Currently in btrfs, for data space reservation, it does not update bytes_may_use in btrfs_update_reserved_bytes() and the decrease operation will be delayed to be done in extent_clear_unlock_delalloc(), for fallocate(2), decrease operation is even delayed to be done in end of btrfs_fallocate(), which is too late. Obviously such delay will cause unnecessary pressure to enospc system, in [PATCH 4/4], there is a simpel test script that can reveal such false ENOSPC bug.
So in this patch set, it will remove RESERVE_FREE, RESERVE_ALLOC and RESERVE_ALLOC_NO_ACCOUNT, and we always update bytes_may_use timely. I'll also commit a fstests test case for this issue. Wang Xiaoguang (4): btrfs: use correct offset for reloc_inode in prealloc_file_extent_cluster() btrfs: divide btrfs_update_reserved_bytes() into two functions btrfs: introduce new EXTENT_CLEAR_DATA_RESV flag btrfs: update btrfs_space_info's bytes_may_use timely fs/btrfs/ctree.h | 2 +- fs/btrfs/extent-tree.c | 131 ++++++++++++++++++++++++------------------------- fs/btrfs/extent_io.h | 1 + fs/btrfs/file.c | 26 +++++----- fs/btrfs/inode-map.c | 3 +- fs/btrfs/inode.c | 37 ++++++++++---- fs/btrfs/relocation.c | 17 +++++-- 7 files changed, 123 insertions(+), 94 deletions(-) -- 2.9.0 -- 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