On Mon, Nov 5, 2018 at 4:33 PM J. Bruce Fields <bfields@xxxxxxxxxxxx> wrote:
>
> On Fri, Oct 19, 2018 at 11:29:02AM -0400, Olga Kornievskaia wrote:
> > From: Olga Kornievskaia <kolga@xxxxxxxxxx>
> >
> > Incoming stateid (used by a READ) could be a saved copy stateid.
> > On first use make it active and check that the copy has started
> > within the allowable lease time.
> >
> > Signed-off-by: Olga Kornievskaia <kolga@xxxxxxxxxx>
> > ---
> > fs/nfsd/nfs4state.c | 43 +++++++++++++++++++++++++++++++++++++++++++
> > 1 file changed, 43 insertions(+)
> >
> > diff --git a/fs/nfsd/nfs4state.c b/fs/nfsd/nfs4state.c
> > index 7764a8b..16359de 100644
> > --- a/fs/nfsd/nfs4state.c
> > +++ b/fs/nfsd/nfs4state.c
> > @@ -5206,6 +5206,47 @@ static __be32 nfsd4_validate_stateid(struct nfs4_client *cl, stateid_t *stateid)
> >
> > return 0;
> > }
> > +/*
> > + * A READ from an inter server to server COPY will have a
> > + * copy stateid. Return the parent nfs4_stid.
> > + */
> > +static __be32 _find_cpntf_state(struct nfsd_net *nn, stateid_t *st,
> > + struct nfs4_cpntf_state **cps)
> > +{
> > + struct nfs4_cpntf_state *state = NULL;
> > +
> > + if (st->si_opaque.so_clid.cl_id != nn->s2s_cp_cl_id)
> > + return nfserr_bad_stateid;
> > + spin_lock(&nn->s2s_cp_lock);
> > + state = idr_find(&nn->s2s_cp_stateids, st->si_opaque.so_id);
> > + spin_unlock(&nn->s2s_cp_lock);
> > + if (!state)
> > + return nfserr_bad_stateid;
> > + *cps = state;
> > + return 0;
> > +}
> > +
> > +static __be32 find_cpntf_state(struct nfsd_net *nn, stateid_t *st,
> > + struct nfs4_stid **stid)
> > +{
> > + __be32 status;
> > + struct nfs4_cpntf_state *cps = NULL;
> > +
> > + status = _find_cpntf_state(nn, st, &cps);
> > + if (status)
> > + return status;
> > +
> > + /* Did the inter server to server copy start in time? */
> > + if (cps->cp_active == false && !time_after(cps->cp_timeout, jiffies))
>
> Is there any other time limit, or can the destination server keeping
> using this stateid indefinitely as long as it manages to get its first
> READ in before cp_timeout?
I don't believe there is anything in the spec that prohibits the
stateid from being used once it was started within a timeout period (a
copy might take a long time). Then it's invalidated either by
OFFLOAD_CANCEL or the parent stateid going away.
>
> > + return nfserr_partner_no_auth;
> > + else
> > + cps->cp_active = true;
> > +
> > + *stid = cps->cp_p_stid;
> > + refcount_inc(&cps->cp_p_stid->sc_count);
>
> Does the caller hold some lock? If not, what's prevnting cps from being
> freed before we get around to incrementing sc_count here?
>
> I would have expected the refcount_inc to happen before we drop
> s2s_cp_lock, but, like I say, maybe I'm missing some other locking.
No you are not missing it, the lock is taken only to find it in the
list. I'll move the refcount_inc into the lock section.
>
> --b.
>
> > +
> > + return nfs_ok;
> > +}
> >
> > /*
> > * Checks for stateid operations
> > @@ -5238,6 +5279,8 @@ static __be32 nfsd4_validate_stateid(struct nfs4_client *cl, stateid_t *stateid)
> > status = nfsd4_lookup_stateid(cstate, stateid,
> > NFS4_DELEG_STID|NFS4_OPEN_STID|NFS4_LOCK_STID,
> > &s, nn);
> > + if (status == nfserr_bad_stateid)
> > + status = find_cpntf_state(nn, stateid, &s);
> > if (status)
> > return status;
> > status = nfsd4_stid_check_stateid_generation(stateid, s,
> > --
> > 1.8.3.1
