On Fri, Sep 12, 2025 at 09:22:05PM +0200, Miklos Szeredi wrote: > You are talking about the disconnected directory case? That can't > happen in this call path, since it's following a normal component from > a parent directory, which by definition isn't disconnected. Huh? No, just a lookup in a different place with server giving you the nodeid of the directory you had in that place in dcache. d_splice_alias() has no choice other than "move the existing alias into the right place" - directory inodes *can't* have multiple hashed dentries, period. > Just realized, that open_last_lookups() will bypass lookup_open() on > cached positive anyway, so really no point in handing that inside > lookup_open(). IIRC, when the last time it came up, it was along the lines of "could we call ->atomic_open() for positive dentries as well?" and I think it had been about FUSE.
