Re: [PATCH] nfsd: add a spinlock to nfs4_stid

["J. Bruce Fields" <bfields@fieldses.org>]

On Thu, Oct 01, 2015 at 09:07:30AM -0400, Jeff Layton wrote:
> On Thu,  1 Oct 2015 09:05:50 -0400
> Jeff Layton <jlayton@poochiereds.net> wrote:
> 
> > ...to allow seqid morphing to be serialized wrt to its copying.
> > 
> > Signed-off-by: Jeff Layton <jeff.layton@primarydata.com>
> > ---
> >  fs/nfsd/nfs4state.c |  1 +
> >  fs/nfsd/state.h     | 13 +++++++------
> >  2 files changed, 8 insertions(+), 6 deletions(-)
> > 
> > diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
> > index bec2cc7eac78..156fc9183728 100644
> > --- a/fs/nfsd/nfs4state.c
> > +++ b/fs/nfsd/nfs4state.c
> > @@ -577,6 +577,7 @@ struct nfs4_stid *nfs4_alloc_stid(struct nfs4_client *cl,
> >  	stid->sc_stateid.si_opaque.so_clid = cl->cl_clientid;
> >  	/* Will be incremented before return to client: */
> >  	atomic_set(&stid->sc_count, 1);
> > +	spin_lock_init(&stid->sc_lock);
> >  
> >  	/*
> >  	 * It shouldn't be a problem to reuse an opaque stateid value.
> > diff --git a/fs/nfsd/state.h b/fs/nfsd/state.h
> > index 6f35fec88517..e3a191ea771c 100644
> > --- a/fs/nfsd/state.h
> > +++ b/fs/nfsd/state.h
> > @@ -84,7 +84,7 @@ struct nfsd4_callback_ops {
> >   * fields that are of general use to any stateid.
> >   */
> >  struct nfs4_stid {
> > -	atomic_t sc_count;
> > +	atomic_t		sc_count;
> >  #define NFS4_OPEN_STID 1
> >  #define NFS4_LOCK_STID 2
> >  #define NFS4_DELEG_STID 4
> > @@ -94,11 +94,12 @@ struct nfs4_stid {
> >  #define NFS4_REVOKED_DELEG_STID 16
> >  #define NFS4_CLOSED_DELEG_STID 32
> >  #define NFS4_LAYOUT_STID 64
> > -	unsigned char sc_type;
> > -	stateid_t sc_stateid;
> > -	struct nfs4_client *sc_client;
> > -	struct nfs4_file *sc_file;
> > -	void (*sc_free)(struct nfs4_stid *);
> > +	unsigned char		sc_type;
> > +	stateid_t		sc_stateid;
> > +	spinlock_t		sc_lock;
> > +	struct nfs4_client	*sc_client;
> > +	struct nfs4_file	*sc_file;
> > +	void			(*sc_free)(struct nfs4_stid *);
> >  };
> >  
> >  /*
> 
> Obviously the previously sent patch should be squashed in with this
> one. I meant to squash them before sending, but forgot. I'll do that if
> I need to respin though.

Thanks, applied; see my for-4.4-incoming branch to make sure I got it
right....

I'm waffling a bit, but currently thinking I'll leave it for 4.4 given
it's a relatively small race that we haven't seen in the wild.

--b.