Hello
I found that after defrag a file, the corresponding inode's i_blocks
become incorrect. I hope
following description address the real reason.
i_blocks is decreased in 'btrfs_drop_extents' when something are
drop/truncated and 'run_delalloc_range' calls 'btrfs_drop_extents' at
first place, so 'run_delalloc_range' may decrease the i_blocks.
('btrfs_drop_extents' doesn't increase i_blocks). In normal write
operation, i_blocks is increased in advance in
'dirty_and_release_pages', then decreased in 'run_delalloc_range' , so
final result is correct. But when do file defrag or mmap write,
i_blocks is only decreased by 'run_delalloc_range', i_blocks may
underflow quickly.
Regards
YZ
_______________________________________________
Btrfs-devel mailing list
[email protected]
http://oss.oracle.com/mailman/listinfo/btrfs-devel
