get_blocks(struct inode *inode, sector_t iblock,
 	/* This figures out the size of the next contiguous block, and
 	 * our logical offset */
 	ret = ocfs2_extent_map_get_blocks(inode, iblock, &p_blkno,
-					  &contig_blocks);
+					  &contig_blocks, &ext_flags);
 	if (ret) {
 		mlog(ML_ERROR, "get_blocks() failed iblock=%llu\n",
 		     (unsigned long long)iblock);
@@ -478,8 +482,10 @@ static int ocfs2_direct_IO_get_blocks(struct inode *inode, sector_t iblock,
 	/*
 	 * get_more_blocks() expects us to describe a hole by clearing
 	 * the mapped bit on bh_result().
+	 *
+	 * Consider an unwritten extent as a hole.
 	 */
-	if (p_blkno)
+	if (p_blkno && !(ext_flags & OCFS2_EXT_UNWRITTEN))
 		map_bh(bh_result, inode->i_sb, p_blkno);
 	else {
 		/*
@@ -1111,7 +1117,8 @@ static ssize_t ocfs2_write(struct file *file, u32 phys, handle_t *handle,
 		}
 	}
 
-	ret = ocfs2_extent_map_get_blocks(inode, v_blkno, &p_blkno, NULL);
+	ret = ocfs2_extent_map_get_blocks(inode, v_blkno, &p_blkno, NULL,
+					  NULL);
 	if (ret < 0) {
 
 		/*
@@ -1215,7 +1222,7 @@ ssize_t ocfs2_buffered_write_cluster(struct file *file, loff_t pos,
 	 */
 	down_write(&OCFS2_I(inode)->ip_alloc_sem);
 
-	ret = ocfs2_get_clusters(inode, wc.w_cpos, &phys, NULL);
+	ret = ocfs2_get_clusters(inode, wc.w_cpos, &phys, NULL, NULL);
 	if (ret) {
 		mlog_errno(ret);
 		goto out_meta;
diff --git a/fs/ocfs2/dir.c b/fs/ocfs2/dir.c
index c91490670ff..8d22e1e4a88 100644
--- a/fs/ocfs2/dir.c
+++ b/fs/ocfs2/dir.c
@@ -379,7 +379,7 @@ int ocfs2_do_extend_dir(struct super_block *sb,
 
 	status = ocfs2_extent_map_get_blocks(dir, (dir->i_blocks >>
 						   (sb->s_blocksize_bits - 9)),
-					     &p_blkno, NULL);
+					     &p_blkno, NULL, NULL);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
diff --git a/fs/ocfs2/extent_map.c b/fs/ocfs2/extent_map.c
index ea0ce41d419..eef6c188770 100644
--- a/fs/ocfs2/extent_map.c
+++ b/fs/ocfs2/extent_map.c
@@ -70,9 +70,11 @@ static int ocfs2_search_extent_list(struct ocfs2_extent_list *el,
 }
 
 int ocfs2_get_clusters(struct inode *inode, u32 v_cluster,
-		       u32 *p_cluster, u32 *num_clusters)
+		       u32 *p_cluster, u32 *num_clusters,
+		       unsigned int *extent_flags)
 {
 	int ret, i;
+	unsigned int flags = 0;
 	struct buffer_head *di_bh = NULL;
 	struct buffer_head *eb_bh = NULL;
 	struct ocfs2_dinode *di;
@@ -142,8 +144,13 @@ int ocfs2_get_clusters(struct inode *inode, u32 v_cluster,
 
 		if (num_clusters)
 			*num_clusters = ocfs2_rec_clusters(el, rec) - coff;
+
+		flags = rec->e_flags;
 	}
 
+	if (extent_flags)
+		*extent_flags = flags;
+
 out:
 	brelse(di_bh);
 	brelse(eb_bh);
@@ -155,7 +162,7 @@ out:
  * all while the map is in the process of being updated.
  */
 int ocfs2_extent_map_get_blocks(struct inode *inode, u64 v_blkno, u64 *p_blkno,
-				int *ret_count)
+				int *ret_count, unsigned int *extent_flags)
 {
 	int ret;
 	int bpc = ocfs2_clusters_to_blocks(inode->i_sb, 1);
@@ -164,7 +171,8 @@ int ocfs2_extent_map_get_blocks(struct inode *inode, u64 v_blkno, u64 *p_blkno,
 
 	cpos = ocfs2_blocks_to_clusters(inode->i_sb, v_blkno);
 
