https://bugzilla.kernel.org/show_bug.cgi?id=196405 --- Comment #27 from Paul Eggert (eggert@xxxxxxxxxxx) --- (In reply to Theodore Tso from comment #22) > Paul is asserting that there is vast > amount of breakage because ext4 can return an st_nlinks value of 1 on a > directory, to the extent that he believes we should withdrawing the > dir_nlinks feature. I think this gets my most recent proposal backwards. At the end of Comment 17, I proposed that the ext4 code act as if dir_nlink is always set. That's what the code has been doing for a decade anyway. All that's missing is documentation which says "the dir_nlink setting is irrelevant, and the file system always acts as if dir_nlink is set", or words to that effect. Although in hindsight perhaps the dir_nlink flag should have been implemented properly, it wasn't and there's little point to implementing it properly now: every application using ext4 must work with the current behavior anyway. -- You are receiving this mail because: You are watching the assignee of the bug.
