Merge commit 'linux-pnfs/nfs41-for-2.6.31' into nfsv41-for-2.6.31

fs/nfs/Kconfig

@@ -74,6 +74,15 @@ config NFS_V4
 
 	  If unsure, say N.
 
+config NFS_V4_1
+	bool "NFS client support for NFSv4.1 (DEVELOPER ONLY)"
+	depends on NFS_V4 && EXPERIMENTAL
+	help
+	  This option enables support for minor version 1 of the NFSv4 protocol
+	  (draft-ietf-nfsv4-minorversion1) in the kernel's NFS client.
+
+	  Unless you're an NFS developer, say N.
+
 config ROOT_NFS
 	bool "Root file system on NFS"
 	depends on NFS_FS=y && IP_PNP

fs/nfs/callback.c

@@ -17,6 +17,9 @@
 #include <linux/freezer.h>
 #include <linux/kthread.h>
 #include <linux/sunrpc/svcauth_gss.h>
+#if defined(CONFIG_NFS_V4_1)
+#include <linux/sunrpc/bc_xprt.h>
+#endif
 
 #include <net/inet_sock.h>
 
@@ -28,11 +31,12 @@
 
 struct nfs_callback_data {
 	unsigned int users;
+	struct svc_serv *serv;
 	struct svc_rqst *rqst;
 	struct task_struct *task;
 };
 
-static struct nfs_callback_data nfs_callback_info;
+static struct nfs_callback_data nfs_callback_info[NFS4_MAX_MINOR_VERSION + 1];
 static DEFINE_MUTEX(nfs_callback_mutex);
 static struct svc_program nfs4_callback_program;
 
@@ -56,10 +60,10 @@ module_param_call(callback_tcpport, param_set_port, param_get_int,
 		 &nfs_callback_set_tcpport, 0644);
 
 /*
- * This is the callback kernel thread.
+ * This is the NFSv4 callback kernel thread.
  */
 static int
-nfs_callback_svc(void *vrqstp)
+nfs4_callback_svc(void *vrqstp)
 {
 	int err, preverr = 0;
 	struct svc_rqst *rqstp = vrqstp;
@@ -97,20 +101,12 @@ nfs_callback_svc(void *vrqstp)
 }
 
 /*
- * Bring up the callback thread if it is not already up.
+ * Prepare to bring up the NFSv4 callback service
  */
-int nfs_callback_up(void)
+struct svc_rqst *
+nfs4_callback_up(struct svc_serv *serv)
 {
-	struct svc_serv *serv = NULL;
-	int ret = 0;
-
-	mutex_lock(&nfs_callback_mutex);
-	if (nfs_callback_info.users++ || nfs_callback_info.task != NULL)
-		goto out;
-	serv = svc_create(&nfs4_callback_program, NFS4_CALLBACK_BUFSIZE, NULL);
-	ret = -ENOMEM;
-	if (!serv)
-		goto out_err;
+	int ret;
 
 	ret = svc_create_xprt(serv, "tcp", PF_INET,
 				nfs_callback_set_tcpport, SVC_SOCK_ANONYMOUS);
@@ -131,23 +127,168 @@ int nfs_callback_up(void)
 		goto out_err;
 #endif	/* defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE) */
 
-	nfs_callback_info.rqst = svc_prepare_thread(serv, &serv->sv_pools[0]);
-	if (IS_ERR(nfs_callback_info.rqst)) {
-		ret = PTR_ERR(nfs_callback_info.rqst);
-		nfs_callback_info.rqst = NULL;
+	return svc_prepare_thread(serv, &serv->sv_pools[0]);
+
+out_err:
+	if (ret == 0)
+		ret = -ENOMEM;
+	return ERR_PTR(ret);
+}
+
+#if defined(CONFIG_NFS_V4_1)
+/*
+ * The callback service for NFSv4.1 callbacks
+ */
+static int
+nfs41_callback_svc(void *vrqstp)
+{
+	struct svc_rqst *rqstp = vrqstp;
+	struct svc_serv *serv = rqstp->rq_server;
+	struct rpc_rqst *req;
+	int error;
+	DEFINE_WAIT(wq);
+
+	set_freezable();
+
+	/*
+	 * FIXME: do we really need to run this under the BKL? If so, please
+	 * add a comment about what it's intended to protect.
+	 */
+	lock_kernel();
+	while (!kthread_should_stop()) {
+		prepare_to_wait(&serv->sv_cb_waitq, &wq, TASK_INTERRUPTIBLE);
+		spin_lock_bh(&serv->sv_cb_lock);
+		if (!list_empty(&serv->sv_cb_list)) {
+			req = list_first_entry(&serv->sv_cb_list,
+					struct rpc_rqst, rq_bc_list);
+			list_del(&req->rq_bc_list);
+			spin_unlock_bh(&serv->sv_cb_lock);
+			dprintk("Invoking bc_svc_process()\n");
+			error = bc_svc_process(serv, req, rqstp);
+			dprintk("bc_svc_process() returned w/ error code= %d\n",
+				error);
+		} else {
+			spin_unlock_bh(&serv->sv_cb_lock);
+			schedule();
+		}
+		finish_wait(&serv->sv_cb_waitq, &wq);
+	}
+	unlock_kernel();
+	return 0;
+}
+
+/*
+ * Bring up the NFSv4.1 callback service
+ */
+struct svc_rqst *
+nfs41_callback_up(struct svc_serv *serv, struct rpc_xprt *xprt)
+{
+	struct svc_xprt *bc_xprt;
+	struct svc_rqst *rqstp = ERR_PTR(-ENOMEM);
+
+	dprintk("--> %s\n", __func__);
+	/* Create a svc_sock for the service */
+	bc_xprt = svc_sock_create(serv, xprt->prot);
+	if (!bc_xprt)
+		goto out;
+
+	/*
+	 * Save the svc_serv in the transport so that it can
+	 * be referenced when the session backchannel is initialized
+	 */
+	serv->bc_xprt = bc_xprt;
+	xprt->bc_serv = serv;
+
+	INIT_LIST_HEAD(&serv->sv_cb_list);
+	spin_lock_init(&serv->sv_cb_lock);
+	init_waitqueue_head(&serv->sv_cb_waitq);
+	rqstp = svc_prepare_thread(serv, &serv->sv_pools[0]);
+	if (IS_ERR(rqstp))
+		svc_sock_destroy(bc_xprt);
+out:
+	dprintk("--> %s return %p\n", __func__, rqstp);
+	return rqstp;
+}
+
+static inline int nfs_minorversion_callback_svc_setup(u32 minorversion,
+		struct svc_serv *serv, struct rpc_xprt *xprt,
+		struct svc_rqst **rqstpp, int (**callback_svc)(void *vrqstp))
+{
+	if (minorversion) {
+		*rqstpp = nfs41_callback_up(serv, xprt);
+		*callback_svc = nfs41_callback_svc;
+	}
+	return minorversion;
+}
+
+static inline void nfs_callback_bc_serv(u32 minorversion, struct rpc_xprt *xprt,
+		struct nfs_callback_data *cb_info)
+{
+	if (minorversion)
+		xprt->bc_serv = cb_info->serv;
+}
+#else
+static inline int nfs_minorversion_callback_svc_setup(u32 minorversion,
+		struct svc_serv *serv, struct rpc_xprt *xprt,
+		struct svc_rqst **rqstpp, int (**callback_svc)(void *vrqstp))
+{
+	return 0;
+}
+
+static inline void nfs_callback_bc_serv(u32 minorversion, struct rpc_xprt *xprt,
+		struct nfs_callback_data *cb_info)
+{
+}
+#endif /* CONFIG_NFS_V4_1 */
+
+/*
+ * Bring up the callback thread if it is not already up.
+ */
+int nfs_callback_up(u32 minorversion, struct rpc_xprt *xprt)
+{
+	struct svc_serv *serv = NULL;
+	struct svc_rqst *rqstp;
+	int (*callback_svc)(void *vrqstp);
+	struct nfs_callback_data *cb_info = &nfs_callback_info[minorversion];
+	char svc_name[12];
+	int ret = 0;
+	int minorversion_setup;
+
+	mutex_lock(&nfs_callback_mutex);
+	if (cb_info->users++ || cb_info->task != NULL) {
+		nfs_callback_bc_serv(minorversion, xprt, cb_info);
+		goto out;
+	}
+	serv = svc_create(&nfs4_callback_program, NFS4_CALLBACK_BUFSIZE, NULL);
+	if (!serv) {
+		ret = -ENOMEM;
+		goto out_err;
+	}
+
+	minorversion_setup =  nfs_minorversion_callback_svc_setup(minorversion,
+					serv, xprt, &rqstp, &callback_svc);
+	if (!minorversion_setup) {
+		/* v4.0 callback setup */
+		rqstp = nfs4_callback_up(serv);
+		callback_svc = nfs4_callback_svc;
+	}
+
+	if (IS_ERR(rqstp)) {
+		ret = PTR_ERR(rqstp);
 		goto out_err;
 	}
 
 	svc_sock_update_bufs(serv);
 
-	nfs_callback_info.task = kthread_run(nfs_callback_svc,
-					     nfs_callback_info.rqst,
-					     "nfsv4-svc");
-	if (IS_ERR(nfs_callback_info.task)) {
-		ret = PTR_ERR(nfs_callback_info.task);
-		svc_exit_thread(nfs_callback_info.rqst);
-		nfs_callback_info.rqst = NULL;
-		nfs_callback_info.task = NULL;
+	sprintf(svc_name, "nfsv4.%u-svc", minorversion);
+	cb_info->serv = serv;
+	cb_info->rqst = rqstp;
+	cb_info->task = kthread_run(callback_svc, cb_info->rqst, svc_name);
+	if (IS_ERR(cb_info->task)) {
+		ret = PTR_ERR(cb_info->task);
+		svc_exit_thread(cb_info->rqst);
+		cb_info->rqst = NULL;
+		cb_info->task = NULL;
 		goto out_err;
 	}
 out:
@@ -164,22 +305,25 @@ out:
 out_err:
 	dprintk("NFS: Couldn't create callback socket or server thread; "
 		"err = %d\n", ret);
-	nfs_callback_info.users--;
+	cb_info->users--;
 	goto out;
 }
 
 /*
  * Kill the callback thread if it's no longer being used.
  */
-void nfs_callback_down(void)
+void nfs_callback_down(int minorversion)
 {
+	struct nfs_callback_data *cb_info = &nfs_callback_info[minorversion];
+
 	mutex_lock(&nfs_callback_mutex);
-	nfs_callback_info.users--;
-	if (nfs_callback_info.users == 0 && nfs_callback_info.task != NULL) {
-		kthread_stop(nfs_callback_info.task);
-		svc_exit_thread(nfs_callback_info.rqst);
-		nfs_callback_info.rqst = NULL;
-		nfs_callback_info.task = NULL;
+	cb_info->users--;
+	if (cb_info->users == 0 && cb_info->task != NULL) {
+		kthread_stop(cb_info->task);
+		svc_exit_thread(cb_info->rqst);
+		cb_info->serv = NULL;
+		cb_info->rqst = NULL;
+		cb_info->task = NULL;
 	}
 	mutex_unlock(&nfs_callback_mutex);
 }

fs/nfs/callback.h

@@ -20,13 +20,24 @@ enum nfs4_callback_procnum {
 enum nfs4_callback_opnum {
 	OP_CB_GETATTR = 3,
 	OP_CB_RECALL  = 4,
+/* Callback operations new to NFSv4.1 */
+	OP_CB_LAYOUTRECALL  = 5,
+	OP_CB_NOTIFY        = 6,
+	OP_CB_PUSH_DELEG    = 7,
+	OP_CB_RECALL_ANY    = 8,
+	OP_CB_RECALLABLE_OBJ_AVAIL = 9,
+	OP_CB_RECALL_SLOT   = 10,
+	OP_CB_SEQUENCE      = 11,
+	OP_CB_WANTS_CANCELLED = 12,
+	OP_CB_NOTIFY_LOCK   = 13,
+	OP_CB_NOTIFY_DEVICEID = 14,
 	OP_CB_ILLEGAL = 10044,
 };
 
 struct cb_compound_hdr_arg {
 	unsigned int taglen;
 	const char *tag;
-	unsigned int callback_ident;
+	unsigned int minorversion;
 	unsigned nops;
 };
 
@@ -59,16 +70,59 @@ struct cb_recallargs {
 	uint32_t truncate;
 };
 
+#if defined(CONFIG_NFS_V4_1)
+
+struct referring_call {
+	uint32_t			rc_sequenceid;
+	uint32_t			rc_slotid;
+};
+
+struct referring_call_list {
+	struct nfs4_sessionid		rcl_sessionid;
+	uint32_t			rcl_nrefcalls;
+	struct referring_call 		*rcl_refcalls;
+};
+
+struct cb_sequenceargs {
+	struct sockaddr			*csa_addr;
+	struct nfs4_sessionid		csa_sessionid;
+	uint32_t			csa_sequenceid;
+	uint32_t			csa_slotid;
+	uint32_t			csa_highestslotid;
+	uint32_t			csa_cachethis;
+	uint32_t			csa_nrclists;
+	struct referring_call_list	*csa_rclists;
+};
+
+struct cb_sequenceres {
+	__be32				csr_status;
+	struct nfs4_sessionid		csr_sessionid;
+	uint32_t			csr_sequenceid;
+	uint32_t			csr_slotid;
+	uint32_t			csr_highestslotid;
+	uint32_t			csr_target_highestslotid;
+};
+
+extern unsigned nfs4_callback_sequence(struct cb_sequenceargs *args,
+				       struct cb_sequenceres *res);
+
+#endif /* CONFIG_NFS_V4_1 */
+
 extern __be32 nfs4_callback_getattr(struct cb_getattrargs *args, struct cb_getattrres *res);
 extern __be32 nfs4_callback_recall(struct cb_recallargs *args, void *dummy);
 
 #ifdef CONFIG_NFS_V4
-extern int nfs_callback_up(void);
-extern void nfs_callback_down(void);
-#else
-#define nfs_callback_up()	(0)
-#define nfs_callback_down()	do {} while(0)
-#endif
+extern int nfs_callback_up(u32 minorversion, struct rpc_xprt *xprt);
+extern void nfs_callback_down(int minorversion);
+#endif /* CONFIG_NFS_V4 */
+
+/*
+ * nfs41: Callbacks are expected to not cause substantial latency,
+ * so we limit their concurrency to 1 by setting up the maximum number
+ * of slots for the backchannel.
+ */
+#define NFS41_BC_MIN_CALLBACKS 1
+#define NFS41_BC_MAX_CALLBACKS 1
 
 extern unsigned int nfs_callback_set_tcpport;
 extern unsigned short nfs_callback_tcpport;

fs/nfs/callback_proc.c

@@ -101,3 +101,130 @@ out:
 	dprintk("%s: exit with status = %d\n", __func__, ntohl(res));
 	return res;
 }
+
+#if defined(CONFIG_NFS_V4_1)
+
+/*
+ * Validate the sequenceID sent by the server.
+ * Return success if the sequenceID is one more than what we last saw on
+ * this slot, accounting for wraparound.  Increments the slot's sequence.
+ *
+ * We don't yet implement a duplicate request cache, so at this time
+ * we will log replays, and process them as if we had not seen them before,
+ * but we don't bump the sequence in the slot.  Not too worried about it,
+ * since we only currently implement idempotent callbacks anyway.
+ *
+ * We have a single slot backchannel at this time, so we don't bother
+ * checking the used_slots bit array on the table.  The lower layer guarantees
+ * a single outstanding callback request at a time.
+ */
+static int
+validate_seqid(struct nfs4_slot_table *tbl, u32 slotid, u32 seqid)
+{
+	struct nfs4_slot *slot;
+
+	dprintk("%s enter. slotid %d seqid %d\n",
+		__func__, slotid, seqid);
+
+	if (slotid > NFS41_BC_MAX_CALLBACKS)
+		return htonl(NFS4ERR_BADSLOT);
+
+	slot = tbl->slots + slotid;
+	dprintk("%s slot table seqid: %d\n", __func__, slot->seq_nr);
+
+	/* Normal */
+	if (likely(seqid == slot->seq_nr + 1)) {
+		slot->seq_nr++;
+		return htonl(NFS4_OK);
+	}
+
+	/* Replay */
+	if (seqid == slot->seq_nr) {
+		dprintk("%s seqid %d is a replay - no DRC available\n",
+			__func__, seqid);
+		return htonl(NFS4_OK);
+	}
+
+	/* Wraparound */
+	if (seqid == 1 && (slot->seq_nr + 1) == 0) {
+		slot->seq_nr = 1;
+		return htonl(NFS4_OK);
+	}
+
+	/* Misordered request */
+	return htonl(NFS4ERR_SEQ_MISORDERED);
+}
+
+/*
+ * Returns a pointer to a held 'struct nfs_client' that matches the server's
+ * address, major version number, and session ID.  It is the caller's
+ * responsibility to release the returned reference.
+ *
+ * Returns NULL if there are no connections with sessions, or if no session
+ * matches the one of interest.
+ */
+ static struct nfs_client *find_client_with_session(
+	const struct sockaddr *addr, u32 nfsversion,
+	struct nfs4_sessionid *sessionid)
+{
+	struct nfs_client *clp;
+
+	clp = nfs_find_client(addr, 4);
+	if (clp == NULL)
+		return NULL;
+
+	do {
+		struct nfs_client *prev = clp;
+
+		if (clp->cl_session != NULL) {
+			if (memcmp(clp->cl_session->sess_id.data,
+					sessionid->data,
+					NFS4_MAX_SESSIONID_LEN) == 0) {
+				/* Returns a held reference to clp */
+				return clp;
+			}
+		}
+		clp = nfs_find_client_next(prev);
+		nfs_put_client(prev);
+	} while (clp != NULL);
+
+	return NULL;
+}
+
+/* FIXME: referring calls should be processed */
+unsigned nfs4_callback_sequence(struct cb_sequenceargs *args,
+				struct cb_sequenceres *res)
+{
+	struct nfs_client *clp;
+	int i, status;
+
+	for (i = 0; i < args->csa_nrclists; i++)
+		kfree(args->csa_rclists[i].rcl_refcalls);
+	kfree(args->csa_rclists);
+
+	status = htonl(NFS4ERR_BADSESSION);
+	clp = find_client_with_session(args->csa_addr, 4, &args->csa_sessionid);
+	if (clp == NULL)
+		goto out;
+
+	status = validate_seqid(&clp->cl_session->bc_slot_table,
+				args->csa_slotid, args->csa_sequenceid);
+	if (status)
+		goto out_putclient;
+
+	memcpy(&res->csr_sessionid, &args->csa_sessionid,
+	       sizeof(res->csr_sessionid));
+	res->csr_sequenceid = args->csa_sequenceid;
+	res->csr_slotid = args->csa_slotid;
+	res->csr_highestslotid = NFS41_BC_MAX_CALLBACKS - 1;
+	res->csr_target_highestslotid = NFS41_BC_MAX_CALLBACKS - 1;
+
+out_putclient:
+	nfs_put_client(clp);
+out:
+	dprintk("%s: exit with status = %d\n", __func__, ntohl(status));
+	res->csr_status = status;
+	return res->csr_status;
+}
+
+#endif /* CONFIG_NFS_V4_1 */

fs/nfs/callback_xdr.c

@@ -20,6 +20,11 @@
 				2 + 2 + 3 + 3)
 #define CB_OP_RECALL_RES_MAXSZ	(CB_OP_HDR_RES_MAXSZ)
 
