nfs_localio: change nfsd_file_put_local() to take a pointer to __rcu pointer

Instead of calling xchg() and unrcu_pointer() before
nfsd_file_put_local(), we now pass pointer to the __rcu pointer and call
xchg() and unrcu_pointer() inside that function.

Where unrcu_pointer() is currently called the internals of "struct
nfsd_file" are not known and that causes older compilers such as gcc-8
to complain.

In some cases we have a __kernel (aka normal) pointer not an __rcu
pointer so we need to cast it to __rcu first. This is strictly a
weakening so no information is lost. Somewhat surprisingly, this cast
is accepted by gcc-8.

This has the pleasing result that the cmpxchg() which sets ro_file and
rw_file, and also the xchg() which clears them, are both now in the nfsd
code.

Reported-by: Pali Rohár <[e-mail]>
Reported-by: Vincent Mailhol <[e-mail]>
Fixes: 86e00412254a ("nfs: cache all open LOCALIO nfsd_file(s) in client")
Signed-off-by: NeilBrown <[e-mail]>
Signed-off-by: Anna Schumaker <[e-mail]>

fs/nfs/localio.c

···
 }
 EXPORT_SYMBOL_GPL(nfs_local_probe_async);
 
-static inline void nfs_local_file_put(struct nfsd_file *nf)
+static inline void nfs_local_file_put(struct nfsd_file *localio)
 {
-	nfs_to_nfsd_file_put_local(nf);
+	/* nfs_to_nfsd_file_put_local() expects an __rcu pointer
+	 * but we have a __kernel pointer.  It is always safe
+	 * to cast a __kernel pointer to an __rcu pointer
+	 * because the cast only weakens what is known about the pointer.
+	 */
+	struct nfsd_file __rcu *nf = (struct nfsd_file __rcu*) localio;
+
+	nfs_to_nfsd_file_put_local(&nf);
 }
 
 /*

fs/nfs_common/nfslocalio.c

···
 	while ((nfl = list_first_entry_or_null(&nfs_uuid->files,
 					       struct nfs_file_localio,
 					       list)) != NULL) {
-		struct nfsd_file *ro_nf;
-		struct nfsd_file *rw_nf;
-
 		/* If nfs_uuid is already NULL, nfs_close_local_fh is
 		 * closing and we must wait, else we unlink and close.
 		 */
···
 			continue;
 		}
 
-		ro_nf = unrcu_pointer(xchg(&nfl->ro_file, NULL));
-		rw_nf = unrcu_pointer(xchg(&nfl->rw_file, NULL));
-
 		/* Remove nfl from nfs_uuid->files list */
 		list_del_init(&nfl->list);
 		spin_unlock(&nfs_uuid->lock);
-		if (ro_nf)
-			nfs_to_nfsd_file_put_local(ro_nf);
-		if (rw_nf)
-			nfs_to_nfsd_file_put_local(rw_nf);
+
+		nfs_to_nfsd_file_put_local(&nfl->ro_file);
+		nfs_to_nfsd_file_put_local(&nfl->rw_file);
 		cond_resched();
+
 		spin_lock(&nfs_uuid->lock);
 		/* Now we can allow racing nfs_close_local_fh() to
 		 * skip the locking.
···
 
 void nfs_close_local_fh(struct nfs_file_localio *nfl)
 {
-	struct nfsd_file *ro_nf;
-	struct nfsd_file *rw_nf;
 	nfs_uuid_t *nfs_uuid;
 
 	rcu_read_lock();
···
 	spin_unlock(&nfs_uuid->lock);
 	rcu_read_unlock();
 
-	ro_nf = unrcu_pointer(xchg(&nfl->ro_file, NULL));
-	rw_nf = unrcu_pointer(xchg(&nfl->rw_file, NULL));
-	if (ro_nf)
-		nfs_to_nfsd_file_put_local(ro_nf);
-	if (rw_nf)
-		nfs_to_nfsd_file_put_local(rw_nf);
+	nfs_to_nfsd_file_put_local(&nfl->ro_file);
+	nfs_to_nfsd_file_put_local(&nfl->rw_file);
 
