Messages in this thread | | | Date | Wed, 9 Oct 2013 19:41:36 -0700 | Subject | Re: [RFC] do we want pipefs et.al. in /proc/filesystems? | From | Linus Torvalds <> |
| |
On Wed, Oct 9, 2013 at 5:47 PM, Al Viro <viro@zeniv.linux.org.uk> wrote: > > These days there's no reason to register the filesystem types > that can only be used for internal mounts - ia64 pfmfs, anon_inodes, > bdev, pipefs and sockfs, in the current tree. The only user-visible > effect of dropping those register_filesystem() would be shorter > /proc/filesystems - that bunch wouldn't be mentioned there anymore. > Does anything care about that? FWIW, the diff eliminating those > would be as below (net/socket.c init leaks on early failures; > I'd left that as-is - it's a separate story).
I like the patch, but I do worry that some user space thing uses this to check whether the particular filesystem module is loaded.
Those kinds of things can't use /proc/modules (because it might be built-in), but /proc/filesystems has traditionally contained everything.
That said, these particular filesystems I really don't think people would ever possibly check for, so I think it's fine.
Linus
| |