+#if defined(CONFIG_NFS_V4_1)
+#define CB_OP_SEQUENCE_RES_MAXSZ	(CB_OP_HDR_RES_MAXSZ + \
+					4 + 1 + 3)
+#endif /* CONFIG_NFS_V4_1 */
+
 #define NFSDBG_FACILITY NFSDBG_CALLBACK
 
 typedef __be32 (*callback_process_op_t)(void *, void *);
@@ -132,7 +137,6 @@ static __be32 decode_stateid(struct xdr_stream *xdr, nfs4_stateid *stateid)
 static __be32 decode_compound_hdr_arg(struct xdr_stream *xdr, struct cb_compound_hdr_arg *hdr)
 {
 	__be32 *p;
-	unsigned int minor_version;
 	__be32 status;
 
 	status = decode_string(xdr, &hdr->taglen, &hdr->tag);
@@ -147,15 +151,19 @@ static __be32 decode_compound_hdr_arg(struct xdr_stream *xdr, struct cb_compound
 	p = read_buf(xdr, 12);
 	if (unlikely(p == NULL))
 		return htonl(NFS4ERR_RESOURCE);
-	minor_version = ntohl(*p++);
-	/* Check minor version is zero. */
-	if (minor_version != 0) {
-		printk(KERN_WARNING "%s: NFSv4 server callback with illegal minor version %u!\n",
-				__func__, minor_version);
+	hdr->minorversion = ntohl(*p++);
+	/* Check minor version is zero or one. */
+	if (hdr->minorversion <= 1) {
+		p++;	/* skip callback_ident */
+	} else {
+		printk(KERN_WARNING "%s: NFSv4 server callback with "
+			"illegal minor version %u!\n",
+			__func__, hdr->minorversion);
 		return htonl(NFS4ERR_MINOR_VERS_MISMATCH);
 	}
-	hdr->callback_ident = ntohl(*p++);
 	hdr->nops = ntohl(*p);
+	dprintk("%s: minorversion %d nops %d\n", __func__,
+		hdr->minorversion, hdr->nops);
 	return 0;
 }
 
@@ -204,6 +212,122 @@ out:
 	return status;
 }
 
+#if defined(CONFIG_NFS_V4_1)
+
+static unsigned decode_sessionid(struct xdr_stream *xdr,
+				 struct nfs4_sessionid *sid)
+{
+	uint32_t *p;
+	int len = NFS4_MAX_SESSIONID_LEN;
+
+	p = read_buf(xdr, len);
+	if (unlikely(p == NULL))
+		return htonl(NFS4ERR_RESOURCE);;
+
+	memcpy(sid->data, p, len);
+	return 0;
+}
+
+static unsigned decode_rc_list(struct xdr_stream *xdr,
+			       struct referring_call_list *rc_list)
+{
+	uint32_t *p;
+	int i;
+	unsigned status;
+
+	status = decode_sessionid(xdr, &rc_list->rcl_sessionid);
+	if (status)
+		goto out;
+
+	status = htonl(NFS4ERR_RESOURCE);
+	p = read_buf(xdr, sizeof(uint32_t));
+	if (unlikely(p == NULL))
+		goto out;
+
+	rc_list->rcl_nrefcalls = ntohl(*p++);
+	if (rc_list->rcl_nrefcalls) {
+		p = read_buf(xdr,
+			     rc_list->rcl_nrefcalls * 2 * sizeof(uint32_t));
+		if (unlikely(p == NULL))
+			goto out;
+		rc_list->rcl_refcalls = kmalloc(rc_list->rcl_nrefcalls *
+						sizeof(*rc_list->rcl_refcalls),
+						GFP_KERNEL);
+		if (unlikely(rc_list->rcl_refcalls == NULL))
+			goto out;
+		for (i = 0; i < rc_list->rcl_nrefcalls; i++) {
+			rc_list->rcl_refcalls[i].rc_sequenceid = ntohl(*p++);
+			rc_list->rcl_refcalls[i].rc_slotid = ntohl(*p++);
+		}
+	}
+	status = 0;
+
+out:
+	return status;
+}
+
+static unsigned decode_cb_sequence_args(struct svc_rqst *rqstp,
+					struct xdr_stream *xdr,
+					struct cb_sequenceargs *args)
+{
+	uint32_t *p;
+	int i;
+	unsigned status;
+
+	status = decode_sessionid(xdr, &args->csa_sessionid);
+	if (status)
+		goto out;
+
+	status = htonl(NFS4ERR_RESOURCE);
+	p = read_buf(xdr, 5 * sizeof(uint32_t));
+	if (unlikely(p == NULL))
+		goto out;
+
+	args->csa_addr = svc_addr(rqstp);
+	args->csa_sequenceid = ntohl(*p++);
+	args->csa_slotid = ntohl(*p++);
+	args->csa_highestslotid = ntohl(*p++);
+	args->csa_cachethis = ntohl(*p++);
+	args->csa_nrclists = ntohl(*p++);
+	args->csa_rclists = NULL;
+	if (args->csa_nrclists) {
+		args->csa_rclists = kmalloc(args->csa_nrclists *
+					    sizeof(*args->csa_rclists),
+					    GFP_KERNEL);
+		if (unlikely(args->csa_rclists == NULL))
+			goto out;
+
+		for (i = 0; i < args->csa_nrclists; i++) {
+			status = decode_rc_list(xdr, &args->csa_rclists[i]);
+			if (status)
+				goto out_free;
+		}
+	}
+	status = 0;
+
+	dprintk("%s: sessionid %x:%x:%x:%x sequenceid %u slotid %u "
+		"highestslotid %u cachethis %d nrclists %u\n",
+		__func__,
+		((u32 *)&args->csa_sessionid)[0],
+		((u32 *)&args->csa_sessionid)[1],
+		((u32 *)&args->csa_sessionid)[2],
+		((u32 *)&args->csa_sessionid)[3],
+		args->csa_sequenceid, args->csa_slotid,
+		args->csa_highestslotid, args->csa_cachethis,
+		args->csa_nrclists);
+out:
+	dprintk("%s: exit with status = %d\n", __func__, ntohl(status));
+	return status;
+
+out_free:
+	for (i = 0; i < args->csa_nrclists; i++)
+		kfree(args->csa_rclists[i].rcl_refcalls);
+	kfree(args->csa_rclists);
+	goto out;
+}
+
+#endif /* CONFIG_NFS_V4_1 */
+
 static __be32 encode_string(struct xdr_stream *xdr, unsigned int len, const char *str)
 {
 	__be32 *p;
@@ -353,31 +477,134 @@ out:
 	return status;
 }
 
-static __be32 process_op(struct svc_rqst *rqstp,
+#if defined(CONFIG_NFS_V4_1)
+
+static unsigned encode_sessionid(struct xdr_stream *xdr,
+				 const struct nfs4_sessionid *sid)
+{
+	uint32_t *p;
+	int len = NFS4_MAX_SESSIONID_LEN;
+
+	p = xdr_reserve_space(xdr, len);
+	if (unlikely(p == NULL))
+		return htonl(NFS4ERR_RESOURCE);
+
+	memcpy(p, sid, len);
+	return 0;
+}
+
+static unsigned encode_cb_sequence_res(struct svc_rqst *rqstp,
+				       struct xdr_stream *xdr,
+				       const struct cb_sequenceres *res)
+{
+	uint32_t *p;
+	unsigned status = res->csr_status;
+
+	if (unlikely(status != 0))
+		goto out;
+
+	encode_sessionid(xdr, &res->csr_sessionid);
+
+	p = xdr_reserve_space(xdr, 4 * sizeof(uint32_t));
+	if (unlikely(p == NULL))
+		return htonl(NFS4ERR_RESOURCE);
+
+	*p++ = htonl(res->csr_sequenceid);
+	*p++ = htonl(res->csr_slotid);
+	*p++ = htonl(res->csr_highestslotid);
+	*p++ = htonl(res->csr_target_highestslotid);
+out:
+	dprintk("%s: exit with status = %d\n", __func__, ntohl(status));
+	return status;
+}
+
+static __be32
+preprocess_nfs41_op(int nop, unsigned int op_nr, struct callback_op **op)
+{
+	if (op_nr == OP_CB_SEQUENCE) {
+		if (nop != 0)
+			return htonl(NFS4ERR_SEQUENCE_POS);
+	} else {
+		if (nop == 0)
+			return htonl(NFS4ERR_OP_NOT_IN_SESSION);
+	}
+
+	switch (op_nr) {
+	case OP_CB_GETATTR:
+	case OP_CB_RECALL:
+	case OP_CB_SEQUENCE:
+		*op = &callback_ops[op_nr];
+		break;
+
+	case OP_CB_LAYOUTRECALL:
+	case OP_CB_NOTIFY_DEVICEID:
+	case OP_CB_NOTIFY:
+	case OP_CB_PUSH_DELEG:
+	case OP_CB_RECALL_ANY:
+	case OP_CB_RECALLABLE_OBJ_AVAIL:
+	case OP_CB_RECALL_SLOT:
+	case OP_CB_WANTS_CANCELLED:
+	case OP_CB_NOTIFY_LOCK:
+		return htonl(NFS4ERR_NOTSUPP);
+
+	default:
+		return htonl(NFS4ERR_OP_ILLEGAL);
+	}
+
+	return htonl(NFS_OK);
+}
+
+#else /* CONFIG_NFS_V4_1 */
+
+static __be32
+preprocess_nfs41_op(int nop, unsigned int op_nr, struct callback_op **op)
+{
+	return htonl(NFS4ERR_MINOR_VERS_MISMATCH);
+}
+
+#endif /* CONFIG_NFS_V4_1 */
+
+static __be32
+preprocess_nfs4_op(unsigned int op_nr, struct callback_op **op)
+{
+	switch (op_nr) {
+	case OP_CB_GETATTR:
+	case OP_CB_RECALL:
+		*op = &callback_ops[op_nr];
+		break;
+	default:
+		return htonl(NFS4ERR_OP_ILLEGAL);
+	}
+
+	return htonl(NFS_OK);
+}
+
+static __be32 process_op(uint32_t minorversion, int nop,
+		struct svc_rqst *rqstp,
 		struct xdr_stream *xdr_in, void *argp,
 		struct xdr_stream *xdr_out, void *resp)
 {
 	struct callback_op *op = &callback_ops[0];
 	unsigned int op_nr = OP_CB_ILLEGAL;
-	__be32 status = 0;
+	__be32 status;
 	long maxlen;
 	__be32 res;
 
 	dprintk("%s: start\n", __func__);
 	status = decode_op_hdr(xdr_in, &op_nr);
-	if (likely(status == 0)) {
-		switch (op_nr) {
-			case OP_CB_GETATTR:
-			case OP_CB_RECALL:
-				op = &callback_ops[op_nr];
-				break;
-			default:
-				op_nr = OP_CB_ILLEGAL;
-				op = &callback_ops[0];
-				status = htonl(NFS4ERR_OP_ILLEGAL);
-		}
+	if (unlikely(status)) {
+		status = htonl(NFS4ERR_OP_ILLEGAL);
+		goto out;
 	}
 
+	dprintk("%s: minorversion=%d nop=%d op_nr=%u\n",
+		__func__, minorversion, nop, op_nr);
+
+	status = minorversion ? preprocess_nfs41_op(nop, op_nr, &op) :
+				preprocess_nfs4_op(op_nr, &op);
+	if (status == htonl(NFS4ERR_OP_ILLEGAL))
+		op_nr = OP_CB_ILLEGAL;
+out:
 	maxlen = xdr_out->end - xdr_out->p;
 	if (maxlen > 0 && maxlen < PAGE_SIZE) {
 		if (likely(status == 0 && op->decode_args != NULL))
@@ -425,7 +652,8 @@ static __be32 nfs4_callback_compound(struct svc_rqst *rqstp, void *argp, void *r
 		return rpc_system_err;
 
 	while (status == 0 && nops != hdr_arg.nops) {
-		status = process_op(rqstp, &xdr_in, argp, &xdr_out, resp);
+		status = process_op(hdr_arg.minorversion, nops,
+				    rqstp, &xdr_in, argp, &xdr_out, resp);
 		nops++;
 	}
 
@@ -452,7 +680,15 @@ static struct callback_op callback_ops[] = {
 		.process_op = (callback_process_op_t)nfs4_callback_recall,
 		.decode_args = (callback_decode_arg_t)decode_recall_args,
 		.res_maxsize = CB_OP_RECALL_RES_MAXSZ,
-	}
+	},
+#if defined(CONFIG_NFS_V4_1)
+	[OP_CB_SEQUENCE] = {
+		.process_op = (callback_process_op_t)nfs4_callback_sequence,
+		.decode_args = (callback_decode_arg_t)decode_cb_sequence_args,
+		.encode_res = (callback_encode_res_t)encode_cb_sequence_res,
+		.res_maxsize = CB_OP_SEQUENCE_RES_MAXSZ,
+	},
+#endif /* CONFIG_NFS_V4_1 */
 };
 
 /*

fs/nfs/client.c

@@ -37,6 +37,7 @@
 #include <linux/in6.h>
 #include <net/ipv6.h>
 #include <linux/nfs_xdr.h>
+#include <linux/sunrpc/bc_xprt.h>
 
 #include <asm/system.h>
 
@@ -102,6 +103,7 @@ struct nfs_client_initdata {
 	size_t addrlen;
 	const struct nfs_rpc_ops *rpc_ops;
 	int proto;
+	u32 minorversion;
 };
 
 /*
@@ -120,12 +122,6 @@ static struct nfs_client *nfs_alloc_client(const struct nfs_client_initdata *cl_
 
 	clp->rpc_ops = cl_init->rpc_ops;
 
-	if (cl_init->rpc_ops->version == 4) {
-		if (nfs_callback_up() < 0)
-			goto error_2;
-		__set_bit(NFS_CS_CALLBACK, &clp->cl_res_state);
-	}
-
 	atomic_set(&clp->cl_count, 1);
 	clp->cl_cons_state = NFS_CS_INITING;
 
@@ -135,7 +131,7 @@ static struct nfs_client *nfs_alloc_client(const struct nfs_client_initdata *cl_
 	if (cl_init->hostname) {
 		clp->cl_hostname = kstrdup(cl_init->hostname, GFP_KERNEL);
 		if (!clp->cl_hostname)
-			goto error_3;
+			goto error_cleanup;
 	}
 
 	INIT_LIST_HEAD(&clp->cl_superblocks);
@@ -150,6 +146,7 @@ static struct nfs_client *nfs_alloc_client(const struct nfs_client_initdata *cl_
 	rpc_init_wait_queue(&clp->cl_rpcwaitq, "NFS client");
 	clp->cl_boot_time = CURRENT_TIME;
 	clp->cl_state = 1 << NFS4CLNT_LEASE_EXPIRED;
+	clp->cl_minorversion = cl_init->minorversion;
 #endif
 	cred = rpc_lookup_machine_cred();
 	if (!IS_ERR(cred))
@@ -159,10 +156,7 @@ static struct nfs_client *nfs_alloc_client(const struct nfs_client_initdata *cl_
 
 	return clp;
 
-error_3:
-	if (__test_and_clear_bit(NFS_CS_CALLBACK, &clp->cl_res_state))
-		nfs_callback_down();
-error_2:
+error_cleanup:
 	kfree(clp);
 error_0:
 	return NULL;
@@ -181,6 +175,35 @@ static void nfs4_shutdown_client(struct nfs_client *clp)
 #endif
 }
 
+/*
+ * Destroy the NFS4 callback service
+ */
+static void nfs4_destroy_callback(struct nfs_client *clp)
+{
+#ifdef CONFIG_NFS_V4
+	if (__test_and_clear_bit(NFS_CS_CALLBACK, &clp->cl_res_state))
+		nfs_callback_down(clp->cl_minorversion);
+#endif /* CONFIG_NFS_V4 */
+}
+
+/*
+ * Clears/puts all minor version specific parts from an nfs_client struct
+ * reverting it to minorversion 0.
+ */
+static void nfs4_clear_client_minor_version(struct nfs_client *clp)
+{
+#ifdef CONFIG_NFS_V4_1
+	if (nfs4_has_session(clp)) {
+		nfs4_destroy_session(clp->cl_session);
+		clp->cl_session = NULL;
+	}
+
+	clp->cl_call_sync = _nfs4_call_sync;
+#endif /* CONFIG_NFS_V4_1 */
+
+	nfs4_destroy_callback(clp);
+}
+
 /*
  * Destroy a shared client record
  */
@@ -188,6 +211,7 @@ static void nfs_free_client(struct nfs_client *clp)
 {
 	dprintk("--> nfs_free_client(%u)\n", clp->rpc_ops->version);
 
+	nfs4_clear_client_minor_version(clp);
 	nfs4_shutdown_client(clp);
 
 	nfs_fscache_release_client_cookie(clp);
@@ -196,9 +220,6 @@ static void nfs_free_client(struct nfs_client *clp)
 	if (!IS_ERR(clp->cl_rpcclient))
 		rpc_shutdown_client(clp->cl_rpcclient);
 
-	if (__test_and_clear_bit(NFS_CS_CALLBACK, &clp->cl_res_state))
-		nfs_callback_down();
-
 	if (clp->cl_machine_cred != NULL)
 		put_rpccred(clp->cl_machine_cred);
 
@@ -347,7 +368,8 @@ struct nfs_client *nfs_find_client(const struct sockaddr *addr, u32 nfsversion)
 		struct sockaddr *clap = (struct sockaddr *)&clp->cl_addr;
 
 		/* Don't match clients that failed to initialise properly */
-		if (clp->cl_cons_state != NFS_CS_READY)
+		if (!(clp->cl_cons_state == NFS_CS_READY ||
+		      clp->cl_cons_state == NFS_CS_SESSION_INITING))
 			continue;
 
 		/* Different NFS versions cannot share the same nfs_client */
@@ -420,7 +442,9 @@ static struct nfs_client *nfs_match_client(const struct nfs_client_initdata *dat
 
 		if (clp->cl_proto != data->proto)
 			continue;
-
+		/* Match nfsv4 minorversion */
+		if (clp->cl_minorversion != data->minorversion)
+			continue;
 		/* Match the full socket address */
 		if (!nfs_sockaddr_cmp(sap, clap))
 			continue;
@@ -478,7 +502,7 @@ found_client:
 		nfs_free_client(new);
 
 	error = wait_event_killable(nfs_client_active_wq,
-				clp->cl_cons_state != NFS_CS_INITING);
+				clp->cl_cons_state < NFS_CS_INITING);
 	if (error < 0) {
 		nfs_put_client(clp);
 		return ERR_PTR(-ERESTARTSYS);
@@ -499,12 +523,28 @@ found_client:
 /*
  * Mark a server as ready or failed
  */
-static void nfs_mark_client_ready(struct nfs_client *clp, int state)
+void nfs_mark_client_ready(struct nfs_client *clp, int state)
 {
 	clp->cl_cons_state = state;
 	wake_up_all(&nfs_client_active_wq);
 }
 
+/*
+ * With sessions, the client is not marked ready until after a
+ * successful EXCHANGE_ID and CREATE_SESSION.
+ *
+ * Map errors cl_cons_state errors to EPROTONOSUPPORT to indicate
+ * other versions of NFS can be tried.
+ */
+int nfs4_check_client_ready(struct nfs_client *clp)
+{
+	if (!nfs4_has_session(clp))
+		return 0;
+	if (clp->cl_cons_state < NFS_CS_READY)
+		return -EPROTONOSUPPORT;
+	return 0;
+}
+
 /*
  * Initialise the timeout values for a connection
  */
@@ -1049,6 +1089,61 @@ error:
 }
 
