[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]
Message-ID: <32654.1504883749@warthog.procyon.org.uk>
Date: Fri, 08 Sep 2017 16:15:49 +0100
From: David Howells <dhowells@...hat.com>
To: Ian Kent <ikent@...hat.com>
Cc: dhowells@...hat.com, NeilBrown <neilb@...e.com>,
"viro@...iv.linux.org.uk" <viro@...iv.linux.org.uk>,
Andrew Morton <akpm@...ux-foundation.org>,
Jeff Layton <jlayton@...hat.com>,
Trond Myklebust <trondmy@...marydata.com>,
"linux-kernel@...r.kernel.org" <linux-kernel@...r.kernel.org>,
"mkoutny@...e.com" <mkoutny@...e.com>,
"linux-nfs@...r.kernel.org" <linux-nfs@...r.kernel.org>,
"linux-fsdevel@...r.kernel.org" <linux-fsdevel@...r.kernel.org>,
"Michael Kerrisk (man-pages)" <mtk.manpages@...il.com>
Subject: Re: Do we really need d_weak_revalidate???
Ian Kent <ikent@...hat.com> wrote:
> So far only David commented about using ENOENT rather than EREMOTE.
>
> I prefer ENOENT for this case myself and he didn't object when I
> explained why, David, any concerns?
Not really - it just seems EREMOTE is a better fit since there is something
there, we're just not allowed to follow it. This is different to a dangling
symlink IMO.
David
Powered by blists - more mailing lists