From: Wang Xiaoguang <wangxg.f...@cn.fujitsu.com>
When running inband dedup with balance, it's possible that inband dedup
still increase ref on extents which are in RO chunk.
This may cause either find_data_references() gives warning, or make
run_delayed_refs() return -EIO and cause trans abort.
The cause is, normal dedup_del() is only called at run_delayed_ref()
time, which is too late for balance case.
This patch fixes this bug by calling dedup_del() at extent searching
time of balance.
Signed-off-by: Wang Xiaoguang <wangxg.f...@cn.fujitsu.com>
---
fs/btrfs/relocation.c | 13 +++++++++++++
1 file changed, 13 insertions(+)
diff --git a/fs/btrfs/relocation.c b/fs/btrfs/relocation.c
index 2bd0011..913a7b1 100644
--- a/fs/btrfs/relocation.c
+++ b/fs/btrfs/relocation.c
@@ -31,6 +31,7 @@
#include "async-thread.h"
#include "free-space-cache.h"
#include "inode-map.h"
+#include "dedup.h"
/*
* backref_node, mapping_node and tree_block start with this
@@ -3909,6 +3910,7 @@ static noinline_for_stack int relocate_block_group(struct
reloc_control *rc)
struct btrfs_trans_handle *trans = NULL;
struct btrfs_path *path;
struct btrfs_extent_item *ei;
+ struct btrfs_fs_info *fs_info = rc->extent_root->fs_info;
u64 flags;
u32 item_size;
int ret;
@@ -4039,6 +4041,17 @@ restart:
if (rc->stage == MOVE_DATA_EXTENTS &&
(flags & BTRFS_EXTENT_FLAG_DATA)) {
rc->found_file_extent = 1;
+ /*
+ * This data extent will be replaced, but normal
+ * dedup_del() will only happen at run_delayed_ref()
+ * time, which is too late, so delete dedup hash early
+ * to prevent its ref get increased.
+ */
+ ret = btrfs_dedup_del(trans, fs_info, key.objectid);
+ if (ret < 0) {
+ err = ret;
+ break;
+ }
ret = relocate_data_extent(rc->data_inode,
&key, &rc->cluster);
if (ret < 0) {
--
2.7.1
--
To unsubscribe from this list: send the line "unsubscribe linux-btrfs" in
the body of a message to majord...@vger.kernel.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