 #ifdef CONFIG_NFS_V4
+/*
+ * Initialize the NFS4 callback service
+ */
+static int nfs4_init_callback(struct nfs_client *clp)
+{
+	int error;
+
+	if (clp->rpc_ops->version == 4) {
+		if (nfs4_has_session(clp)) {
+			error = xprt_setup_backchannel(
+						clp->cl_rpcclient->cl_xprt,
+						NFS41_BC_MIN_CALLBACKS);
+			if (error < 0)
+				return error;
+		}
+
+		error = nfs_callback_up(clp->cl_minorversion,
+					clp->cl_rpcclient->cl_xprt);
+		if (error < 0) {
+			dprintk("%s: failed to start callback. Error = %d\n",
+				__func__, error);
+			return error;
+		}
+		__set_bit(NFS_CS_CALLBACK, &clp->cl_res_state);
+	}
+	return 0;
+}
+
+/*
+ * Initialize the minor version specific parts of an NFS4 client record
+ */
+static int nfs4_init_client_minor_version(struct nfs_client *clp)
+{
+	clp->cl_call_sync = _nfs4_call_sync;
+
+#if defined(CONFIG_NFS_V4_1)
+	if (clp->cl_minorversion) {
+		struct nfs4_session *session = NULL;
+		/*
+		 * Create the session and mark it expired.
+		 * When a SEQUENCE operation encounters the expired session
+		 * it will do session recovery to initialize it.
+		 */
+		session = nfs4_alloc_session(clp);
+		if (!session)
+			return -ENOMEM;
+
+		clp->cl_session = session;
+		clp->cl_call_sync = _nfs4_call_sync_session;
+	}
+#endif /* CONFIG_NFS_V4_1 */
+
+	return nfs4_init_callback(clp);
+}
+
 /*
  * Initialise an NFS4 client record
  */
@@ -1083,7 +1178,12 @@ static int nfs4_init_client(struct nfs_client *clp,
 	}
 	__set_bit(NFS_CS_IDMAP, &clp->cl_res_state);
 
-	nfs_mark_client_ready(clp, NFS_CS_READY);
+	error = nfs4_init_client_minor_version(clp);
+	if (error < 0)
+		goto error;
+
+	if (!nfs4_has_session(clp))
+		nfs_mark_client_ready(clp, NFS_CS_READY);
 	return 0;
 
 error:
@@ -1101,7 +1201,8 @@ static int nfs4_set_client(struct nfs_server *server,
 		const size_t addrlen,
 		const char *ip_addr,
 		rpc_authflavor_t authflavour,
-		int proto, const struct rpc_timeout *timeparms)
+		int proto, const struct rpc_timeout *timeparms,
+		u32 minorversion)
 {
 	struct nfs_client_initdata cl_init = {
 		.hostname = hostname,
@@ -1109,6 +1210,7 @@ static int nfs4_set_client(struct nfs_server *server,
 		.addrlen = addrlen,
 		.rpc_ops = &nfs_v4_clientops,
 		.proto = proto,
+		.minorversion = minorversion,
 	};
 	struct nfs_client *clp;
 	int error;
@@ -1137,6 +1239,36 @@ error:
 	return error;
 }
 
+/*
+ * Initialize a session.
+ * Note: save the mount rsize and wsize for create_server negotiation.
+ */
+static void nfs4_init_session(struct nfs_client *clp,
+			      unsigned int wsize, unsigned int rsize)
+{
+#if defined(CONFIG_NFS_V4_1)
+	if (nfs4_has_session(clp)) {
+		clp->cl_session->fc_attrs.max_rqst_sz = wsize;
+		clp->cl_session->fc_attrs.max_resp_sz = rsize;
+	}
+#endif /* CONFIG_NFS_V4_1 */
+}
+
+/*
+ * Session has been established, and the client marked ready.
+ * Set the mount rsize and wsize with negotiated fore channel
+ * attributes which will be bound checked in nfs_server_set_fsinfo.
+ */
+static void nfs4_session_set_rwsize(struct nfs_server *server)
+{
+#ifdef CONFIG_NFS_V4_1
+	if (!nfs4_has_session(server->nfs_client))
+		return;
+	server->rsize = server->nfs_client->cl_session->fc_attrs.max_resp_sz;
+	server->wsize = server->nfs_client->cl_session->fc_attrs.max_rqst_sz;
+#endif /* CONFIG_NFS_V4_1 */
+}
+
 /*
  * Create a version 4 volume record
  */
@@ -1164,7 +1296,8 @@ static int nfs4_init_server(struct nfs_server *server,
 			data->client_address,
 			data->auth_flavors[0],
 			data->nfs_server.protocol,
-			&timeparms);
+			&timeparms,
+			data->minorversion);
 	if (error < 0)
 		goto error;
 
@@ -1214,6 +1347,8 @@ struct nfs_server *nfs4_create_server(const struct nfs_parsed_mount_data *data,
 	BUG_ON(!server->nfs_client->rpc_ops);
 	BUG_ON(!server->nfs_client->rpc_ops->file_inode_ops);
 
+	nfs4_init_session(server->nfs_client, server->wsize, server->rsize);
+
 	/* Probe the root fh to retrieve its FSID */
 	error = nfs4_path_walk(server, mntfh, data->nfs_server.export_path);
 	if (error < 0)
@@ -1224,6 +1359,8 @@ struct nfs_server *nfs4_create_server(const struct nfs_parsed_mount_data *data,
 		(unsigned long long) server->fsid.minor);
 	dprintk("Mount FH: %d\n", mntfh->size);
 
+	nfs4_session_set_rwsize(server);
+
 	error = nfs_probe_fsinfo(server, mntfh, &fattr);
 	if (error < 0)
 		goto error;
@@ -1282,7 +1419,8 @@ struct nfs_server *nfs4_create_referral_server(struct nfs_clone_mount *data,
 				parent_client->cl_ipaddr,
 				data->authflavor,
 				parent_server->client->cl_xprt->prot,
-				parent_server->client->cl_timeout);
+				parent_server->client->cl_timeout,
+				parent_client->cl_minorversion);
 	if (error < 0)
 		goto error;
 

fs/nfs/direct.c

@@ -259,6 +259,9 @@ static void nfs_direct_read_release(void *calldata)
 }
 
 static const struct rpc_call_ops nfs_read_direct_ops = {
+#if defined(CONFIG_NFS_V4_1)
+	.rpc_call_prepare = nfs_read_prepare,
+#endif /* CONFIG_NFS_V4_1 */
 	.rpc_call_done = nfs_direct_read_result,
 	.rpc_release = nfs_direct_read_release,
 };
@@ -535,6 +538,9 @@ static void nfs_direct_commit_release(void *calldata)
 }
 
 static const struct rpc_call_ops nfs_commit_direct_ops = {
+#if defined(CONFIG_NFS_V4_1)
+	.rpc_call_prepare = nfs_write_prepare,
+#endif /* CONFIG_NFS_V4_1 */
 	.rpc_call_done = nfs_direct_commit_result,
 	.rpc_release = nfs_direct_commit_release,
 };
@@ -673,6 +679,9 @@ out_unlock:
 }
 
 static const struct rpc_call_ops nfs_write_direct_ops = {
+#if defined(CONFIG_NFS_V4_1)
+	.rpc_call_prepare = nfs_write_prepare,
+#endif /* CONFIG_NFS_V4_1 */
 	.rpc_call_done = nfs_direct_write_result,
 	.rpc_release = nfs_direct_write_release,
 };

fs/nfs/internal.h

@@ -2,6 +2,7 @@
  * NFS internal definitions
  */
 
+#include "nfs4_fs.h"
 #include <linux/mount.h>
 #include <linux/security.h>
 
@@ -17,6 +18,18 @@ struct nfs_string;
  */
 #define NFS_MAX_READAHEAD	(RPC_DEF_SLOT_TABLE - 1)
 
+/*
+ * Determine if sessions are in use.
+ */
+static inline int nfs4_has_session(const struct nfs_client *clp)
+{
+#ifdef CONFIG_NFS_V4_1
+	if (clp->cl_session)
+		return 1;
+#endif /* CONFIG_NFS_V4_1 */
+	return 0;
+}
+
 struct nfs_clone_mount {
 	const struct super_block *sb;
 	const struct dentry *dentry;
@@ -44,6 +57,7 @@ struct nfs_parsed_mount_data {
 	unsigned int		auth_flavor_len;
 	rpc_authflavor_t	auth_flavors[1];
 	char			*client_address;
+	unsigned int		minorversion;
 	char			*fscache_uniq;
 
 	struct {
@@ -99,6 +113,8 @@ extern void nfs_free_server(struct nfs_server *server);
 extern struct nfs_server *nfs_clone_server(struct nfs_server *,
 					   struct nfs_fh *,
 					   struct nfs_fattr *);
+extern void nfs_mark_client_ready(struct nfs_client *clp, int state);
+extern int nfs4_check_client_ready(struct nfs_client *clp);
 #ifdef CONFIG_PROC_FS
 extern int __init nfs_fs_proc_init(void);
 extern void nfs_fs_proc_exit(void);
@@ -146,6 +162,20 @@ extern __be32 * nfs_decode_dirent(__be32 *, struct nfs_entry *, int);
 extern struct rpc_procinfo nfs3_procedures[];
 extern __be32 *nfs3_decode_dirent(__be32 *, struct nfs_entry *, int);
 
+/* nfs4proc.c */
+static inline void nfs4_restart_rpc(struct rpc_task *task,
+				    const struct nfs_client *clp)
+{
+#ifdef CONFIG_NFS_V4_1
+	if (nfs4_has_session(clp) &&
+	    test_bit(NFS4CLNT_SESSION_SETUP, &clp->cl_state)) {
+		rpc_restart_call_prepare(task);
+		return;
+	}
+#endif /* CONFIG_NFS_V4_1 */
+	rpc_restart_call(task);
+}
+
 /* nfs4xdr.c */
 #ifdef CONFIG_NFS_V4
 extern __be32 *nfs4_decode_dirent(__be32 *p, struct nfs_entry *entry, int plus);
@@ -205,6 +235,38 @@ extern int nfs4_path_walk(struct nfs_server *server,
 			  const char *path);
 #endif
 
+/* read.c */
+extern void nfs_read_prepare(struct rpc_task *task, void *calldata);
+
+/* write.c */
+extern void nfs_write_prepare(struct rpc_task *task, void *calldata);
+
+/* nfs4proc.c */
+extern int _nfs4_call_sync(struct nfs_server *server,
+			   struct rpc_message *msg,
+			   struct nfs4_sequence_args *args,
+			   struct nfs4_sequence_res *res,
+			   int cache_reply);
+extern int _nfs4_call_sync_session(struct nfs_server *server,
+				   struct rpc_message *msg,
+				   struct nfs4_sequence_args *args,
+				   struct nfs4_sequence_res *res,
+				   int cache_reply);
+
+#ifdef CONFIG_NFS_V4_1
+extern void nfs41_sequence_free_slot(const struct nfs_client *,
+				     struct nfs4_sequence_res *res);
+#endif /* CONFIG_NFS_V4_1 */
+
+static inline void nfs4_sequence_free_slot(const struct nfs_client *clp,
+					   struct nfs4_sequence_res *res)
+{
+#ifdef CONFIG_NFS_V4_1
+	if (nfs4_has_session(clp))
+		nfs41_sequence_free_slot(clp, res);
+#endif /* CONFIG_NFS_V4_1 */
+}
+
 /*
  * Determine the device name as a string
  */

fs/nfs/nfs4_fs.h

@@ -44,6 +44,7 @@ enum nfs4_client_state {
 	NFS4CLNT_RECLAIM_REBOOT,
 	NFS4CLNT_RECLAIM_NOGRACE,
 	NFS4CLNT_DELEGRETURN,
+	NFS4CLNT_SESSION_SETUP,
 };
 
 /*
@@ -177,6 +178,14 @@ struct nfs4_state_recovery_ops {
 	int state_flag_bit;
 	int (*recover_open)(struct nfs4_state_owner *, struct nfs4_state *);
 	int (*recover_lock)(struct nfs4_state *, struct file_lock *);
+	int (*establish_clid)(struct nfs_client *, struct rpc_cred *);
+	struct rpc_cred * (*get_clid_cred)(struct nfs_client *);
+};
+
+struct nfs4_state_maintenance_ops {
+	int (*sched_state_renewal)(struct nfs_client *, struct rpc_cred *);
+	struct rpc_cred * (*get_state_renewal_cred_locked)(struct nfs_client *);
+	int (*renew_lease)(struct nfs_client *, struct rpc_cred *);
 };
 
 extern const struct dentry_operations nfs4_dentry_operations;
@@ -193,6 +202,7 @@ extern int nfs4_proc_setclientid(struct nfs_client *, u32, unsigned short, struc
 extern int nfs4_proc_setclientid_confirm(struct nfs_client *, struct rpc_cred *);
 extern int nfs4_proc_async_renew(struct nfs_client *, struct rpc_cred *);
 extern int nfs4_proc_renew(struct nfs_client *, struct rpc_cred *);
+extern int nfs4_init_clientid(struct nfs_client *, struct rpc_cred *);
 extern int nfs4_do_close(struct path *path, struct nfs4_state *state, int wait);
 extern struct dentry *nfs4_atomic_open(struct inode *, struct dentry *, struct nameidata *);
 extern int nfs4_open_revalidate(struct inode *, struct dentry *, int, struct nameidata *);
@@ -200,8 +210,26 @@ extern int nfs4_server_capabilities(struct nfs_server *server, struct nfs_fh *fh
 extern int nfs4_proc_fs_locations(struct inode *dir, const struct qstr *name,
 		struct nfs4_fs_locations *fs_locations, struct page *page);
 
-extern struct nfs4_state_recovery_ops nfs4_reboot_recovery_ops;
-extern struct nfs4_state_recovery_ops nfs4_nograce_recovery_ops;
+extern struct nfs4_state_recovery_ops *nfs4_reboot_recovery_ops[];
+extern struct nfs4_state_recovery_ops *nfs4_nograce_recovery_ops[];
+#if defined(CONFIG_NFS_V4_1)
+extern int nfs4_setup_sequence(struct nfs_client *clp,
+		struct nfs4_sequence_args *args, struct nfs4_sequence_res *res,
+		int cache_reply, struct rpc_task *task);
+extern void nfs4_destroy_session(struct nfs4_session *session);
+extern struct nfs4_session *nfs4_alloc_session(struct nfs_client *clp);
+extern int nfs4_proc_create_session(struct nfs_client *, int reset);
+extern int nfs4_proc_destroy_session(struct nfs4_session *);
+#else /* CONFIG_NFS_v4_1 */
+static inline int nfs4_setup_sequence(struct nfs_client *clp,
+		struct nfs4_sequence_args *args, struct nfs4_sequence_res *res,
+		int cache_reply, struct rpc_task *task)
+{
+	return 0;
+}
+#endif /* CONFIG_NFS_V4_1 */
+
+extern struct nfs4_state_maintenance_ops *nfs4_state_renewal_ops[];
 
 extern const u32 nfs4_fattr_bitmap[2];
 extern const u32 nfs4_statfs_bitmap[2];
@@ -216,7 +244,12 @@ extern void nfs4_kill_renewd(struct nfs_client *);
 extern void nfs4_renew_state(struct work_struct *);
 
 /* nfs4state.c */
+struct rpc_cred *nfs4_get_setclientid_cred(struct nfs_client *clp);
 struct rpc_cred *nfs4_get_renew_cred_locked(struct nfs_client *clp);
+#if defined(CONFIG_NFS_V4_1)
+struct rpc_cred *nfs4_get_machine_cred_locked(struct nfs_client *clp);
+struct rpc_cred *nfs4_get_exchange_id_cred(struct nfs_client *clp);
+#endif /* CONFIG_NFS_V4_1 */
 
 extern struct nfs4_state_owner * nfs4_get_state_owner(struct nfs_server *, struct rpc_cred *);
 extern void nfs4_put_state_owner(struct nfs4_state_owner *);

fs/nfs/nfs4renewd.c

@@ -59,12 +59,14 @@
 void
 nfs4_renew_state(struct work_struct *work)
 {
+	struct nfs4_state_maintenance_ops *ops;
 	struct nfs_client *clp =
 		container_of(work, struct nfs_client, cl_renewd.work);
 	struct rpc_cred *cred;
 	long lease, timeout;
 	unsigned long last, now;
 
+	ops = nfs4_state_renewal_ops[clp->cl_minorversion];
 	dprintk("%s: start\n", __func__);
 	/* Are there any active superblocks? */
 	if (list_empty(&clp->cl_superblocks))
@@ -76,7 +78,7 @@ nfs4_renew_state(struct work_struct *work)
 	timeout = (2 * lease) / 3 + (long)last - (long)now;
 	/* Are we close to a lease timeout? */
 	if (time_after(now, last + lease/3)) {
-		cred = nfs4_get_renew_cred_locked(clp);
+		cred = ops->get_state_renewal_cred_locked(clp);
 		spin_unlock(&clp->cl_lock);
 		if (cred == NULL) {
 			if (list_empty(&clp->cl_delegations)) {
@@ -86,7 +88,7 @@ nfs4_renew_state(struct work_struct *work)
 			nfs_expire_all_delegations(clp);
 		} else {
 			/* Queue an asynchronous RENEW. */
-			nfs4_proc_async_renew(clp, cred);
+			ops->sched_state_renewal(clp, cred);
 			put_rpccred(cred);
 		}
 		timeout = (2 * lease) / 3;

fs/nfs/nfs4state.c

@@ -60,7 +60,7 @@ const nfs4_stateid zero_stateid;
 
 static LIST_HEAD(nfs4_clientid_list);
 
-static int nfs4_init_client(struct nfs_client *clp, struct rpc_cred *cred)
+int nfs4_init_clientid(struct nfs_client *clp, struct rpc_cred *cred)
 {
 	unsigned short port;
 	int status;
@@ -77,7 +77,7 @@ static int nfs4_init_client(struct nfs_client *clp, struct rpc_cred *cred)
 	return status;
 }
 
-static struct rpc_cred *nfs4_get_machine_cred_locked(struct nfs_client *clp)
+struct rpc_cred *nfs4_get_machine_cred_locked(struct nfs_client *clp)
 {
 	struct rpc_cred *cred = NULL;
 
@@ -114,17 +114,21 @@ struct rpc_cred *nfs4_get_renew_cred_locked(struct nfs_client *clp)
 	return cred;
 }
 
-static struct rpc_cred *nfs4_get_renew_cred(struct nfs_client *clp)
+#if defined(CONFIG_NFS_V4_1)
+
+struct rpc_cred *nfs4_get_exchange_id_cred(struct nfs_client *clp)
 {
 	struct rpc_cred *cred;
 
 	spin_lock(&clp->cl_lock);
-	cred = nfs4_get_renew_cred_locked(clp);
+	cred = nfs4_get_machine_cred_locked(clp);
 	spin_unlock(&clp->cl_lock);
 	return cred;
 }
 
-static struct rpc_cred *nfs4_get_setclientid_cred(struct nfs_client *clp)
+#endif /* CONFIG_NFS_V4_1 */
+
+struct rpc_cred *nfs4_get_setclientid_cred(struct nfs_client *clp)
 {
 	struct nfs4_state_owner *sp;
 	struct rb_node *pos;
@@ -738,12 +742,14 @@ static void nfs_increment_seqid(int status, struct nfs_seqid *seqid)
 
 void nfs_increment_open_seqid(int status, struct nfs_seqid *seqid)
 {
-	if (status == -NFS4ERR_BAD_SEQID) {
-		struct nfs4_state_owner *sp = container_of(seqid->sequence,
-				struct nfs4_state_owner, so_seqid);
+	struct nfs4_state_owner *sp = container_of(seqid->sequence,
+					struct nfs4_state_owner, so_seqid);
+	struct nfs_server *server = sp->so_server;
+
+	if (status == -NFS4ERR_BAD_SEQID)
 		nfs4_drop_state_owner(sp);
-	}
-	nfs_increment_seqid(status, seqid);
+	if (!nfs4_has_session(server->nfs_client))
+		nfs_increment_seqid(status, seqid);
 }
 
 /*
@@ -1042,6 +1048,14 @@ static void nfs4_recovery_handle_error(struct nfs_client *clp, int error)
 		case -NFS4ERR_EXPIRED:
 			set_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state);
 			nfs4_state_start_reclaim_nograce(clp);
+		case -NFS4ERR_BADSESSION:
+		case -NFS4ERR_BADSLOT:
+		case -NFS4ERR_BAD_HIGH_SLOT:
+		case -NFS4ERR_DEADSESSION:
+		case -NFS4ERR_CONN_NOT_BOUND_TO_SESSION:
+		case -NFS4ERR_SEQ_FALSE_RETRY:
+		case -NFS4ERR_SEQ_MISORDERED:
+			set_bit(NFS4CLNT_SESSION_SETUP, &clp->cl_state);
 	}
 }
 
