Re: Does ceph_fill_inode() mishandle I_NEW?
From: David Howells
Date: Thu Mar 13 2025 - 19:48:24 EST
Viacheslav Dubeyko <Slava.Dubeyko@xxxxxxx> wrote:
> As far as I can see, ceph_fill_inode() has comment: "Populate an inode based
> on info from mds. May be called on new or existing inodes". It sounds to me
> that particular CephFS kernel client could have obsolete state of inode
> compared with MDS's state. And we need to "re-new" the existing inode with
> the actual state that we received from MDS side. My vision is that we need
> to take into account the distributed nature of Ceph and inode metadata can
> be updated from multiple CephFS kernel client instances. Am I right here?
As I mentioned in my reply to Jeff, I'm thinking of what happens in the event
that we have a file that has hard links in several directories in a situation
where several of those links are looked up simultaneously. Can we end up with
ceph_fill_inode() being run in parallel on several threads on the same inode?
Actually, the use of ci->i_ceph_lock looks like it should make it safe, now
that I look at it again.
David