-	ret = ocfs2_get_clusters(inode, cpos, &p_cluster, &num_clusters);
+	ret = ocfs2_get_clusters(inode, cpos, &p_cluster, &num_clusters,
+				 extent_flags);
 	if (ret) {
 		mlog_errno(ret);
 		goto out;
diff --git a/fs/ocfs2/extent_map.h b/fs/ocfs2/extent_map.h
index 625d0ee5e04..0031c59c347 100644
--- a/fs/ocfs2/extent_map.h
+++ b/fs/ocfs2/extent_map.h
@@ -26,8 +26,8 @@
 #define _EXTENT_MAP_H
 
 int ocfs2_get_clusters(struct inode *inode, u32 v_cluster, u32 *p_cluster,
-		       u32 *num_clusters);
+		       u32 *num_clusters, unsigned int *extent_flags);
 int ocfs2_extent_map_get_blocks(struct inode *inode, u64 v_blkno, u64 *p_blkno,
-				int *ret_count);
+				int *ret_count, unsigned int *extent_flags);
 
 #endif  /* _EXTENT_MAP_H */
diff --git a/fs/ocfs2/file.c b/fs/ocfs2/file.c
index 36176018b4b..f516619a374 100644
--- a/fs/ocfs2/file.c
+++ b/fs/ocfs2/file.c
@@ -1127,6 +1127,7 @@ static int ocfs2_check_range_for_holes(struct inode *inode, loff_t pos,
 				       size_t count)
 {
 	int ret = 0;
+	unsigned int extent_flags;
 	u32 cpos, clusters, extent_len, phys_cpos;
 	struct super_block *sb = inode->i_sb;
 
@@ -1134,13 +1135,14 @@ static int ocfs2_check_range_for_holes(struct inode *inode, loff_t pos,
 	clusters = ocfs2_clusters_for_bytes(sb, pos + count) - cpos;
 
 	while (clusters) {
-		ret = ocfs2_get_clusters(inode, cpos, &phys_cpos, &extent_len);
+		ret = ocfs2_get_clusters(inode, cpos, &phys_cpos, &extent_len,
+					 &extent_flags);
 		if (ret < 0) {
 			mlog_errno(ret);
 			goto out;
 		}
 
-		if (phys_cpos == 0) {
+		if (phys_cpos == 0 || (extent_flags & OCFS2_EXT_UNWRITTEN)) {
 			ret = 1;
 			break;
 		}
diff --git a/fs/ocfs2/inode.c b/fs/ocfs2/inode.c
index 78c99b5050d..310049bf7f6 100644
--- a/fs/ocfs2/inode.c
+++ b/fs/ocfs2/inode.c
@@ -1105,7 +1105,8 @@ struct buffer_head *ocfs2_bread(struct inode *inode,
 		return NULL;
 	}
 
-	tmperr = ocfs2_extent_map_get_blocks(inode, block, &p_blkno, NULL);
+	tmperr = ocfs2_extent_map_get_blocks(inode, block, &p_blkno, NULL,
+					     NULL);
 	if (tmperr < 0) {
 		mlog_errno(tmperr);
 		goto fail;
diff --git a/fs/ocfs2/journal.c b/fs/ocfs2/journal.c
index 2e2e04fe973..db77e0996bb 100644
--- a/fs/ocfs2/journal.c
+++ b/fs/ocfs2/journal.c
@@ -670,7 +670,7 @@ static int ocfs2_force_read_journal(struct inode *inode)
 	       (inode->i_blocks >> (inode->i_sb->s_blocksize_bits - 9))) {
 
 		status = ocfs2_extent_map_get_blocks(inode, v_blkno,
-						     &p_blkno, &p_blocks);
+						     &p_blkno, &p_blocks, NULL);
 		if (status < 0) {
 			mlog_errno(status);
 			goto bail;
diff --git a/fs/ocfs2/namei.c b/fs/ocfs2/namei.c
index 5755e074825..395859edb51 100644
--- a/fs/ocfs2/namei.c
+++ b/fs/ocfs2/namei.c
@@ -1511,7 +1511,8 @@ static int ocfs2_create_symlink_data(struct ocfs2_super *osb,
 		goto bail;
 	}
 
-	status = ocfs2_extent_map_get_blocks(inode, 0, &p_blkno, &p_blocks);
+	status = ocfs2_extent_map_get_blocks(inode, 0, &p_blkno, &p_blocks,
+					     NULL);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;
diff --git a/fs/ocfs2/slot_map.c b/fs/ocfs2/slot_map.c
index f4416e7330e..d921a28329d 100644
--- a/fs/ocfs2/slot_map.c
+++ b/fs/ocfs2/slot_map.c
@@ -197,7 +197,7 @@ int ocfs2_init_slot_info(struct ocfs2_super *osb)
 		goto bail;
 	}
 
-	status = ocfs2_extent_map_get_blocks(inode, 0ULL, &blkno, NULL);
+	status = ocfs2_extent_map_get_blocks(inode, 0ULL, &blkno, NULL, NULL);
 	if (status < 0) {
 		mlog_errno(status);
 		goto bail;