@@ -1075,18 +1089,22 @@ restart:
 static int nfs4_check_lease(struct nfs_client *clp)
 {
 	struct rpc_cred *cred;
+	struct nfs4_state_maintenance_ops *ops =
+		nfs4_state_renewal_ops[clp->cl_minorversion];
 	int status = -NFS4ERR_EXPIRED;
 
 	/* Is the client already known to have an expired lease? */
 	if (test_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state))
 		return 0;
-	cred = nfs4_get_renew_cred(clp);
+	spin_lock(&clp->cl_lock);
+	cred = ops->get_state_renewal_cred_locked(clp);
+	spin_unlock(&clp->cl_lock);
 	if (cred == NULL) {
 		cred = nfs4_get_setclientid_cred(clp);
 		if (cred == NULL)
 			goto out;
 	}
-	status = nfs4_proc_renew(clp, cred);
+	status = ops->renew_lease(clp, cred);
 	put_rpccred(cred);
 out:
 	nfs4_recovery_handle_error(clp, status);
@@ -1096,21 +1114,98 @@ out:
 static int nfs4_reclaim_lease(struct nfs_client *clp)
 {
 	struct rpc_cred *cred;
+	struct nfs4_state_recovery_ops *ops =
+		nfs4_reboot_recovery_ops[clp->cl_minorversion];
 	int status = -ENOENT;
 
-	cred = nfs4_get_setclientid_cred(clp);
+	cred = ops->get_clid_cred(clp);
 	if (cred != NULL) {
-		status = nfs4_init_client(clp, cred);
+		status = ops->establish_clid(clp, cred);
 		put_rpccred(cred);
 		/* Handle case where the user hasn't set up machine creds */
 		if (status == -EACCES && cred == clp->cl_machine_cred) {
 			nfs4_clear_machine_cred(clp);
 			status = -EAGAIN;
 		}
+		if (status == -NFS4ERR_MINOR_VERS_MISMATCH)
+			status = -EPROTONOSUPPORT;
 	}
 	return status;
 }
 
+#ifdef CONFIG_NFS_V4_1
+static void nfs4_session_recovery_handle_error(struct nfs_client *clp, int err)
+{
+	switch (err) {
+	case -NFS4ERR_STALE_CLIENTID:
+		set_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state);
+		set_bit(NFS4CLNT_SESSION_SETUP, &clp->cl_state);
+	}
+}
+
+static int nfs4_reset_session(struct nfs_client *clp)
+{
+	int status;
+
+	status = nfs4_proc_destroy_session(clp->cl_session);
+	if (status && status != -NFS4ERR_BADSESSION &&
+	    status != -NFS4ERR_DEADSESSION) {
+		nfs4_session_recovery_handle_error(clp, status);
+		goto out;
+	}
+
+	memset(clp->cl_session->sess_id.data, 0, NFS4_MAX_SESSIONID_LEN);
+	status = nfs4_proc_create_session(clp, 1);
+	if (status)
+		nfs4_session_recovery_handle_error(clp, status);
+		/* fall through*/
+out:
+	/* Wake up the next rpc task even on error */
+	rpc_wake_up_next(&clp->cl_session->fc_slot_table.slot_tbl_waitq);
+	return status;
+}
+
+static int nfs4_initialize_session(struct nfs_client *clp)
+{
+	int status;
+
+	status = nfs4_proc_create_session(clp, 0);
+	if (!status) {
+		nfs_mark_client_ready(clp, NFS_CS_READY);
+	} else if (status == -NFS4ERR_STALE_CLIENTID) {
+		set_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state);
+		set_bit(NFS4CLNT_SESSION_SETUP, &clp->cl_state);
+	} else {
+		nfs_mark_client_ready(clp, status);
+	}
+	return status;
+}
+#else /* CONFIG_NFS_V4_1 */
+static int nfs4_reset_session(struct nfs_client *clp) { return 0; }
+static int nfs4_initialize_session(struct nfs_client *clp) { return 0; }
+#endif /* CONFIG_NFS_V4_1 */
+
+/* Set NFS4CLNT_LEASE_EXPIRED for all v4.0 errors and for recoverable errors
+ * on EXCHANGE_ID for v4.1
+ */
+static void nfs4_set_lease_expired(struct nfs_client *clp, int status)
+{
+	if (nfs4_has_session(clp)) {
+		switch (status) {
+		case -NFS4ERR_DELAY:
+		case -NFS4ERR_CLID_INUSE:
+		case -EAGAIN:
+			break;
+
+		case -NFS4ERR_NOT_SAME: /* FixMe: implement recovery
+					 * in nfs4_exchange_id */
+		default:
+			return;
+		}
+	}
+	set_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state);
+}
+
 static void nfs4_state_manager(struct nfs_client *clp)
 {
 	int status = 0;
@@ -1121,9 +1216,12 @@ static void nfs4_state_manager(struct nfs_client *clp)
 			/* We're going to have to re-establish a clientid */
 			status = nfs4_reclaim_lease(clp);
 			if (status) {
-				set_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state);
+				nfs4_set_lease_expired(clp, status);
 				if (status == -EAGAIN)
 					continue;
+				if (clp->cl_cons_state ==
+							NFS_CS_SESSION_INITING)
+					nfs_mark_client_ready(clp, status);
 				goto out_error;
 			}
 			clear_bit(NFS4CLNT_CHECK_LEASE, &clp->cl_state);
@@ -1134,25 +1232,44 @@ static void nfs4_state_manager(struct nfs_client *clp)
 			if (status != 0)
 				continue;
 		}
-
+		/* Initialize or reset the session */
+		if (nfs4_has_session(clp) &&
+		   test_and_clear_bit(NFS4CLNT_SESSION_SETUP, &clp->cl_state)) {
+			if (clp->cl_cons_state == NFS_CS_SESSION_INITING)
+				status = nfs4_initialize_session(clp);
+			else
+				status = nfs4_reset_session(clp);
+			if (status) {
+				if (status == -NFS4ERR_STALE_CLIENTID)
+					continue;
+				goto out_error;
+			}
+		}
 		/* First recover reboot state... */
 		if (test_and_clear_bit(NFS4CLNT_RECLAIM_REBOOT, &clp->cl_state)) {
-			status = nfs4_do_reclaim(clp, &nfs4_reboot_recovery_ops);
+			status = nfs4_do_reclaim(clp,
+				nfs4_reboot_recovery_ops[clp->cl_minorversion]);
 			if (status == -NFS4ERR_STALE_CLIENTID)
 				continue;
+			if (test_bit(NFS4CLNT_SESSION_SETUP, &clp->cl_state))
+				continue;
 			nfs4_state_end_reclaim_reboot(clp);
 			continue;
 		}
 
 		/* Now recover expired state... */
 		if (test_and_clear_bit(NFS4CLNT_RECLAIM_NOGRACE, &clp->cl_state)) {
-			status = nfs4_do_reclaim(clp, &nfs4_nograce_recovery_ops);
+			status = nfs4_do_reclaim(clp,
+				nfs4_nograce_recovery_ops[clp->cl_minorversion]);
 			if (status < 0) {
 				set_bit(NFS4CLNT_RECLAIM_NOGRACE, &clp->cl_state);
 				if (status == -NFS4ERR_STALE_CLIENTID)
 					continue;
 				if (status == -NFS4ERR_EXPIRED)
 					continue;
+				if (test_bit(NFS4CLNT_SESSION_SETUP,
+								&clp->cl_state))
+					continue;
 				goto out_error;
 			} else
 				nfs4_state_end_reclaim_nograce(clp);

fs/nfs/read.c

@@ -22,6 +22,7 @@
 
 #include <asm/system.h>
 
+#include "nfs4_fs.h"
 #include "internal.h"
 #include "iostat.h"
 #include "fscache.h"
@@ -46,6 +47,7 @@ struct nfs_read_data *nfs_readdata_alloc(unsigned int pagecount)
 		memset(p, 0, sizeof(*p));
 		INIT_LIST_HEAD(&p->pages);
 		p->npages = pagecount;
+		p->res.seq_res.sr_slotid = NFS4_MAX_SLOT_TABLE;
 		if (pagecount <= ARRAY_SIZE(p->page_array))
 			p->pagevec = p->page_array;
 		else {
@@ -357,19 +359,25 @@ static void nfs_readpage_retry(struct rpc_task *task, struct nfs_read_data *data
 	struct nfs_readres *resp = &data->res;
 
 	if (resp->eof || resp->count == argp->count)
-		return;
+		goto out;
 
 	/* This is a short read! */
 	nfs_inc_stats(data->inode, NFSIOS_SHORTREAD);
 	/* Has the server at least made some progress? */
 	if (resp->count == 0)
-		return;
+		goto out;
 
 	/* Yes, so retry the read at the end of the data */
 	argp->offset += resp->count;
 	argp->pgbase += resp->count;
 	argp->count -= resp->count;
-	rpc_restart_call(task);
+	nfs4_restart_rpc(task, NFS_SERVER(data->inode)->nfs_client);
+	return;
+out:
+	nfs4_sequence_free_slot(NFS_SERVER(data->inode)->nfs_client,
+				&data->res.seq_res);
+	return;
+
 }
 
 /*
@@ -406,7 +414,23 @@ static void nfs_readpage_release_partial(void *calldata)
 	nfs_readdata_release(calldata);
 }
 
+#if defined(CONFIG_NFS_V4_1)
+void nfs_read_prepare(struct rpc_task *task, void *calldata)
+{
+	struct nfs_read_data *data = calldata;
+
+	if (nfs4_setup_sequence(NFS_SERVER(data->inode)->nfs_client,
+				&data->args.seq_args, &data->res.seq_res,
+				0, task))
+		return;
+	rpc_call_start(task);
+}
+#endif /* CONFIG_NFS_V4_1 */
+
 static const struct rpc_call_ops nfs_read_partial_ops = {
+#if defined(CONFIG_NFS_V4_1)
+	.rpc_call_prepare = nfs_read_prepare,
+#endif /* CONFIG_NFS_V4_1 */
 	.rpc_call_done = nfs_readpage_result_partial,
 	.rpc_release = nfs_readpage_release_partial,
 };
@@ -470,6 +494,9 @@ static void nfs_readpage_release_full(void *calldata)
 }
 
 static const struct rpc_call_ops nfs_read_full_ops = {
+#if defined(CONFIG_NFS_V4_1)
+	.rpc_call_prepare = nfs_read_prepare,
+#endif /* CONFIG_NFS_V4_1 */
 	.rpc_call_done = nfs_readpage_result_full,
 	.rpc_release = nfs_readpage_release_full,
 };

fs/nfs/super.c

@@ -90,6 +90,7 @@ enum {
 	Opt_mountport,
 	Opt_mountvers,
 	Opt_nfsvers,
+	Opt_minorversion,
 
 	/* Mount options that take string arguments */
 	Opt_sec, Opt_proto, Opt_mountproto, Opt_mounthost,
@@ -155,6 +156,7 @@ static const match_table_t nfs_mount_option_tokens = {
 	{ Opt_mountvers, "mountvers=%u" },
 	{ Opt_nfsvers, "nfsvers=%u" },
 	{ Opt_nfsvers, "vers=%u" },
+	{ Opt_minorversion, "minorversion=%u" },
 
 	{ Opt_sec, "sec=%s" },
 	{ Opt_proto, "proto=%s" },
@@ -1211,6 +1213,13 @@ static int nfs_parse_mount_options(char *raw,
 				nfs_parse_invalid_value("nfsvers");
 			}
 			break;
+		case Opt_minorversion:
+			if (match_int(args, &option))
+				return 0;
+			if (option < 0 || option > NFS4_MAX_MINOR_VERSION)
+				return 0;
+			mnt->minorversion = option;
+			break;
 
 		/*
 		 * options that take text values
@@ -2263,6 +2272,7 @@ static int nfs4_validate_mount_data(void *options,
 	args->nfs_server.port	= NFS_PORT; /* 2049 unless user set port= */
 	args->auth_flavors[0]	= RPC_AUTH_UNIX;
 	args->auth_flavor_len	= 0;
+	args->minorversion	= 0;
 
 	switch (data->version) {
 	case 1:
@@ -2477,12 +2487,13 @@ static void nfs4_kill_super(struct super_block *sb)
 {
 	struct nfs_server *server = NFS_SB(sb);
 
+	dprintk("--> %s\n", __func__);
 	nfs_super_return_all_delegations(sb);
 	kill_anon_super(sb);
-
 	nfs4_renewd_prepare_shutdown(server);
 	nfs_fscache_release_super_cookie(sb);
 	nfs_free_server(server);
+	dprintk("<-- %s\n", __func__);
 }
 
 /*

fs/nfs/unlink.c

@@ -15,6 +15,7 @@
 #include <linux/wait.h>
 
 #include "internal.h"
+#include "nfs4_fs.h"
 
 struct nfs_unlinkdata {
 	struct hlist_node list;
@@ -82,7 +83,7 @@ static void nfs_async_unlink_done(struct rpc_task *task, void *calldata)
 	struct inode *dir = data->dir;
 
 	if (!NFS_PROTO(dir)->unlink_done(task, dir))
-		rpc_restart_call(task);
+		nfs4_restart_rpc(task, NFS_SERVER(dir)->nfs_client);
 }
 
 /**
@@ -102,9 +103,25 @@ static void nfs_async_unlink_release(void *calldata)
 	nfs_sb_deactive(sb);
 }
 
+#if defined(CONFIG_NFS_V4_1)
+void nfs_unlink_prepare(struct rpc_task *task, void *calldata)
+{
+	struct nfs_unlinkdata *data = calldata;
+	struct nfs_server *server = NFS_SERVER(data->dir);
+
+	if (nfs4_setup_sequence(server->nfs_client, &data->args.seq_args,
+				&data->res.seq_res, 1, task))
+		return;
+	rpc_call_start(task);
+}
+#endif /* CONFIG_NFS_V4_1 */
+
 static const struct rpc_call_ops nfs_unlink_ops = {
 	.rpc_call_done = nfs_async_unlink_done,
 	.rpc_release = nfs_async_unlink_release,
+#if defined(CONFIG_NFS_V4_1)
+	.rpc_call_prepare = nfs_unlink_prepare,
+#endif /* CONFIG_NFS_V4_1 */
 };
 
 static int nfs_do_call_unlink(struct dentry *parent, struct inode *dir, struct nfs_unlinkdata *data)
@@ -241,6 +258,7 @@ nfs_async_unlink(struct inode *dir, struct dentry *dentry)
 		status = PTR_ERR(data->cred);
 		goto out_free;
 	}
+	data->res.seq_res.sr_slotid = NFS4_MAX_SLOT_TABLE;
 
 	status = -EBUSY;
 	spin_lock(&dentry->d_lock);

fs/nfs/write.c

@@ -25,6 +25,7 @@
 #include "delegation.h"
 #include "internal.h"
 #include "iostat.h"
+#include "nfs4_fs.h"
 
 #define NFSDBG_FACILITY		NFSDBG_PAGECACHE
 
@@ -52,6 +53,7 @@ struct nfs_write_data *nfs_commitdata_alloc(void)
 	if (p) {
 		memset(p, 0, sizeof(*p));
 		INIT_LIST_HEAD(&p->pages);
+		p->res.seq_res.sr_slotid = NFS4_MAX_SLOT_TABLE;
 	}
 	return p;
 }
@@ -71,6 +73,7 @@ struct nfs_write_data *nfs_writedata_alloc(unsigned int pagecount)
 		memset(p, 0, sizeof(*p));
 		INIT_LIST_HEAD(&p->pages);
 		p->npages = pagecount;
+		p->res.seq_res.sr_slotid = NFS4_MAX_SLOT_TABLE;
 		if (pagecount <= ARRAY_SIZE(p->page_array))
 			p->pagevec = p->page_array;
 		else {
@@ -1048,7 +1051,23 @@ out:
 	nfs_writedata_release(calldata);
 }
 
+#if defined(CONFIG_NFS_V4_1)
+void nfs_write_prepare(struct rpc_task *task, void *calldata)
+{
+	struct nfs_write_data *data = calldata;
+	struct nfs_client *clp = (NFS_SERVER(data->inode))->nfs_client;
+
+	if (nfs4_setup_sequence(clp, &data->args.seq_args,
+				&data->res.seq_res, 1, task))
+		return;
+	rpc_call_start(task);
+}
+#endif /* CONFIG_NFS_V4_1 */
+
 static const struct rpc_call_ops nfs_write_partial_ops = {
+#if defined(CONFIG_NFS_V4_1)
+	.rpc_call_prepare = nfs_write_prepare,
+#endif /* CONFIG_NFS_V4_1 */
 	.rpc_call_done = nfs_writeback_done_partial,
 	.rpc_release = nfs_writeback_release_partial,
 };
@@ -1111,6 +1130,9 @@ remove_request:
 }
 
 static const struct rpc_call_ops nfs_write_full_ops = {
+#if defined(CONFIG_NFS_V4_1)
+	.rpc_call_prepare = nfs_write_prepare,
+#endif /* CONFIG_NFS_V4_1 */
 	.rpc_call_done = nfs_writeback_done_full,
 	.rpc_release = nfs_writeback_release_full,
 };
@@ -1123,6 +1145,7 @@ int nfs_writeback_done(struct rpc_task *task, struct nfs_write_data *data)
 {
 	struct nfs_writeargs	*argp = &data->args;
 	struct nfs_writeres	*resp = &data->res;
+	struct nfs_server	*server = NFS_SERVER(data->inode);
 	int status;
 
 	dprintk("NFS: %5u nfs_writeback_done (status %d)\n",
@@ -1155,7 +1178,7 @@ int nfs_writeback_done(struct rpc_task *task, struct nfs_write_data *data)
 		if (time_before(complain, jiffies)) {
 			dprintk("NFS:       faulty NFS server %s:"
 				" (committed = %d) != (stable = %d)\n",
-				NFS_SERVER(data->inode)->nfs_client->cl_hostname,
+				server->nfs_client->cl_hostname,
 				resp->verf->committed, argp->stable);
 			complain = jiffies + 300 * HZ;
 		}
@@ -1181,7 +1204,7 @@ int nfs_writeback_done(struct rpc_task *task, struct nfs_write_data *data)
 				 */
 				argp->stable = NFS_FILE_SYNC;
 			}
-			rpc_restart_call(task);
+			nfs4_restart_rpc(task, server->nfs_client);
 			return -EAGAIN;
 		}
 		if (time_before(complain, jiffies)) {
@@ -1193,6 +1216,7 @@ int nfs_writeback_done(struct rpc_task *task, struct nfs_write_data *data)
 		/* Can't do anything about it except throw an error. */
 		task->tk_status = -EIO;
 	}
+	nfs4_sequence_free_slot(server->nfs_client, &data->res.seq_res);
 	return 0;
 }
 
@@ -1349,6 +1373,9 @@ static void nfs_commit_release(void *calldata)
 }
 
 static const struct rpc_call_ops nfs_commit_ops = {
+#if defined(CONFIG_NFS_V4_1)
+	.rpc_call_prepare = nfs_write_prepare,
+#endif /* CONFIG_NFS_V4_1 */
 	.rpc_call_done = nfs_commit_done,
 	.rpc_release = nfs_commit_release,
 };

include/linux/nfs4.h

@@ -21,6 +21,7 @@
 #define NFS4_FHSIZE		128
 #define NFS4_MAXPATHLEN		PATH_MAX
 #define NFS4_MAXNAMLEN		NAME_MAX
+#define NFS4_OPAQUE_LIMIT	1024
 #define NFS4_MAX_SESSIONID_LEN	16
 
 #define NFS4_ACCESS_READ        0x0001
@@ -130,6 +131,16 @@
 
 #define NFS4_MAX_UINT64	(~(u64)0)
 