 	/* Remove nfl from nfs_uuid->files list and signal nfs_uuid_put()
 	 * that we are done.  The moment we drop the spinlock the

fs/nfsd/filecache.c

···
  * the reference of the nfsd_file.
  */
 struct net *
-nfsd_file_put_local(struct nfsd_file *nf)
+nfsd_file_put_local(struct nfsd_file __rcu **pnf)
 {
-	struct net *net = nf->nf_net;
+	struct nfsd_file *nf;
+	struct net *net = NULL;
 
-	nfsd_file_put(nf);
+	nf = unrcu_pointer(xchg(pnf, NULL));
+	if (nf) {
+		net = nf->nf_net;
+		nfsd_file_put(nf);
+	}
 	return net;
 }
 

fs/nfsd/filecache.h

···
 int nfsd_file_cache_start_net(struct net *net);
 void nfsd_file_cache_shutdown_net(struct net *net);
 void nfsd_file_put(struct nfsd_file *nf);
-struct net *nfsd_file_put_local(struct nfsd_file *nf);
+struct net *nfsd_file_put_local(struct nfsd_file __rcu **nf);
 struct nfsd_file *nfsd_file_get_local(struct nfsd_file *nf);
 struct nfsd_file *nfsd_file_get(struct nfsd_file *nf);
 struct file *nfsd_file_file(struct nfsd_file *nf);

include/linux/nfslocalio.h

···
 				    spinlock_t *nn_local_clients_lock);
 
 /* localio needs to map filehandle -> struct nfsd_file */
-extern struct nfsd_file *
-nfsd_open_local_fh(struct net *, struct auth_domain *, struct rpc_clnt *,
-		   const struct cred *, const struct nfs_fh *,
-		   const fmode_t) __must_hold(rcu);
 void nfs_close_local_fh(struct nfs_file_localio *);
 
 struct nfsd_localio_operations {
···
 						struct rpc_clnt *,
 						const struct cred *,
 						const struct nfs_fh *,
+						struct nfsd_file __rcu **pnf,
 						const fmode_t);
-	struct net *(*nfsd_file_put_local)(struct nfsd_file *);
+	struct net *(*nfsd_file_put_local)(struct nfsd_file __rcu **);
 	struct nfsd_file *(*nfsd_file_get_local)(struct nfsd_file *);
 	struct file *(*nfsd_file_file)(struct nfsd_file *);
 } ____cacheline_aligned;
···
 struct nfsd_file *nfs_open_local_fh(nfs_uuid_t *,
 		   struct rpc_clnt *, const struct cred *,
 		   const struct nfs_fh *, struct nfs_file_localio *,
+		   struct nfsd_file __rcu **pnf,
 		   const fmode_t);
 
 static inline void nfs_to_nfsd_net_put(struct net *net)
···
 	rcu_read_unlock();
 }
 
-static inline void nfs_to_nfsd_file_put_local(struct nfsd_file *localio)
+static inline void nfs_to_nfsd_file_put_local(struct nfsd_file __rcu **localio)
 {
 	/*
-	 * Must not hold RCU otherwise nfsd_file_put() can easily trigger:
-	 * "Voluntary context switch within RCU read-side critical section!"
-	 * by scheduling deep in underlying filesystem (e.g. XFS).
+	 * Either *localio must be guaranteed to be non-NULL, or caller
+	 * must prevent nfsd shutdown from completing as nfs_close_local_fh()
+	 * does by blocking the nfs_uuid from being finally put.
 	 */
-	struct net *net = nfs_to->nfsd_file_put_local(localio);
+	struct net *net;
 
-	nfs_to_nfsd_net_put(net);
+	net = nfs_to->nfsd_file_put_local(localio);
+
+	if (net)
+		nfs_to_nfsd_net_put(net);
 }
 
 #else   /* CONFIG_NFS_LOCALIO */