[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <252465.1582142281@warthog.procyon.org.uk>
Date: Wed, 19 Feb 2020 19:58:01 +0000
From: David Howells <dhowells@...hat.com>
To: Linus Torvalds <torvalds@...ux-foundation.org>
Cc: dhowells@...hat.com, Geert Uytterhoeven <geert@...ux-m68k.org>,
Al Viro <viro@...iv.linux.org.uk>, coda@...cmu.edu,
linux-afs@...ts.infradead.org, CIFS <linux-cifs@...r.kernel.org>,
"open list:NFS, SUNRPC, AND..." <linux-nfs@...r.kernel.org>,
linux-btrfs <linux-btrfs@...r.kernel.org>,
Linux FS Devel <linux-fsdevel@...r.kernel.org>,
Linux Kernel Mailing List <linux-kernel@...r.kernel.org>
Subject: Re: [RFC PATCH] vfs: syscalls: Add create_automount() and remove_automount()
Linus Torvalds <torvalds@...ux-foundation.org> wrote:
> > Why don't you just use mkdir with S_ISVTX set, or something like that?
>
> Actually, since this is apparently a different filetype, the _logical_
> thing to do is to use "mknod()".
Actually, in many ways, they're more akin to symlinks (and are implemented as
symlinks with funny attributes). It's a shame that symlinkat() doesn't have
an at_flags parameter.
mknod() isn't otherwise supported on AFS as there aren't any UNIX special
files.
> You presumably need a new type _anyway_ for stat() and/or the filldir
> d_type field. Or do you always want to make it look exactly like a
> directory to all user space?
That's already dealt with. They're presented as directories with
STATX_ATTR_AUTOMOUNT showing when you call statx() on them. You can also use
readlink() to extract the target if they haven't been mounted over yet.
Inside the kernel, they have no ->lookup() op, so DCACHE_AUTODIR_TYPE is set
on the dentry and there's a ->d_automount() op. The inode has S_AUTOMOUNT
set. That's all taken care of when the inode is created and the dentry is
instantiated.
Davod
Powered by blists - more mailing lists