+/* An NFS4 sessions server must support at least NFS4_MAX_OPS operations.
+ * If a compound requires more operations, adjust NFS4_MAX_OPS accordingly.
+ */
+#define NFS4_MAX_OPS   8
+
+/* Our NFS4 client back channel server only wants the cb_sequene and the
+ * actual operation per compound
+ */
+#define NFS4_MAX_BACK_CHANNEL_OPS 2
+
 enum nfs4_acl_whotype {
 	NFS4_ACL_WHO_NAMED = 0,
 	NFS4_ACL_WHO_OWNER,
@@ -462,6 +473,13 @@ enum lock_type4 {
 #define NFSPROC4_NULL 0
 #define NFSPROC4_COMPOUND 1
 #define NFS4_MINOR_VERSION 0
+
+#if defined(CONFIG_NFS_V4_1)
+#define NFS4_MAX_MINOR_VERSION 1
+#else
+#define NFS4_MAX_MINOR_VERSION 0
+#endif /* CONFIG_NFS_V4_1 */
+
 #define NFS4_DEBUG 1
 
 /* Index of predefined Linux client operations */

include/linux/nfs_fs_sb.h

@@ -4,11 +4,17 @@
 #include <linux/list.h>
 #include <linux/backing-dev.h>
 #include <linux/wait.h>
+#include <linux/nfs_xdr.h>
+#include <linux/sunrpc/xprt.h>
 
 #include <asm/atomic.h>
 
+struct nfs4_session;
 struct nfs_iostats;
 struct nlm_host;
+struct nfs4_sequence_args;
+struct nfs4_sequence_res;
+struct nfs_server;
 
 /*
  * The nfs_client identifies our client state to the server.
@@ -18,6 +24,7 @@ struct nfs_client {
 	int			cl_cons_state;	/* current construction state (-ve: init error) */
 #define NFS_CS_READY		0		/* ready to be used */
 #define NFS_CS_INITING		1		/* busy initialising */
+#define NFS_CS_SESSION_INITING	2		/* busy initialising  session */
 	unsigned long		cl_res_state;	/* NFS resources state */
 #define NFS_CS_CALLBACK		1		/* - callback started */
 #define NFS_CS_IDMAP		2		/* - idmap started */
@@ -32,6 +39,7 @@ struct nfs_client {
 	const struct nfs_rpc_ops *rpc_ops;	/* NFS protocol vector */
 	int			cl_proto;	/* Network transport protocol */
 
+	u32			cl_minorversion;/* NFSv4 minorversion */
 	struct rpc_cred		*cl_machine_cred;
 
 #ifdef CONFIG_NFS_V4
@@ -63,7 +71,22 @@ struct nfs_client {
 	 */
 	char			cl_ipaddr[48];
 	unsigned char		cl_id_uniquifier;
-#endif
+	int		     (* cl_call_sync)(struct nfs_server *server,
+					      struct rpc_message *msg,
+					      struct nfs4_sequence_args *args,
+					      struct nfs4_sequence_res *res,
+					      int cache_reply);
+#endif /* CONFIG_NFS_V4 */
+
+#ifdef CONFIG_NFS_V4_1
+	/* clientid returned from EXCHANGE_ID, used by session operations */
+	u64			cl_ex_clid;
+	/* The sequence id to use for the next CREATE_SESSION */
+	u32			cl_seqid;
+	/* The flags used for obtaining the clientid during EXCHANGE_ID */
+	u32			cl_exchange_flags;
+	struct nfs4_session	*cl_session; 	/* sharred session */
+#endif /* CONFIG_NFS_V4_1 */
 
 #ifdef CONFIG_NFS_FSCACHE
 	struct fscache_cookie	*fscache;	/* client index cache cookie */
@@ -145,4 +168,46 @@ struct nfs_server {
 #define NFS_CAP_ACLS		(1U << 3)
 #define NFS_CAP_ATOMIC_OPEN	(1U << 4)
 
+
+/* maximum number of slots to use */
+#define NFS4_MAX_SLOT_TABLE RPC_MAX_SLOT_TABLE
+
+#if defined(CONFIG_NFS_V4_1)
+
+/* Sessions */
+#define SLOT_TABLE_SZ (NFS4_MAX_SLOT_TABLE/(8*sizeof(long)))
+struct nfs4_slot_table {
+	struct nfs4_slot *slots;		/* seqid per slot */
+	unsigned long   used_slots[SLOT_TABLE_SZ]; /* used/unused bitmap */
+	spinlock_t	slot_tbl_lock;
+	struct rpc_wait_queue	slot_tbl_waitq;	/* allocators may wait here */
+	int		max_slots;		/* # slots in table */
+	int		highest_used_slotid;	/* sent to server on each SEQ.
+						 * op for dynamic resizing */
+};
+
+static inline int slot_idx(struct nfs4_slot_table *tbl, struct nfs4_slot *sp)
+{
+	return sp - tbl->slots;
+}
+
+/*
+ * Session related parameters
+ */
+struct nfs4_session {
+	struct nfs4_sessionid		sess_id;
+	u32				flags;
+	unsigned long			session_state;
+	u32				hash_alg;
+	u32				ssv_len;
+
+	/* The fore and back channel */
+	struct nfs4_channel_attrs	fc_attrs;
+	struct nfs4_slot_table		fc_slot_table;
+	struct nfs4_channel_attrs	bc_attrs;
+	struct nfs4_slot_table		bc_slot_table;
+	struct nfs_client		*clp;
+};
+
+#endif /* CONFIG_NFS_V4_1 */
 #endif

include/linux/nfs_xdr.h

@@ -145,6 +145,44 @@ struct nfs4_change_info {
 };
 
 struct nfs_seqid;
+
+/* nfs41 sessions channel attributes */
+struct nfs4_channel_attrs {
+	u32			headerpadsz;
+	u32			max_rqst_sz;
+	u32			max_resp_sz;
+	u32			max_resp_sz_cached;
+	u32			max_ops;
+	u32			max_reqs;
+};
+
+/* nfs41 sessions slot seqid */
+struct nfs4_slot {
+	u32		 	seq_nr;
+};
+
+struct nfs4_sequence_args {
+	struct nfs4_session	*sa_session;
+	u8			sa_slotid;
+	u8			sa_cache_this;
+};
+
+struct nfs4_sequence_res {
+	struct nfs4_session	*sr_session;
+	u8			sr_slotid;	/* slot used to send request */
+	unsigned long		sr_renewal_time;
+	int			sr_status;	/* sequence operation status */
+};
+
+struct nfs4_get_lease_time_args {
+	struct nfs4_sequence_args	la_seq_args;
+};
+
+struct nfs4_get_lease_time_res {
+	struct nfs_fsinfo	       *lr_fsinfo;
+	struct nfs4_sequence_res	lr_seq_res;
+};
+
 /*
  * Arguments to the open call.
  */
@@ -165,6 +203,7 @@ struct nfs_openargs {
 	const struct nfs_server *server;	 /* Needed for ID mapping */
 	const u32 *		bitmask;
 	__u32			claim;
+	struct nfs4_sequence_args	seq_args;
 };
 
 struct nfs_openres {
@@ -181,6 +220,7 @@ struct nfs_openres {
 	__u32			do_recall;
 	__u64			maxsize;
 	__u32			attrset[NFS4_BITMAP_SIZE];
+	struct nfs4_sequence_res	seq_res;
 };
 
 /*
@@ -206,6 +246,7 @@ struct nfs_closeargs {
 	struct nfs_seqid *	seqid;
 	fmode_t			fmode;
 	const u32 *		bitmask;
+	struct nfs4_sequence_args	seq_args;
 };
 
 struct nfs_closeres {
@@ -213,6 +254,7 @@ struct nfs_closeres {
 	struct nfs_fattr *	fattr;
 	struct nfs_seqid *	seqid;
 	const struct nfs_server *server;
+	struct nfs4_sequence_res	seq_res;
 };
 /*
  *  * Arguments to the lock,lockt, and locku call.
@@ -233,12 +275,14 @@ struct nfs_lock_args {
 	unsigned char		block : 1;
 	unsigned char		reclaim : 1;
 	unsigned char		new_lock_owner : 1;
+	struct nfs4_sequence_args	seq_args;
 };
 
 struct nfs_lock_res {
 	nfs4_stateid		stateid;
 	struct nfs_seqid *	lock_seqid;
 	struct nfs_seqid *	open_seqid;
+	struct nfs4_sequence_res	seq_res;
 };
 
 struct nfs_locku_args {
@@ -246,32 +290,38 @@ struct nfs_locku_args {
 	struct file_lock *	fl;
 	struct nfs_seqid *	seqid;
 	nfs4_stateid *		stateid;
+	struct nfs4_sequence_args	seq_args;
 };
 
 struct nfs_locku_res {
 	nfs4_stateid		stateid;
 	struct nfs_seqid *	seqid;
+	struct nfs4_sequence_res	seq_res;
 };
 
 struct nfs_lockt_args {
 	struct nfs_fh *		fh;
 	struct file_lock *	fl;
 	struct nfs_lowner	lock_owner;
+	struct nfs4_sequence_args	seq_args;
 };
 
 struct nfs_lockt_res {
 	struct file_lock *	denied; /* LOCK, LOCKT failed */
+	struct nfs4_sequence_res	seq_res;
 };
 
 struct nfs4_delegreturnargs {
 	const struct nfs_fh *fhandle;
 	const nfs4_stateid *stateid;
 	const u32 * bitmask;
+	struct nfs4_sequence_args	seq_args;
 };
 
 struct nfs4_delegreturnres {
 	struct nfs_fattr * fattr;
 	const struct nfs_server *server;
+	struct nfs4_sequence_res	seq_res;
 };
 
 /*
@@ -284,12 +334,14 @@ struct nfs_readargs {
 	__u32			count;
 	unsigned int		pgbase;
 	struct page **		pages;
+	struct nfs4_sequence_args	seq_args;
 };
 
 struct nfs_readres {
 	struct nfs_fattr *	fattr;
 	__u32			count;
 	int                     eof;
+	struct nfs4_sequence_res	seq_res;
 };
 
 /*
@@ -304,6 +356,7 @@ struct nfs_writeargs {
 	unsigned int		pgbase;
 	struct page **		pages;
 	const u32 *		bitmask;
+	struct nfs4_sequence_args	seq_args;
 };
 
 struct nfs_writeverf {
@@ -316,6 +369,7 @@ struct nfs_writeres {
 	struct nfs_writeverf *	verf;
 	__u32			count;
 	const struct nfs_server *server;
+	struct nfs4_sequence_res	seq_res;
 };
 
 /*
@@ -325,12 +379,14 @@ struct nfs_removeargs {
 	const struct nfs_fh	*fh;
 	struct qstr		name;
 	const u32 *		bitmask;
+	struct nfs4_sequence_args	seq_args;
 };
 
 struct nfs_removeres {
 	const struct nfs_server *server;
 	struct nfs4_change_info	cinfo;
 	struct nfs_fattr	dir_attr;
+	struct nfs4_sequence_res 	seq_res;
 };
 
 /*
@@ -383,6 +439,7 @@ struct nfs_setattrargs {
 	struct iattr *                  iap;
 	const struct nfs_server *	server; /* Needed for name mapping */
 	const u32 *			bitmask;
+	struct nfs4_sequence_args 	seq_args;
 };
 
 struct nfs_setaclargs {
@@ -390,6 +447,11 @@ struct nfs_setaclargs {
 	size_t				acl_len;
 	unsigned int			acl_pgbase;
 	struct page **			acl_pages;
+	struct nfs4_sequence_args	seq_args;
+};
+
+struct nfs_setaclres {
+	struct nfs4_sequence_res	seq_res;
 };
 
 struct nfs_getaclargs {
@@ -397,11 +459,18 @@ struct nfs_getaclargs {
 	size_t				acl_len;
 	unsigned int			acl_pgbase;
 	struct page **			acl_pages;
+	struct nfs4_sequence_args 	seq_args;
+};
+
+struct nfs_getaclres {
+	size_t				acl_len;
+	struct nfs4_sequence_res	seq_res;
 };
 
 struct nfs_setattrres {
 	struct nfs_fattr *              fattr;
 	const struct nfs_server *	server;
+	struct nfs4_sequence_res	seq_res;
 };
 
 struct nfs_linkargs {
@@ -583,6 +652,7 @@ struct nfs4_accessargs {
 	const struct nfs_fh *		fh;
 	const u32 *			bitmask;
 	u32				access;
+	struct nfs4_sequence_args	seq_args;
 };
 
 struct nfs4_accessres {
@@ -590,6 +660,7 @@ struct nfs4_accessres {
 	struct nfs_fattr *		fattr;
 	u32				supported;
 	u32				access;
+	struct nfs4_sequence_res	seq_res;
 };
 
 struct nfs4_create_arg {
@@ -609,6 +680,7 @@ struct nfs4_create_arg {
 	const struct iattr *		attrs;
 	const struct nfs_fh *		dir_fh;
 	const u32 *			bitmask;
+	struct nfs4_sequence_args 	seq_args;
 };
 
 struct nfs4_create_res {
@@ -617,21 +689,30 @@ struct nfs4_create_res {
 	struct nfs_fattr *		fattr;
 	struct nfs4_change_info		dir_cinfo;
 	struct nfs_fattr *		dir_fattr;
+	struct nfs4_sequence_res	seq_res;
 };
 
 struct nfs4_fsinfo_arg {
 	const struct nfs_fh *		fh;
 	const u32 *			bitmask;
+	struct nfs4_sequence_args	seq_args;
+};
+
+struct nfs4_fsinfo_res {
+	struct nfs_fsinfo	       *fsinfo;
+	struct nfs4_sequence_res	seq_res;
 };
 
 struct nfs4_getattr_arg {
 	const struct nfs_fh *		fh;
 	const u32 *			bitmask;
+	struct nfs4_sequence_args	seq_args;
 };
 
 struct nfs4_getattr_res {
 	const struct nfs_server *	server;
 	struct nfs_fattr *		fattr;
+	struct nfs4_sequence_res	seq_res;
 };
 
 struct nfs4_link_arg {
@@ -639,6 +720,7 @@ struct nfs4_link_arg {
 	const struct nfs_fh *		dir_fh;
 	const struct qstr *		name;
 	const u32 *			bitmask;
+	struct nfs4_sequence_args 	seq_args;
 };
 
 struct nfs4_link_res {
@@ -646,6 +728,7 @@ struct nfs4_link_res {
 	struct nfs_fattr *		fattr;
 	struct nfs4_change_info		cinfo;
 	struct nfs_fattr *		dir_attr;
+	struct nfs4_sequence_res	seq_res;
 };
 
 
@@ -653,21 +736,30 @@ struct nfs4_lookup_arg {
 	const struct nfs_fh *		dir_fh;
 	const struct qstr *		name;
 	const u32 *			bitmask;
+	struct nfs4_sequence_args	seq_args;
 };
 
 struct nfs4_lookup_res {
 	const struct nfs_server *	server;
 	struct nfs_fattr *		fattr;
 	struct nfs_fh *			fh;
+	struct nfs4_sequence_res	seq_res;
 };
 
 struct nfs4_lookup_root_arg {
 	const u32 *			bitmask;
+	struct nfs4_sequence_args	seq_args;
 };
 
 struct nfs4_pathconf_arg {
 	const struct nfs_fh *		fh;
 	const u32 *			bitmask;
+	struct nfs4_sequence_args	seq_args;
+};
+
+struct nfs4_pathconf_res {
+	struct nfs_pathconf	       *pathconf;
+	struct nfs4_sequence_res	seq_res;
 };
 
 struct nfs4_readdir_arg {
@@ -678,11 +770,13 @@ struct nfs4_readdir_arg {
 	struct page **			pages;	/* zero-copy data */
 	unsigned int			pgbase;	/* zero-copy data */
 	const u32 *			bitmask;
+	struct nfs4_sequence_args	seq_args;
 };
 
 struct nfs4_readdir_res {
 	nfs4_verifier			verifier;
 	unsigned int			pgbase;
+	struct nfs4_sequence_res	seq_res;
 };
 
 struct nfs4_readlink {
@@ -690,6 +784,11 @@ struct nfs4_readlink {
 	unsigned int			pgbase;
 	unsigned int			pglen;   /* zero-copy data */
 	struct page **			pages;   /* zero-copy data */
+	struct nfs4_sequence_args	seq_args;
+};
+
+struct nfs4_readlink_res {
+	struct nfs4_sequence_res	seq_res;
 };
 
 struct nfs4_rename_arg {
@@ -698,6 +797,7 @@ struct nfs4_rename_arg {
 	const struct qstr *		old_name;
 	const struct qstr *		new_name;
 	const u32 *			bitmask;
+	struct nfs4_sequence_args	seq_args;
 };
 
 struct nfs4_rename_res {
@@ -706,6 +806,7 @@ struct nfs4_rename_res {
 	struct nfs_fattr *		old_fattr;
 	struct nfs4_change_info		new_cinfo;
 	struct nfs_fattr *		new_fattr;
+	struct nfs4_sequence_res	seq_res;
 };
 
 #define NFS4_SETCLIENTID_NAMELEN	(127)
@@ -724,6 +825,17 @@ struct nfs4_setclientid {
 struct nfs4_statfs_arg {
 	const struct nfs_fh *		fh;
 	const u32 *			bitmask;
+	struct nfs4_sequence_args	seq_args;
+};
+
+struct nfs4_statfs_res {
+	struct nfs_fsstat	       *fsstat;
+	struct nfs4_sequence_res	seq_res;
+};
+
+struct nfs4_server_caps_arg {
+	struct nfs_fh		       *fhandle;
+	struct nfs4_sequence_args	seq_args;
 };
 
 struct nfs4_server_caps_res {
@@ -731,6 +843,7 @@ struct nfs4_server_caps_res {
 	u32				acl_bitmask;
 	u32				has_links;
 	u32				has_symlinks;
+	struct nfs4_sequence_res	seq_res;
 };
 
 struct nfs4_string {
@@ -765,10 +878,68 @@ struct nfs4_fs_locations_arg {
 	const struct qstr *name;
 	struct page *page;
 	const u32 *bitmask;
+	struct nfs4_sequence_args	seq_args;
+};
+
+struct nfs4_fs_locations_res {
+	struct nfs4_fs_locations       *fs_locations;
+	struct nfs4_sequence_res	seq_res;
 };
 
 #endif /* CONFIG_NFS_V4 */
 
+struct nfstime4 {
+	u64	seconds;
+	u32	nseconds;
+};
+
+#ifdef CONFIG_NFS_V4_1
+struct nfs_impl_id4 {
+	u32		domain_len;
+	char		*domain;
+	u32		name_len;
+	char		*name;
+	struct nfstime4	date;
+};
+
+#define NFS4_EXCHANGE_ID_LEN	(48)
+struct nfs41_exchange_id_args {
+	struct nfs_client		*client;
+	nfs4_verifier			*verifier;
+	unsigned int 			id_len;
+	char 				id[NFS4_EXCHANGE_ID_LEN];
+	u32				flags;
+};
+
+struct server_owner {
+	uint64_t			minor_id;
+	uint32_t			major_id_sz;
+	char				major_id[NFS4_OPAQUE_LIMIT];
+};
+
+struct server_scope {
+	uint32_t			server_scope_sz;
+	char 				server_scope[NFS4_OPAQUE_LIMIT];
+};
+
+struct nfs41_exchange_id_res {
+	struct nfs_client		*client;
+	u32				flags;
+};
+
+struct nfs41_create_session_args {
+	struct nfs_client	       *client;
+	uint32_t			flags;
+	uint32_t			cb_program;
+	struct nfs4_channel_attrs	fc_attrs;	/* Fore Channel */
+	struct nfs4_channel_attrs	bc_attrs;	/* Back Channel */
+};
+
+struct nfs41_create_session_res {
+	struct nfs_client	       *client;
+};
+#endif /* CONFIG_NFS_V4_1 */
+
 struct nfs_page;
 
