On Fri, Aug 29, 2025 at 05:37:17PM +0100, Al Viro wrote: > It would be in v3, but I didn't feel like sending another 63-patch > mailbomb for the sake of these 4 changed commits (well, and a cosmetical > change in #33, with matching modification in #35, ending with both > being cleaner - with the same resulting tree after #35). > > These 4 do repace #59..#62 in v3. Speaking of v3 - does anybody have objections to the following? 1) allow ->show_path() to return -EOPNOTSUPP, interpreted as "fall back to default seq_path(...)"? E.g. kernfs_sop_show_path() could return that if there's no ->scops->show_path(). 2) pass the sodding escape set as explicit argument, made an argument of fs/namespace.c:show_path() as well. 3) similar for ->show_devname(). 4) ... and to hell with those string_unescape_inplace() calls in there.
