Messages in this thread Patch in this message | | | From | Jaegeuk Kim <> | Subject | [PATCH 06/11] f2fs: skip unnecessary data writes during fsync | Date | Fri, 25 Jul 2014 15:47:20 -0700 |
| |
This patch intends to improve the fsync performance by skipping remaining the recovery information, only when there is no data that we should recover.
Signed-off-by: Jaegeuk Kim <jaegeuk@kernel.org> --- fs/f2fs/file.c | 16 ++++++++++++++++ 1 file changed, 16 insertions(+)
diff --git a/fs/f2fs/file.c b/fs/f2fs/file.c index 7c652b3..121689a 100644 --- a/fs/f2fs/file.c +++ b/fs/f2fs/file.c @@ -133,6 +133,17 @@ int f2fs_sync_file(struct file *file, loff_t start, loff_t end, int datasync) return ret; } + /* + * if there is no written data, don't waste time to write recovery info. + */ + if (!is_inode_flag_set(fi, FI_APPEND_WRITE) && + !exist_written_data(sbi, inode->i_ino, APPEND_INO)) { + if (is_inode_flag_set(fi, FI_UPDATE_WRITE) && + exist_written_data(sbi, inode->i_ino, UPDATE_INO)) + goto flush_out; + goto out; + } + /* guarantee free sections for fsync */ f2fs_balance_fs(sbi); @@ -188,6 +199,11 @@ int f2fs_sync_file(struct file *file, loff_t start, loff_t end, int datasync) ret = wait_on_node_pages_writeback(sbi, inode->i_ino); if (ret) goto out; + + /* once recovery info is written, don't need to tack this */ + remove_dirty_inode(sbi, inode->i_ino, APPEND_INO); +flush_out: + remove_dirty_inode(sbi, inode->i_ino, UPDATE_INO); ret = f2fs_issue_flush(F2FS_SB(inode->i_sb)); } out: -- 1.8.5.2 (Apple Git-48)
| |