 #define NFS_PAGEVEC_SIZE	(8U)

include/linux/nfsd/state.h

@@ -41,7 +41,6 @@
 #include <linux/kref.h>
 #include <linux/sunrpc/clnt.h>
 
-#define NFS4_OPAQUE_LIMIT 1024
 typedef struct {
 	u32             cl_boot;
 	u32             cl_id;

include/linux/sunrpc/bc_xprt.h

@@ -0,0 +1,49 @@
+/******************************************************************************
+
+(c) 2008 NetApp.  All Rights Reserved.
+
+NetApp provides this source code under the GPL v2 License.
+The GPL v2 license is available at
+http://opensource.org/licenses/gpl-license.php.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
+CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
+EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+******************************************************************************/
+
+/*
+ * Functions to create and manage the backchannel
+ */
+
+#ifndef _LINUX_SUNRPC_BC_XPRT_H
+#define _LINUX_SUNRPC_BC_XPRT_H
+
+#include <linux/sunrpc/svcsock.h>
+#include <linux/sunrpc/xprt.h>
+#include <linux/sunrpc/sched.h>
+
+#ifdef CONFIG_NFS_V4_1
+struct rpc_rqst *xprt_alloc_bc_request(struct rpc_xprt *xprt);
+void xprt_free_bc_request(struct rpc_rqst *req);
+int xprt_setup_backchannel(struct rpc_xprt *, unsigned int min_reqs);
+void xprt_destroy_backchannel(struct rpc_xprt *, int max_reqs);
+void bc_release_request(struct rpc_task *);
+int bc_send(struct rpc_rqst *req);
+#else /* CONFIG_NFS_V4_1 */
+static inline int xprt_setup_backchannel(struct rpc_xprt *xprt,
+					 unsigned int min_reqs)
+{
+	return 0;
+}
+#endif /* CONFIG_NFS_V4_1 */
+#endif /* _LINUX_SUNRPC_BC_XPRT_H */
+

include/linux/sunrpc/clnt.h

@@ -143,6 +143,7 @@ int		rpc_call_sync(struct rpc_clnt *clnt,
 			      const struct rpc_message *msg, int flags);
 struct rpc_task *rpc_call_null(struct rpc_clnt *clnt, struct rpc_cred *cred,
 			       int flags);
+void		rpc_restart_call_prepare(struct rpc_task *);
 void		rpc_restart_call(struct rpc_task *);
 void		rpc_setbufsize(struct rpc_clnt *, unsigned int, unsigned int);
 size_t		rpc_max_payload(struct rpc_clnt *);

include/linux/sunrpc/sched.h

@@ -210,6 +210,8 @@ struct rpc_wait_queue {
  */
 struct rpc_task *rpc_new_task(const struct rpc_task_setup *);
 struct rpc_task *rpc_run_task(const struct rpc_task_setup *);
+struct rpc_task *rpc_run_bc_task(struct rpc_rqst *req,
+				const struct rpc_call_ops *ops);
 void		rpc_put_task(struct rpc_task *);
 void		rpc_exit_task(struct rpc_task *);
 void		rpc_release_calldata(const struct rpc_call_ops *, void *);
@@ -237,6 +239,7 @@ void		rpc_show_tasks(void);
 int		rpc_init_mempool(void);
 void		rpc_destroy_mempool(void);
 extern struct workqueue_struct *rpciod_workqueue;
+void		rpc_prepare_task(struct rpc_task *task);
 
 static inline void rpc_exit(struct rpc_task *task, int status)
 {

include/linux/sunrpc/svc.h

@@ -96,6 +96,15 @@ struct svc_serv {
 	svc_thread_fn		sv_function;	/* main function for threads */
 	unsigned int		sv_drc_max_pages; /* Total pages for DRC */
 	unsigned int		sv_drc_pages_used;/* DRC pages used */
+#if defined(CONFIG_NFS_V4_1)
+	struct list_head	sv_cb_list;	/* queue for callback requests
+						 * that arrive over the same
+						 * connection */
+	spinlock_t		sv_cb_lock;	/* protects the svc_cb_list */
+	wait_queue_head_t	sv_cb_waitq;	/* sleep here if there are no
+						 * entries in the svc_cb_list */
+	struct svc_xprt		*bc_xprt;
+#endif /* CONFIG_NFS_V4_1 */
 };
 
 /*
@@ -411,6 +420,8 @@ int		   svc_set_num_threads(struct svc_serv *, struct svc_pool *, int);
 int		   svc_pool_stats_open(struct svc_serv *serv, struct file *file);
 void		   svc_destroy(struct svc_serv *);
 int		   svc_process(struct svc_rqst *);
+int		   bc_svc_process(struct svc_serv *, struct rpc_rqst *,
+			struct svc_rqst *);
 int		   svc_register(const struct svc_serv *, const int,
 				const unsigned short, const unsigned short);
 

include/linux/sunrpc/svcsock.h

@@ -42,6 +42,8 @@ int		svc_sock_names(char *buf, struct svc_serv *serv, char *toclose);
 int		svc_addsock(struct svc_serv *serv, int fd, char *name_return);
 void		svc_init_xprt_sock(void);
 void		svc_cleanup_xprt_sock(void);
+struct svc_xprt *svc_sock_create(struct svc_serv *serv, int prot);
+void		svc_sock_destroy(struct svc_xprt *);
 
 /*
  * svc_makesock socket characteristics

include/linux/sunrpc/xprt.h

@@ -67,7 +67,8 @@ struct rpc_rqst {
 	struct rpc_task *	rq_task;	/* RPC task data */
 	__be32			rq_xid;		/* request XID */
 	int			rq_cong;	/* has incremented xprt->cong */
-	int			rq_received;	/* receive completed */
+	int			rq_reply_bytes_recvd;	/* number of reply */
+							/* bytes received */
 	u32			rq_seqno;	/* gss seq no. used on req. */
 	int			rq_enc_pages_num;
 	struct page		**rq_enc_pages;	/* scratch pages for use by
@@ -97,6 +98,12 @@ struct rpc_rqst {
 
 	unsigned long		rq_xtime;	/* when transmitted */
 	int			rq_ntrans;
+
+#if defined(CONFIG_NFS_V4_1)
+	struct list_head	rq_bc_list;	/* Callback service list */
+	unsigned long		rq_bc_pa_state;	/* Backchannel prealloc state */
+	struct list_head	rq_bc_pa_list;	/* Backchannel prealloc list */
+#endif /* CONFIG_NFS_V4_1 */
 };
 #define rq_svec			rq_snd_buf.head
 #define rq_slen			rq_snd_buf.len
@@ -174,6 +181,15 @@ struct rpc_xprt {
 	spinlock_t		reserve_lock;	/* lock slot table */
 	u32			xid;		/* Next XID value to use */
 	struct rpc_task *	snd_task;	/* Task blocked in send */
+#if defined(CONFIG_NFS_V4_1)
+	struct svc_serv		*bc_serv;       /* The RPC service which will */
+						/* process the callback */
+	unsigned int		bc_alloc_count;	/* Total number of preallocs */
+	spinlock_t		bc_pa_lock;	/* Protects the preallocated
+						 * items */
+	struct list_head	bc_pa_list;	/* List of preallocated
+						 * backchannel rpc_rqst's */
+#endif /* CONFIG_NFS_V4_1 */
 	struct list_head	recv;
 
 	struct {
@@ -192,6 +208,26 @@ struct rpc_xprt {
 	const char		*address_strings[RPC_DISPLAY_MAX];
 };
 
+#if defined(CONFIG_NFS_V4_1)
+/*
+ * Backchannel flags
+ */
+#define	RPC_BC_PA_IN_USE	0x0001		/* Preallocated backchannel */
+						/* buffer in use */
+#endif /* CONFIG_NFS_V4_1 */
+
+#if defined(CONFIG_NFS_V4_1)
+static inline int bc_prealloc(struct rpc_rqst *req)
+{
+	return test_bit(RPC_BC_PA_IN_USE, &req->rq_bc_pa_state);
+}
+#else
+static inline int bc_prealloc(struct rpc_rqst *req)
+{
+	return 0;
+}
+#endif /* CONFIG_NFS_V4_1 */
+
 struct xprt_create {
 	int			ident;		/* XPRT_TRANSPORT identifier */
 	struct sockaddr *	srcaddr;	/* optional local address */

net/sunrpc/Makefile

@@ -13,5 +13,6 @@ sunrpc-y := clnt.o xprt.o socklib.o xprtsock.o sched.o \
 	    rpcb_clnt.o timer.o xdr.o \
 	    sunrpc_syms.o cache.o rpc_pipe.o \
 	    svc_xprt.o
+sunrpc-$(CONFIG_NFS_V4_1) += backchannel_rqst.o bc_svc.o
 sunrpc-$(CONFIG_PROC_FS) += stats.o
 sunrpc-$(CONFIG_SYSCTL) += sysctl.o

net/sunrpc/backchannel_rqst.c

@@ -0,0 +1,278 @@
+/******************************************************************************
+
+(c) 2007 Network Appliance, Inc.  All Rights Reserved.
+(c) 2009 NetApp.  All Rights Reserved.
+
+NetApp provides this source code under the GPL v2 License.
+The GPL v2 license is available at
+http://opensource.org/licenses/gpl-license.php.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
+CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
+EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+******************************************************************************/
+
+#include <linux/tcp.h>
+#include <linux/sunrpc/xprt.h>
+
+#ifdef RPC_DEBUG
+#define RPCDBG_FACILITY	RPCDBG_TRANS
+#endif
+
+#if defined(CONFIG_NFS_V4_1)
+
+/*
+ * Helper routines that track the number of preallocation elements
+ * on the transport.
+ */
+static inline int xprt_need_to_requeue(struct rpc_xprt *xprt)
+{
+	return xprt->bc_alloc_count > 0;
+}
+
+static inline void xprt_inc_alloc_count(struct rpc_xprt *xprt, unsigned int n)
+{
+	xprt->bc_alloc_count += n;
+}
+
+static inline int xprt_dec_alloc_count(struct rpc_xprt *xprt, unsigned int n)
+{
+	return xprt->bc_alloc_count -= n;
+}
+
+/*
+ * Free the preallocated rpc_rqst structure and the memory
+ * buffers hanging off of it.
+ */
+static void xprt_free_allocation(struct rpc_rqst *req)
+{
+	struct xdr_buf *xbufp;
+
+	dprintk("RPC:        free allocations for req= %p\n", req);
+	BUG_ON(test_bit(RPC_BC_PA_IN_USE, &req->rq_bc_pa_state));
+	xbufp = &req->rq_private_buf;
+	free_page((unsigned long)xbufp->head[0].iov_base);
+	xbufp = &req->rq_snd_buf;
+	free_page((unsigned long)xbufp->head[0].iov_base);
+	list_del(&req->rq_bc_pa_list);
+	kfree(req);
+}
+
+/*
+ * Preallocate up to min_reqs structures and related buffers for use
+ * by the backchannel.  This function can be called multiple times
+ * when creating new sessions that use the same rpc_xprt.  The
+ * preallocated buffers are added to the pool of resources used by
+ * the rpc_xprt.  Anyone of these resources may be used used by an
+ * incoming callback request.  It's up to the higher levels in the
+ * stack to enforce that the maximum number of session slots is not
+ * being exceeded.
+ *
+ * Some callback arguments can be large.  For example, a pNFS server
+ * using multiple deviceids.  The list can be unbound, but the client
+ * has the ability to tell the server the maximum size of the callback
+ * requests.  Each deviceID is 16 bytes, so allocate one page
+ * for the arguments to have enough room to receive a number of these
+ * deviceIDs.  The NFS client indicates to the pNFS server that its
+ * callback requests can be up to 4096 bytes in size.
+ */
+int xprt_setup_backchannel(struct rpc_xprt *xprt, unsigned int min_reqs)
+{
+	struct page *page_rcv = NULL, *page_snd = NULL;
+	struct xdr_buf *xbufp = NULL;
+	struct rpc_rqst *req, *tmp;
+	struct list_head tmp_list;
+	int i;
+
+	dprintk("RPC:       setup backchannel transport\n");
+
+	/*
+	 * We use a temporary list to keep track of the preallocated
+	 * buffers.  Once we're done building the list we splice it
+	 * into the backchannel preallocation list off of the rpc_xprt
+	 * struct.  This helps minimize the amount of time the list
+	 * lock is held on the rpc_xprt struct.  It also makes cleanup
+	 * easier in case of memory allocation errors.
+	 */
+	INIT_LIST_HEAD(&tmp_list);
+	for (i = 0; i < min_reqs; i++) {
+		/* Pre-allocate one backchannel rpc_rqst */
+		req = kzalloc(sizeof(struct rpc_rqst), GFP_KERNEL);
+		if (req == NULL) {
+			printk(KERN_ERR "Failed to create bc rpc_rqst\n");
+			goto out_free;
+		}
+
+		/* Add the allocated buffer to the tmp list */
+		dprintk("RPC:       adding req= %p\n", req);
+		list_add(&req->rq_bc_pa_list, &tmp_list);
+
+		req->rq_xprt = xprt;
+		INIT_LIST_HEAD(&req->rq_list);
+		INIT_LIST_HEAD(&req->rq_bc_list);
+
+		/* Preallocate one XDR receive buffer */
+		page_rcv = alloc_page(GFP_KERNEL);
+		if (page_rcv == NULL) {
+			printk(KERN_ERR "Failed to create bc receive xbuf\n");
+			goto out_free;
+		}
+		xbufp = &req->rq_rcv_buf;
+		xbufp->head[0].iov_base = page_address(page_rcv);
+		xbufp->head[0].iov_len = PAGE_SIZE;
+		xbufp->tail[0].iov_base = NULL;
+		xbufp->tail[0].iov_len = 0;
+		xbufp->page_len = 0;
+		xbufp->len = PAGE_SIZE;
+		xbufp->buflen = PAGE_SIZE;
+
+		/* Preallocate one XDR send buffer */
+		page_snd = alloc_page(GFP_KERNEL);
+		if (page_snd == NULL) {
+			printk(KERN_ERR "Failed to create bc snd xbuf\n");
+			goto out_free;
+		}
+
+		xbufp = &req->rq_snd_buf;
+		xbufp->head[0].iov_base = page_address(page_snd);
+		xbufp->head[0].iov_len = 0;
+		xbufp->tail[0].iov_base = NULL;
+		xbufp->tail[0].iov_len = 0;
+		xbufp->page_len = 0;
+		xbufp->len = 0;
+		xbufp->buflen = PAGE_SIZE;
+	}
+
+	/*
+	 * Add the temporary list to the backchannel preallocation list
+	 */
+	spin_lock_bh(&xprt->bc_pa_lock);
+	list_splice(&tmp_list, &xprt->bc_pa_list);
+	xprt_inc_alloc_count(xprt, min_reqs);
+	spin_unlock_bh(&xprt->bc_pa_lock);
+
+	dprintk("RPC:       setup backchannel transport done\n");
+	return 0;
+
+out_free:
+	/*
+	 * Memory allocation failed, free the temporary list
+	 */
+	list_for_each_entry_safe(req, tmp, &tmp_list, rq_bc_pa_list)
+		xprt_free_allocation(req);
+
+	dprintk("RPC:       setup backchannel transport failed\n");
+	return -1;
+}
+EXPORT_SYMBOL(xprt_setup_backchannel);
+
+/*
+ * Destroys the backchannel preallocated structures.
+ * Since these structures may have been allocated by multiple calls
+ * to xprt_setup_backchannel, we only destroy up to the maximum number
+ * of reqs specified by the caller.
+ * @xprt:	the transport holding the preallocated strucures
+ * @max_reqs	the maximum number of preallocated structures to destroy
+ */
+void xprt_destroy_backchannel(struct rpc_xprt *xprt, unsigned int max_reqs)
+{
+	struct rpc_rqst *req = NULL, *tmp = NULL;
+
+	dprintk("RPC:        destroy backchannel transport\n");
+
+	BUG_ON(max_reqs == 0);
+	spin_lock_bh(&xprt->bc_pa_lock);
+	xprt_dec_alloc_count(xprt, max_reqs);
+	list_for_each_entry_safe(req, tmp, &xprt->bc_pa_list, rq_bc_pa_list) {
+		dprintk("RPC:        req=%p\n", req);
+		xprt_free_allocation(req);
+		if (--max_reqs == 0)
+			break;
+	}
+	spin_unlock_bh(&xprt->bc_pa_lock);
+
+	dprintk("RPC:        backchannel list empty= %s\n",
+		list_empty(&xprt->bc_pa_list) ? "true" : "false");
+}
+EXPORT_SYMBOL(xprt_destroy_backchannel);
+
+/*
+ * One or more rpc_rqst structure have been preallocated during the
+ * backchannel setup.  Buffer space for the send and private XDR buffers
+ * has been preallocated as well.  Use xprt_alloc_bc_request to allocate
+ * to this request.  Use xprt_free_bc_request to return it.
+ *
+ * Return an available rpc_rqst, otherwise NULL if non are available.
+ */
+struct rpc_rqst *xprt_alloc_bc_request(struct rpc_xprt *xprt)
+{
+	struct rpc_rqst *req;
+
+	dprintk("RPC:       allocate a backchannel request\n");
+	spin_lock_bh(&xprt->bc_pa_lock);
+	if (!list_empty(&xprt->bc_pa_list)) {
+		req = list_first_entry(&xprt->bc_pa_list, struct rpc_rqst,
+				rq_bc_pa_list);
+		list_del(&req->rq_bc_pa_list);
+	} else {
+		req = NULL;
+	}
+	spin_unlock_bh(&xprt->bc_pa_lock);
+
+	if (req != NULL) {
+		set_bit(RPC_BC_PA_IN_USE, &req->rq_bc_pa_state);
+		req->rq_reply_bytes_recvd = 0;
+		req->rq_bytes_sent = 0;
+		memcpy(&req->rq_private_buf, &req->rq_rcv_buf,
+			sizeof(req->rq_private_buf));
+	}
+	dprintk("RPC:       backchannel req=%p\n", req);
+	return req;
+}
+
+/*
+ * Return the preallocated rpc_rqst structure and XDR buffers
+ * associated with this rpc_task.
+ */
+void xprt_free_bc_request(struct rpc_rqst *req)
+{
+	struct rpc_xprt *xprt = req->rq_xprt;
+
+	dprintk("RPC:       free backchannel req=%p\n", req);
+
+	smp_mb__before_clear_bit();
+	BUG_ON(!test_bit(RPC_BC_PA_IN_USE, &req->rq_bc_pa_state));
+	clear_bit(RPC_BC_PA_IN_USE, &req->rq_bc_pa_state);
+	smp_mb__after_clear_bit();
+
+	if (!xprt_need_to_requeue(xprt)) {
+		/*
+		 * The last remaining session was destroyed while this
+		 * entry was in use.  Free the entry and don't attempt
+		 * to add back to the list because there is no need to
+		 * have anymore preallocated entries.
+		 */
+		dprintk("RPC:       Last session removed req=%p\n", req);
+		xprt_free_allocation(req);
+		return;
+	}
+
+	/*
+	 * Return it to the list of preallocations so that it
+	 * may be reused by a new callback request.
+	 */
+	spin_lock_bh(&xprt->bc_pa_lock);
+	list_add(&req->rq_bc_pa_list, &xprt->bc_pa_list);
+	spin_unlock_bh(&xprt->bc_pa_lock);
+}
+
+#endif /* CONFIG_NFS_V4_1 */

net/sunrpc/bc_svc.c

@@ -0,0 +1,81 @@
+/******************************************************************************
+
+(c) 2007 Network Appliance, Inc.  All Rights Reserved.
+(c) 2009 NetApp.  All Rights Reserved.
+
+NetApp provides this source code under the GPL v2 License.
+The GPL v2 license is available at
+http://opensource.org/licenses/gpl-license.php.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
+CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
+EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+******************************************************************************/
+
+/*
+ * The NFSv4.1 callback service helper routines.
+ * They implement the transport level processing required to send the
+ * reply over an existing open connection previously established by the client.
+ */
+
+#if defined(CONFIG_NFS_V4_1)
+
+#include <linux/module.h>
+
+#include <linux/sunrpc/xprt.h>
+#include <linux/sunrpc/sched.h>
+#include <linux/sunrpc/bc_xprt.h>
+
+#define RPCDBG_FACILITY	RPCDBG_SVCDSP
+
+void bc_release_request(struct rpc_task *task)
+{
+	struct rpc_rqst *req = task->tk_rqstp;
+
+	dprintk("RPC:       bc_release_request: task= %p\n", task);
+
+	/*
+	 * Release this request only if it's a backchannel
+	 * preallocated request
+	 */
+	if (!bc_prealloc(req))
+		return;
+	xprt_free_bc_request(req);
+}
+
+/* Empty callback ops */
+static const struct rpc_call_ops nfs41_callback_ops = {
+};
+
+
+/*
+ * Send the callback reply
+ */
+int bc_send(struct rpc_rqst *req)
+{
+	struct rpc_task *task;
+	int ret;
+
+	dprintk("RPC:       bc_send req= %p\n", req);
+	task = rpc_run_bc_task(req, &nfs41_callback_ops);
+	if (IS_ERR(task))
+		ret = PTR_ERR(task);
+	else {
+		BUG_ON(atomic_read(&task->tk_count) != 1);
+		ret = task->tk_status;
+		rpc_put_task(task);
+	}
+	return ret;
+	dprintk("RPC:       bc_send ret= %d \n", ret);
+}
+
+#endif /* CONFIG_NFS_V4_1 */

net/sunrpc/clnt.c

@@ -36,7 +36,9 @@
 #include <linux/sunrpc/clnt.h>
 #include <linux/sunrpc/rpc_pipe_fs.h>
 #include <linux/sunrpc/metrics.h>
+#include <linux/sunrpc/bc_xprt.h>
 
