Btrfs: fix truncate down when no_holes feature is enabled
authorLiu Bo <bo.li.liu@oracle.com>
Thu, 1 Dec 2016 21:43:31 +0000 (13:43 -0800)
committerDavid Sterba <dsterba@suse.com>
Thu, 19 Jan 2017 17:02:22 +0000 (18:02 +0100)
For such a file mapping,

[0-4k][hole][8k-12k]

In NO_HOLES mode, we don't have the [hole] extent any more.
Commit c1aa45759e90 ("Btrfs: fix shrinking truncate when the no_holes feature is enabled")
 fixed disk isize not being updated in NO_HOLES mode when data is not flushed.

However, even if data has been flushed, we can still have trouble
in updating disk isize since we updated disk isize to 'start' of
the last evicted extent.

Reviewed-by: Chris Mason <clm@fb.com>
Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
Signed-off-by: David Sterba <dsterba@suse.com>
fs/btrfs/inode.c

index bed6cf5f67a43a2695c4a23d06142915abe0b41e..177e7284909c72d4911ad719db0e15480ec98798 100644 (file)
@@ -4505,8 +4505,19 @@ search_again:
                if (found_type > min_type) {
                        del_item = 1;
                } else {
-                       if (item_end < new_size)
+                       if (item_end < new_size) {
+                               /*
+                                * With NO_HOLES mode, for the following mapping
+                                *
+                                * [0-4k][hole][8k-12k]
+                                *
+                                * if truncating isize down to 6k, it ends up
+                                * isize being 8k.
+                                */
+                               if (btrfs_fs_incompat(root->fs_info, NO_HOLES))
+                                       last_size = new_size;
                                break;
+                       }
                        if (found_key.offset >= new_size)
                                del_item = 1;
                        else