]> git.ipfire.org Git - thirdparty/xfsprogs-dev.git/commitdiff
writeback: Drop I_DIRTY_TIME_EXPIRE
authorJan Kara <jack@suse.cz>
Tue, 15 Sep 2020 19:59:38 +0000 (15:59 -0400)
committerEric Sandeen <sandeen@sandeen.net>
Tue, 15 Sep 2020 19:59:38 +0000 (15:59 -0400)
Source kernel commit: 5fcd57505c002efc5823a7355e21f48dd02d5a51

The only use of I_DIRTY_TIME_EXPIRE is to detect in
__writeback_single_inode() that inode got there because flush worker
decided it's time to writeback the dirty inode time stamps (either
because we are syncing or because of age). However we can detect this
directly in __writeback_single_inode() and there's no need for the
strange propagation with I_DIRTY_TIME_EXPIRE flag.

Reviewed-by: Christoph Hellwig <hch@lst.de>
Signed-off-by: Jan Kara <jack@suse.cz>
Signed-off-by: Eric Sandeen <sandeen@sandeen.net>
libxfs/xfs_trans_inode.c

index 034eef9c74b1d9ddc64756648acc2d47cbe06422..82716ea27f926840e399656eb9d9236cf53467a0 100644 (file)
@@ -107,9 +107,9 @@ xfs_trans_log_inode(
         * to log the timestamps, or will clear already cleared fields in the
         * worst case.
         */
-       if (inode->i_state & (I_DIRTY_TIME | I_DIRTY_TIME_EXPIRED)) {
+       if (inode->i_state & I_DIRTY_TIME) {
                spin_lock(&inode->i_lock);
-               inode->i_state &= ~(I_DIRTY_TIME | I_DIRTY_TIME_EXPIRED);
+               inode->i_state &= ~I_DIRTY_TIME;
                spin_unlock(&inode->i_lock);
        }