+#include "sunrpc.h"
 
 #ifdef RPC_DEBUG
 # define RPCDBG_FACILITY	RPCDBG_CALL
@@ -63,6 +65,9 @@ static void	call_decode(struct rpc_task *task);
 static void	call_bind(struct rpc_task *task);
 static void	call_bind_status(struct rpc_task *task);
 static void	call_transmit(struct rpc_task *task);
+#if defined(CONFIG_NFS_V4_1)
+static void	call_bc_transmit(struct rpc_task *task);
+#endif /* CONFIG_NFS_V4_1 */
 static void	call_status(struct rpc_task *task);
 static void	call_transmit_status(struct rpc_task *task);
 static void	call_refresh(struct rpc_task *task);
@@ -613,6 +618,50 @@ rpc_call_async(struct rpc_clnt *clnt, const struct rpc_message *msg, int flags,
 }
 EXPORT_SYMBOL_GPL(rpc_call_async);
 
+#if defined(CONFIG_NFS_V4_1)
+/**
+ * rpc_run_bc_task - Allocate a new RPC task for backchannel use, then run
+ * rpc_execute against it
+ * @ops: RPC call ops
+ */
+struct rpc_task *rpc_run_bc_task(struct rpc_rqst *req,
+					const struct rpc_call_ops *tk_ops)
+{
+	struct rpc_task *task;
+	struct xdr_buf *xbufp = &req->rq_snd_buf;
+	struct rpc_task_setup task_setup_data = {
+		.callback_ops = tk_ops,
+	};
+
+	dprintk("RPC: rpc_run_bc_task req= %p\n", req);
+	/*
+	 * Create an rpc_task to send the data
+	 */
+	task = rpc_new_task(&task_setup_data);
+	if (!task) {
+		xprt_free_bc_request(req);
+		goto out;
+	}
+	task->tk_rqstp = req;
+
+	/*
+	 * Set up the xdr_buf length.
+	 * This also indicates that the buffer is XDR encoded already.
+	 */
+	xbufp->len = xbufp->head[0].iov_len + xbufp->page_len +
+			xbufp->tail[0].iov_len;
+
+	task->tk_action = call_bc_transmit;
+	atomic_inc(&task->tk_count);
+	BUG_ON(atomic_read(&task->tk_count) != 2);
+	rpc_execute(task);
+
+out:
+	dprintk("RPC: rpc_run_bc_task: task= %p\n", task);
+	return task;
+}
+#endif /* CONFIG_NFS_V4_1 */
+
 void
 rpc_call_start(struct rpc_task *task)
 {
@@ -694,6 +743,19 @@ void rpc_force_rebind(struct rpc_clnt *clnt)
 }
 EXPORT_SYMBOL_GPL(rpc_force_rebind);
 
+/*
+ * Restart an (async) RPC call from the call_prepare state.
+ * Usually called from within the exit handler.
+ */
+void
+rpc_restart_call_prepare(struct rpc_task *task)
+{
+	if (RPC_ASSASSINATED(task))
+		return;
+	task->tk_action = rpc_prepare_task;
+}
+EXPORT_SYMBOL_GPL(rpc_restart_call_prepare);
+
 /*
  * Restart an (async) RPC call. Usually called from within the
  * exit handler.
@@ -1085,7 +1147,7 @@ call_transmit(struct rpc_task *task)
 	 * in order to allow access to the socket to other RPC requests.
 	 */
 	call_transmit_status(task);
-	if (task->tk_msg.rpc_proc->p_decode != NULL)
+	if (rpc_reply_expected(task))
 		return;
 	task->tk_action = rpc_exit_task;
 	rpc_wake_up_queued_task(&task->tk_xprt->pending, task);
@@ -1120,6 +1182,72 @@ call_transmit_status(struct rpc_task *task)
 	}
 }
 
+#if defined(CONFIG_NFS_V4_1)
+/*
+ * 5b.	Send the backchannel RPC reply.  On error, drop the reply.  In
+ * addition, disconnect on connectivity errors.
+ */
+static void
+call_bc_transmit(struct rpc_task *task)
+{
+	struct rpc_rqst *req = task->tk_rqstp;
+
+	BUG_ON(task->tk_status != 0);
+	task->tk_status = xprt_prepare_transmit(task);
+	if (task->tk_status == -EAGAIN) {
+		/*
+		 * Could not reserve the transport. Try again after the
+		 * transport is released.
+		 */
+		task->tk_status = 0;
+		task->tk_action = call_bc_transmit;
+		return;
+	}
+
+	task->tk_action = rpc_exit_task;
+	if (task->tk_status < 0) {
+		printk(KERN_NOTICE "RPC: Could not send backchannel reply "
+			"error: %d\n", task->tk_status);
+		return;
+	}
+
+	xprt_transmit(task);
+	xprt_end_transmit(task);
+	dprint_status(task);
+	switch (task->tk_status) {
+	case 0:
+		/* Success */
+		break;
+	case -EHOSTDOWN:
+	case -EHOSTUNREACH:
+	case -ENETUNREACH:
+	case -ETIMEDOUT:
+		/*
+		 * Problem reaching the server.  Disconnect and let the
+		 * forechannel reestablish the connection.  The server will
+		 * have to retransmit the backchannel request and we'll
+		 * reprocess it.  Since these ops are idempotent, there's no
+		 * need to cache our reply at this time.
+		 */
+		printk(KERN_NOTICE "RPC: Could not send backchannel reply "
+			"error: %d\n", task->tk_status);
+		xprt_conditional_disconnect(task->tk_xprt,
+			req->rq_connect_cookie);
+		break;
+	default:
+		/*
+		 * We were unable to reply and will have to drop the
+		 * request.  The server should reconnect and retransmit.
+		 */
+		BUG_ON(task->tk_status == -EAGAIN);
+		printk(KERN_NOTICE "RPC: Could not send backchannel reply "
+			"error: %d\n", task->tk_status);
+		break;
+	}
+	rpc_wake_up_queued_task(&req->rq_xprt->pending, task);
+}
+#endif /* CONFIG_NFS_V4_1 */
+
 /*
  * 6.	Sort out the RPC call status
  */
@@ -1130,8 +1258,8 @@ call_status(struct rpc_task *task)
 	struct rpc_rqst	*req = task->tk_rqstp;
 	int		status;
 
-	if (req->rq_received > 0 && !req->rq_bytes_sent)
-		task->tk_status = req->rq_received;
+	if (req->rq_reply_bytes_recvd > 0 && !req->rq_bytes_sent)
+		task->tk_status = req->rq_reply_bytes_recvd;
 
 	dprint_status(task);
 
@@ -1248,7 +1376,7 @@ call_decode(struct rpc_task *task)
 
 	/*
 	 * Ensure that we see all writes made by xprt_complete_rqst()
-	 * before it changed req->rq_received.
+	 * before it changed req->rq_reply_bytes_recvd.
 	 */
 	smp_rmb();
 	req->rq_rcv_buf.len = req->rq_private_buf.len;
@@ -1289,7 +1417,7 @@ out_retry:
 	task->tk_status = 0;
 	/* Note: rpc_verify_header() may have freed the RPC slot */
 	if (task->tk_rqstp == req) {
-		req->rq_received = req->rq_rcv_buf.len = 0;
+		req->rq_reply_bytes_recvd = req->rq_rcv_buf.len = 0;
 		if (task->tk_client->cl_discrtry)
 			xprt_conditional_disconnect(task->tk_xprt,
 					req->rq_connect_cookie);
@@ -1377,13 +1505,14 @@ rpc_verify_header(struct rpc_task *task)
 	}
 	if ((len -= 3) < 0)
 		goto out_overflow;
-	p += 1;	/* skip XID */
 
+	p += 1; /* skip XID */
 	if ((n = ntohl(*p++)) != RPC_REPLY) {
 		dprintk("RPC: %5u %s: not an RPC reply: %x\n",
-				task->tk_pid, __func__, n);
+			task->tk_pid, __func__, n);
 		goto out_garbage;
 	}
+
 	if ((n = ntohl(*p++)) != RPC_MSG_ACCEPTED) {
 		if (--len < 0)
 			goto out_overflow;

net/sunrpc/sched.c

@@ -569,7 +569,7 @@ EXPORT_SYMBOL_GPL(rpc_delay);
 /*
  * Helper to call task->tk_ops->rpc_call_prepare
  */
-static void rpc_prepare_task(struct rpc_task *task)
+void rpc_prepare_task(struct rpc_task *task)
 {
 	task->tk_ops->rpc_call_prepare(task, task->tk_calldata);
 }

net/sunrpc/stats.c

@@ -141,12 +141,14 @@ EXPORT_SYMBOL_GPL(rpc_free_iostats);
 void rpc_count_iostats(struct rpc_task *task)
 {
 	struct rpc_rqst *req = task->tk_rqstp;
-	struct rpc_iostats *stats = task->tk_client->cl_metrics;
+	struct rpc_iostats *stats;
 	struct rpc_iostats *op_metrics;
 	long rtt, execute, queue;
 
-	if (!stats || !req)
+	if (!task->tk_client || !task->tk_client->cl_metrics || !req)
 		return;
+
+	stats = task->tk_client->cl_metrics;
 	op_metrics = &stats[task->tk_msg.rpc_proc->p_statidx];
 
 	op_metrics->om_ops++;
@@ -154,7 +156,7 @@ void rpc_count_iostats(struct rpc_task *task)
 	op_metrics->om_timeouts += task->tk_timeouts;
 
 	op_metrics->om_bytes_sent += task->tk_bytes_sent;
-	op_metrics->om_bytes_recv += req->rq_received;
+	op_metrics->om_bytes_recv += req->rq_reply_bytes_recvd;
 
 	queue = (long)req->rq_xtime - task->tk_start;
 	if (queue < 0)

net/sunrpc/sunrpc.h

@@ -0,0 +1,37 @@
+/******************************************************************************
+
+(c) 2008 NetApp.  All Rights Reserved.
+
+NetApp provides this source code under the GPL v2 License.
+The GPL v2 license is available at
+http://opensource.org/licenses/gpl-license.php.
+
+THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS
+"AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT
+LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR
+A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR
+CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
+EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
+PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
+PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
+LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
+NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
+SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
+
+******************************************************************************/
+
+/*
+ * Functions and macros used internally by RPC
+ */
+
+#ifndef _NET_SUNRPC_SUNRPC_H
+#define _NET_SUNRPC_SUNRPC_H
+
+static inline int rpc_reply_expected(struct rpc_task *task)
+{
+	return (task->tk_msg.rpc_proc != NULL) &&
+		(task->tk_msg.rpc_proc->p_decode != NULL);
+}
+
+#endif /* _NET_SUNRPC_SUNRPC_H */
+

net/sunrpc/svc.c

@@ -25,6 +25,7 @@
 #include <linux/sunrpc/stats.h>
 #include <linux/sunrpc/svcsock.h>
 #include <linux/sunrpc/clnt.h>
+#include <linux/sunrpc/bc_xprt.h>
 
 #define RPCDBG_FACILITY	RPCDBG_SVCDSP
 
@@ -486,6 +487,10 @@ svc_destroy(struct svc_serv *serv)
 	if (svc_serv_is_pooled(serv))
 		svc_pool_map_put();
 
+#if defined(CONFIG_NFS_V4_1)
+	svc_sock_destroy(serv->bc_xprt);
+#endif /* CONFIG_NFS_V4_1 */
+
 	svc_unregister(serv);
 	kfree(serv->sv_pools);
 	kfree(serv);
@@ -970,20 +975,18 @@ svc_printk(struct svc_rqst *rqstp, const char *fmt, ...)
 }
 
 /*
- * Process the RPC request.
+ * Common routine for processing the RPC request.
  */
-int
-svc_process(struct svc_rqst *rqstp)
+static int
+svc_process_common(struct svc_rqst *rqstp, struct kvec *argv, struct kvec *resv)
 {
 	struct svc_program	*progp;
 	struct svc_version	*versp = NULL;	/* compiler food */
 	struct svc_procedure	*procp = NULL;
-	struct kvec *		argv = &rqstp->rq_arg.head[0];
-	struct kvec *		resv = &rqstp->rq_res.head[0];
 	struct svc_serv		*serv = rqstp->rq_server;
 	kxdrproc_t		xdr;
 	__be32			*statp;
-	u32			dir, prog, vers, proc;
+	u32			prog, vers, proc;
 	__be32			auth_stat, rpc_stat;
 	int			auth_res;
 	__be32			*reply_statp;
@@ -993,19 +996,6 @@ svc_process(struct svc_rqst *rqstp)
 	if (argv->iov_len < 6*4)
 		goto err_short_len;
 
-	/* setup response xdr_buf.
-	 * Initially it has just one page
-	 */
-	rqstp->rq_resused = 1;
-	resv->iov_base = page_address(rqstp->rq_respages[0]);
-	resv->iov_len = 0;
-	rqstp->rq_res.pages = rqstp->rq_respages + 1;
-	rqstp->rq_res.len = 0;
-	rqstp->rq_res.page_base = 0;
-	rqstp->rq_res.page_len = 0;
-	rqstp->rq_res.buflen = PAGE_SIZE;
-	rqstp->rq_res.tail[0].iov_base = NULL;
-	rqstp->rq_res.tail[0].iov_len = 0;
 	/* Will be turned off only in gss privacy case: */
 	rqstp->rq_splice_ok = 1;
 	/* Will be turned off only when NFSv4 Sessions are used */
@@ -1014,17 +1004,13 @@ svc_process(struct svc_rqst *rqstp)
 	/* Setup reply header */
 	rqstp->rq_xprt->xpt_ops->xpo_prep_reply_hdr(rqstp);
 
-	rqstp->rq_xid = svc_getu32(argv);
 	svc_putu32(resv, rqstp->rq_xid);
 
-	dir  = svc_getnl(argv);
 	vers = svc_getnl(argv);
 
 	/* First words of reply: */
 	svc_putnl(resv, 1);		/* REPLY */
 
-	if (dir != 0)		/* direction != CALL */
-		goto err_bad_dir;
 	if (vers != 2)		/* RPC version number */
 		goto err_bad_rpc;
 
@@ -1147,7 +1133,7 @@ svc_process(struct svc_rqst *rqstp)
  sendit:
 	if (svc_authorise(rqstp))
 		goto dropit;
-	return svc_send(rqstp);
+	return 1;		/* Caller can now send it */
 
  dropit:
 	svc_authorise(rqstp);	/* doesn't hurt to call this twice */
@@ -1161,12 +1147,6 @@ err_short_len:
 
 	goto dropit;			/* drop request */
 
-err_bad_dir:
-	svc_printk(rqstp, "bad direction %d, dropping request\n", dir);
-
-	serv->sv_stats->rpcbadfmt++;
-	goto dropit;			/* drop request */
-
 err_bad_rpc:
 	serv->sv_stats->rpcbadfmt++;
 	svc_putnl(resv, 1);	/* REJECT */
@@ -1219,6 +1199,100 @@ err_bad:
 }
 EXPORT_SYMBOL_GPL(svc_process);
 
+/*
+ * Process the RPC request.
+ */
+int
+svc_process(struct svc_rqst *rqstp)
+{
+	struct kvec		*argv = &rqstp->rq_arg.head[0];
+	struct kvec		*resv = &rqstp->rq_res.head[0];
+	struct svc_serv		*serv = rqstp->rq_server;
+	u32			dir;
+	int			error;
+
+	/*
+	 * Setup response xdr_buf.
+	 * Initially it has just one page
+	 */
+	rqstp->rq_resused = 1;
+	resv->iov_base = page_address(rqstp->rq_respages[0]);
+	resv->iov_len = 0;
+	rqstp->rq_res.pages = rqstp->rq_respages + 1;
+	rqstp->rq_res.len = 0;
+	rqstp->rq_res.page_base = 0;
+	rqstp->rq_res.page_len = 0;
+	rqstp->rq_res.buflen = PAGE_SIZE;
+	rqstp->rq_res.tail[0].iov_base = NULL;
+	rqstp->rq_res.tail[0].iov_len = 0;
+
+	rqstp->rq_xid = svc_getu32(argv);
+
+	dir  = svc_getnl(argv);
+	if (dir != 0) {
+		/* direction != CALL */
+		svc_printk(rqstp, "bad direction %d, dropping request\n", dir);
+		serv->sv_stats->rpcbadfmt++;
+		svc_drop(rqstp);
+		return 0;
+	}
+
+	error = svc_process_common(rqstp, argv, resv);
+	if (error <= 0)
+		return error;
+
+	return svc_send(rqstp);
+}
+
+#if defined(CONFIG_NFS_V4_1)
+/*
+ * Process a backchannel RPC request that arrived over an existing
+ * outbound connection
+ */
+int
+bc_svc_process(struct svc_serv *serv, struct rpc_rqst *req,
+	       struct svc_rqst *rqstp)
+{
+	struct kvec	*argv = &rqstp->rq_arg.head[0];
+	struct kvec	*resv = &rqstp->rq_res.head[0];
+	int 		error;
+
+	/* Build the svc_rqst used by the common processing routine */
+	rqstp->rq_xprt = serv->bc_xprt;
+	rqstp->rq_xid = req->rq_xid;
+	rqstp->rq_prot = req->rq_xprt->prot;
+	rqstp->rq_server = serv;
+
+	rqstp->rq_addrlen = sizeof(req->rq_xprt->addr);
+	memcpy(&rqstp->rq_addr, &req->rq_xprt->addr, rqstp->rq_addrlen);
+	memcpy(&rqstp->rq_arg, &req->rq_rcv_buf, sizeof(rqstp->rq_arg));
+	memcpy(&rqstp->rq_res, &req->rq_snd_buf, sizeof(rqstp->rq_res));
+
+	/* reset result send buffer "put" position */
+	resv->iov_len = 0;
+
+	if (rqstp->rq_prot != IPPROTO_TCP) {
+		printk(KERN_ERR "No support for Non-TCP transports!\n");
+		BUG();
+	}
+
+	/*
+	 * Skip the next two words because they've already been
+	 * processed in the trasport
+	 */
+	svc_getu32(argv);	/* XID */
+	svc_getnl(argv);	/* CALLDIR */
+
+	error = svc_process_common(rqstp, argv, resv);
+	if (error <= 0)
+		return error;
+
+	memcpy(&req->rq_snd_buf, &rqstp->rq_res, sizeof(req->rq_snd_buf));
+	return bc_send(req);
+}
+EXPORT_SYMBOL(bc_svc_process);
+#endif /* CONFIG_NFS_V4_1 */
+
 /*
  * Return (transport-specific) limit on the rpc payload.
  */

net/sunrpc/svcsock.c

@@ -1327,3 +1327,42 @@ static void svc_sock_free(struct svc_xprt *xprt)
 		sock_release(svsk->sk_sock);
 	kfree(svsk);
 }
+
+/*
+ * Create a svc_xprt.
+ *
+ * For internal use only (e.g. nfsv4.1 backchannel).
+ * Callers should typically use the xpo_create() method.
+ */
+struct svc_xprt *svc_sock_create(struct svc_serv *serv, int prot)
+{
+	struct svc_sock *svsk;
+	struct svc_xprt *xprt = NULL;
+
+	dprintk("svc: %s\n", __func__);
+	svsk = kzalloc(sizeof(*svsk), GFP_KERNEL);
+	if (!svsk)
+		goto out;
+
+	xprt = &svsk->sk_xprt;
+	if (prot == IPPROTO_TCP)
+		svc_xprt_init(&svc_tcp_class, xprt, serv);
+	else if (prot == IPPROTO_UDP)
+		svc_xprt_init(&svc_udp_class, xprt, serv);
+	else
+		BUG();
+out:
+	dprintk("svc: %s return %p\n", __func__, xprt);
+	return xprt;
+}
+EXPORT_SYMBOL_GPL(svc_sock_create);
+
+/*
+ * Destroy a svc_sock.
+ */
+void svc_sock_destroy(struct svc_xprt *xprt)
+{
+	if (xprt)
+		kfree(container_of(xprt, struct svc_sock, sk_xprt));
+}
+EXPORT_SYMBOL_GPL(svc_sock_destroy);

net/sunrpc/xprt.c

@@ -12,8 +12,9 @@
  *  -	Next, the caller puts together the RPC message, stuffs it into
  *	the request struct, and calls xprt_transmit().
  *  -	xprt_transmit sends the message and installs the caller on the
- *	transport's wait list. At the same time, it installs a timer that
- *	is run after the packet's timeout has expired.
+ *	transport's wait list. At the same time, if a reply is expected,
+ *	it installs a timer that is run after the packet's timeout has
+ *	expired.
  *  -	When a packet arrives, the data_ready handler walks the list of
  *	pending requests for that transport. If a matching XID is found, the
  *	caller is woken up, and the timer removed.
@@ -46,6 +47,8 @@
 #include <linux/sunrpc/clnt.h>
 #include <linux/sunrpc/metrics.h>
 
+#include "sunrpc.h"
+
 /*
  * Local variables
  */
@@ -192,8 +195,8 @@ EXPORT_SYMBOL_GPL(xprt_load_transport);
  */
 int xprt_reserve_xprt(struct rpc_task *task)
 {
-	struct rpc_xprt	*xprt = task->tk_xprt;
 	struct rpc_rqst *req = task->tk_rqstp;
+	struct rpc_xprt	*xprt = req->rq_xprt;
 
 	if (test_and_set_bit(XPRT_LOCKED, &xprt->state)) {
 		if (task == xprt->snd_task)
@@ -803,9 +806,10 @@ void xprt_complete_rqst(struct rpc_task *task, int copied)
 
 	list_del_init(&req->rq_list);
 	req->rq_private_buf.len = copied;
-	/* Ensure all writes are done before we update req->rq_received */
+	/* Ensure all writes are done before we update */
+	/* req->rq_reply_bytes_recvd */
 	smp_wmb();
-	req->rq_received = copied;
+	req->rq_reply_bytes_recvd = copied;
 	rpc_wake_up_queued_task(&xprt->pending, task);
 }
 EXPORT_SYMBOL_GPL(xprt_complete_rqst);
@@ -820,7 +824,7 @@ static void xprt_timer(struct rpc_task *task)
 	dprintk("RPC: %5u xprt_timer\n", task->tk_pid);
 
 	spin_lock_bh(&xprt->transport_lock);
-	if (!req->rq_received) {
+	if (!req->rq_reply_bytes_recvd) {
 		if (xprt->ops->timer)
 			xprt->ops->timer(task);
 	} else
@@ -842,8 +846,8 @@ int xprt_prepare_transmit(struct rpc_task *task)
 	dprintk("RPC: %5u xprt_prepare_transmit\n", task->tk_pid);
 
 	spin_lock_bh(&xprt->transport_lock);
-	if (req->rq_received && !req->rq_bytes_sent) {
-		err = req->rq_received;
+	if (req->rq_reply_bytes_recvd && !req->rq_bytes_sent) {
+		err = req->rq_reply_bytes_recvd;
 		goto out_unlock;
 	}
 	if (!xprt->ops->reserve_xprt(task))
@@ -855,7 +859,7 @@ out_unlock:
 
 void xprt_end_transmit(struct rpc_task *task)
 {
-	xprt_release_write(task->tk_xprt, task);
+	xprt_release_write(task->tk_rqstp->rq_xprt, task);
 }
 
 /**
@@ -872,8 +876,11 @@ void xprt_transmit(struct rpc_task *task)
 
 	dprintk("RPC: %5u xprt_transmit(%u)\n", task->tk_pid, req->rq_slen);
 
-	if (!req->rq_received) {
-		if (list_empty(&req->rq_list)) {
+	if (!req->rq_reply_bytes_recvd) {
+		if (list_empty(&req->rq_list) && rpc_reply_expected(task)) {
+			/*
+			 * Add to the list only if we're expecting a reply
+			 */
 			spin_lock_bh(&xprt->transport_lock);
 			/* Update the softirq receive buffer */
 			memcpy(&req->rq_private_buf, &req->rq_rcv_buf,
@@ -908,8 +915,13 @@ void xprt_transmit(struct rpc_task *task)
 	/* Don't race with disconnect */
 	if (!xprt_connected(xprt))
 		task->tk_status = -ENOTCONN;
-	else if (!req->rq_received)
+	else if (!req->rq_reply_bytes_recvd && rpc_reply_expected(task)) {
+		/*
+		 * Sleep on the pending queue since
+		 * we're expecting a reply.
+		 */
 		rpc_sleep_on(&xprt->pending, task, xprt_timer);
+	}
 	spin_unlock_bh(&xprt->transport_lock);
 }
 
@@ -982,11 +994,17 @@ static void xprt_request_init(struct rpc_task *task, struct rpc_xprt *xprt)
  */
 void xprt_release(struct rpc_task *task)
 {
-	struct rpc_xprt	*xprt = task->tk_xprt;
+	struct rpc_xprt	*xprt;
 	struct rpc_rqst	*req;
+	int is_bc_request;
 
 	if (!(req = task->tk_rqstp))
 		return;
+
+	/* Preallocated backchannel request? */
+	is_bc_request = bc_prealloc(req);
+
+	xprt = req->rq_xprt;
 	rpc_count_iostats(task);
 	spin_lock_bh(&xprt->transport_lock);
 	xprt->ops->release_xprt(xprt, task);
@@ -999,10 +1017,19 @@ void xprt_release(struct rpc_task *task)
 		mod_timer(&xprt->timer,
 				xprt->last_used + xprt->idle_timeout);
 	spin_unlock_bh(&xprt->transport_lock);
-	xprt->ops->buf_free(req->rq_buffer);
+	if (!bc_prealloc(req))
+		xprt->ops->buf_free(req->rq_buffer);
 	task->tk_rqstp = NULL;
 	if (req->rq_release_snd_buf)
 		req->rq_release_snd_buf(req);
+
+	/*
+	 * Early exit if this is a backchannel preallocated request.
+	 * There is no need to have it added to the RPC slot list.
+	 */
+	if (is_bc_request)
+		return;
+
 	memset(req, 0, sizeof(*req));	/* mark unused */
 
 	dprintk("RPC: %5u release request %p\n", task->tk_pid, req);
@@ -1049,6 +1076,11 @@ found:
 
 	INIT_LIST_HEAD(&xprt->free);
 	INIT_LIST_HEAD(&xprt->recv);
+#if defined(CONFIG_NFS_V4_1)
+	spin_lock_init(&xprt->bc_pa_lock);
+	INIT_LIST_HEAD(&xprt->bc_pa_list);
+#endif /* CONFIG_NFS_V4_1 */
+
 	INIT_WORK(&xprt->task_cleanup, xprt_autoclose);
 	setup_timer(&xprt->timer, xprt_init_autodisconnect,
 			(unsigned long)xprt);

net/sunrpc/xprtsock.c

@@ -34,6 +34,9 @@
 #include <linux/sunrpc/sched.h>
 #include <linux/sunrpc/xprtsock.h>
 #include <linux/file.h>
+#ifdef CONFIG_NFS_V4_1
+#include <linux/sunrpc/bc_xprt.h>
+#endif
 
 #include <net/sock.h>
 #include <net/checksum.h>
@@ -270,6 +273,13 @@ struct sock_xprt {
 #define TCP_RCV_COPY_FRAGHDR	(1UL << 1)
 #define TCP_RCV_COPY_XID	(1UL << 2)
 #define TCP_RCV_COPY_DATA	(1UL << 3)
+#define TCP_RCV_READ_CALLDIR	(1UL << 4)
+#define TCP_RCV_COPY_CALLDIR	(1UL << 5)
+
+/*
+ * TCP RPC flags
+ */
+#define TCP_RPC_REPLY		(1UL << 6)
 
 static inline struct sockaddr *xs_addr(struct rpc_xprt *xprt)
 {
@@ -956,7 +966,7 @@ static inline void xs_tcp_read_fraghdr(struct rpc_xprt *xprt, struct xdr_skb_rea
 	transport->tcp_offset = 0;
 
 	/* Sanity check of the record length */
-	if (unlikely(transport->tcp_reclen < 4)) {
+	if (unlikely(transport->tcp_reclen < 8)) {
 		dprintk("RPC:       invalid TCP record fragment length\n");
 		xprt_force_disconnect(xprt);
 		return;
@@ -991,33 +1001,77 @@ static inline void xs_tcp_read_xid(struct sock_xprt *transport, struct xdr_skb_r
 	if (used != len)
 		return;
 	transport->tcp_flags &= ~TCP_RCV_COPY_XID;
-	transport->tcp_flags |= TCP_RCV_COPY_DATA;
+	transport->tcp_flags |= TCP_RCV_READ_CALLDIR;
 	transport->tcp_copied = 4;
-	dprintk("RPC:       reading reply for XID %08x\n",
+	dprintk("RPC:       reading %s XID %08x\n",
+			(transport->tcp_flags & TCP_RPC_REPLY) ? "reply for"
+							      : "request with",
 			ntohl(transport->tcp_xid));
 	xs_tcp_check_fraghdr(transport);
 }
 
-static inline void xs_tcp_read_request(struct rpc_xprt *xprt, struct xdr_skb_reader *desc)
+static inline void xs_tcp_read_calldir(struct sock_xprt *transport,
+				       struct xdr_skb_reader *desc)
 {
-	struct sock_xprt *transport = container_of(xprt, struct sock_xprt, xprt);
-	struct rpc_rqst *req;
+	size_t len, used;
+	u32 offset;
+	__be32	calldir;
+
+	/*
+	 * We want transport->tcp_offset to be 8 at the end of this routine
+	 * (4 bytes for the xid and 4 bytes for the call/reply flag).
+	 * When this function is called for the first time,
+	 * transport->tcp_offset is 4 (after having already read the xid).
+	 */
+	offset = transport->tcp_offset - sizeof(transport->tcp_xid);
+	len = sizeof(calldir) - offset;
+	dprintk("RPC:       reading CALL/REPLY flag (%Zu bytes)\n", len);
+	used = xdr_skb_read_bits(desc, &calldir, len);
+	transport->tcp_offset += used;
+	if (used != len)
+		return;
+	transport->tcp_flags &= ~TCP_RCV_READ_CALLDIR;
+	transport->tcp_flags |= TCP_RCV_COPY_CALLDIR;
+	transport->tcp_flags |= TCP_RCV_COPY_DATA;
+	/*
+	 * We don't yet have the XDR buffer, so we will write the calldir
+	 * out after we get the buffer from the 'struct rpc_rqst'
+	 */
+	if (ntohl(calldir) == RPC_REPLY)
+		transport->tcp_flags |= TCP_RPC_REPLY;
+	else
+		transport->tcp_flags &= ~TCP_RPC_REPLY;
+	dprintk("RPC:       reading %s CALL/REPLY flag %08x\n",
+			(transport->tcp_flags & TCP_RPC_REPLY) ?
+				"reply for" : "request with", calldir);
+	xs_tcp_check_fraghdr(transport);
+}
+
+static inline void xs_tcp_read_common(struct rpc_xprt *xprt,
+				     struct xdr_skb_reader *desc,
+				     struct rpc_rqst *req)
+{
+	struct sock_xprt *transport =
+				container_of(xprt, struct sock_xprt, xprt);
 	struct xdr_buf *rcvbuf;
 	size_t len;
 	ssize_t r;
 
-	/* Find and lock the request corresponding to this xid */
-	spin_lock(&xprt->transport_lock);
-	req = xprt_lookup_rqst(xprt, transport->tcp_xid);
-	if (!req) {
-		transport->tcp_flags &= ~TCP_RCV_COPY_DATA;
-		dprintk("RPC:       XID %08x request not found!\n",
-				ntohl(transport->tcp_xid));
-		spin_unlock(&xprt->transport_lock);
-		return;
+	rcvbuf = &req->rq_private_buf;
+
+	if (transport->tcp_flags & TCP_RCV_COPY_CALLDIR) {
+		/*
+		 * Save the RPC direction in the XDR buffer
+		 */
+		__be32	calldir = transport->tcp_flags & TCP_RPC_REPLY ?
+					htonl(RPC_REPLY) : 0;
+
+		memcpy(rcvbuf->head[0].iov_base + transport->tcp_copied,
+			&calldir, sizeof(calldir));
+		transport->tcp_copied += sizeof(calldir);
+		transport->tcp_flags &= ~TCP_RCV_COPY_CALLDIR;
 	}
 
-	rcvbuf = &req->rq_private_buf;
 	len = desc->count;
 	if (len > transport->tcp_reclen - transport->tcp_offset) {
 		struct xdr_skb_reader my_desc;
@@ -1054,7 +1108,7 @@ static inline void xs_tcp_read_request(struct rpc_xprt *xprt, struct xdr_skb_rea
 				"tcp_offset = %u, tcp_reclen = %u\n",
 				xprt, transport->tcp_copied,
 				transport->tcp_offset, transport->tcp_reclen);
-		goto out;
+		return;
 	}
 
 	dprintk("RPC:       XID %08x read %Zd bytes\n",
@@ -1070,11 +1124,125 @@ static inline void xs_tcp_read_request(struct rpc_xprt *xprt, struct xdr_skb_rea
 			transport->tcp_flags &= ~TCP_RCV_COPY_DATA;
 	}
 
-out:
+	return;
+}
+
+/*
+ * Finds the request corresponding to the RPC xid and invokes the common
+ * tcp read code to read the data.
+ */
+static inline int xs_tcp_read_reply(struct rpc_xprt *xprt,
+				    struct xdr_skb_reader *desc)
+{
+	struct sock_xprt *transport =
+				container_of(xprt, struct sock_xprt, xprt);
+	struct rpc_rqst *req;
+
+	dprintk("RPC:       read reply XID %08x\n", ntohl(transport->tcp_xid));
+
+	/* Find and lock the request corresponding to this xid */
+	spin_lock(&xprt->transport_lock);
+	req = xprt_lookup_rqst(xprt, transport->tcp_xid);
+	if (!req) {
+		dprintk("RPC:       XID %08x request not found!\n",
+				ntohl(transport->tcp_xid));
+		spin_unlock(&xprt->transport_lock);
+		return -1;
+	}
+
+	xs_tcp_read_common(xprt, desc, req);
+
 	if (!(transport->tcp_flags & TCP_RCV_COPY_DATA))
 		xprt_complete_rqst(req->rq_task, transport->tcp_copied);
+
 	spin_unlock(&xprt->transport_lock);
-	xs_tcp_check_fraghdr(transport);
+	return 0;
+}
+
+#if defined(CONFIG_NFS_V4_1)
+/*
+ * Obtains an rpc_rqst previously allocated and invokes the common
+ * tcp read code to read the data.  The result is placed in the callback
+ * queue.
+ * If we're unable to obtain the rpc_rqst we schedule the closing of the
+ * connection and return -1.
+ */
+static inline int xs_tcp_read_callback(struct rpc_xprt *xprt,
+				       struct xdr_skb_reader *desc)
+{
+	struct sock_xprt *transport =
+				container_of(xprt, struct sock_xprt, xprt);
+	struct rpc_rqst *req;
+
+	req = xprt_alloc_bc_request(xprt);
+	if (req == NULL) {
+		printk(KERN_WARNING "Callback slot table overflowed\n");
+		xprt_force_disconnect(xprt);
+		return -1;
+	}
+
+	req->rq_xid = transport->tcp_xid;
+	dprintk("RPC:       read callback  XID %08x\n", ntohl(req->rq_xid));
+	xs_tcp_read_common(xprt, desc, req);
+
+	if (!(transport->tcp_flags & TCP_RCV_COPY_DATA)) {
+		struct svc_serv *bc_serv = xprt->bc_serv;
+
+		/*
+		 * Add callback request to callback list.  The callback
+		 * service sleeps on the sv_cb_waitq waiting for new
+		 * requests.  Wake it up after adding enqueing the
+		 * request.
+		 */
+		dprintk("RPC:       add callback request to list\n");
+		spin_lock(&bc_serv->sv_cb_lock);
+		list_add(&req->rq_bc_list, &bc_serv->sv_cb_list);
+		spin_unlock(&bc_serv->sv_cb_lock);
+		wake_up(&bc_serv->sv_cb_waitq);
+	}
+
+	req->rq_private_buf.len = transport->tcp_copied;
+
+	return 0;
+}
+
+static inline int _xs_tcp_read_data(struct rpc_xprt *xprt,
+					struct xdr_skb_reader *desc)
+{
+	struct sock_xprt *transport =
+				container_of(xprt, struct sock_xprt, xprt);
+
+	return (transport->tcp_flags & TCP_RPC_REPLY) ?
+		xs_tcp_read_reply(xprt, desc) :
+		xs_tcp_read_callback(xprt, desc);
+}
+#else
+static inline int _xs_tcp_read_data(struct rpc_xprt *xprt,
+					struct xdr_skb_reader *desc)
+{
+	return xs_tcp_read_reply(xprt, desc);
+}
+#endif /* CONFIG_NFS_V4_1 */
+
+/*
+ * Read data off the transport.  This can be either an RPC_CALL or an
+ * RPC_REPLY.  Relay the processing to helper functions.
+ */
+static void xs_tcp_read_data(struct rpc_xprt *xprt,
+				    struct xdr_skb_reader *desc)
+{
+	struct sock_xprt *transport =
+				container_of(xprt, struct sock_xprt, xprt);
+
+	if (_xs_tcp_read_data(xprt, desc) == 0)
+		xs_tcp_check_fraghdr(transport);
+	else {
+		/*
+		 * The transport_lock protects the request handling.
+		 * There's no need to hold it to update the tcp_flags.
+		 */
+		transport->tcp_flags &= ~TCP_RCV_COPY_DATA;
+	}
 }
 
 static inline void xs_tcp_read_discard(struct sock_xprt *transport, struct xdr_skb_reader *desc)
@@ -1114,9 +1282,14 @@ static int xs_tcp_data_recv(read_descriptor_t *rd_desc, struct sk_buff *skb, uns
 			xs_tcp_read_xid(transport, &desc);
 			continue;
 		}
+		/* Read in the call/reply flag */
+		if (transport->tcp_flags & TCP_RCV_READ_CALLDIR) {
+			xs_tcp_read_calldir(transport, &desc);
+			continue;
+		}
 		/* Read in the request data */
 		if (transport->tcp_flags & TCP_RCV_COPY_DATA) {
-			xs_tcp_read_request(xprt, &desc);
+			xs_tcp_read_data(xprt, &desc);
 			continue;
 		}
 		/* Skip over any trailing bytes on short reads */
@@ -2010,6 +2183,9 @@ static struct rpc_xprt_ops xs_tcp_ops = {
 	.buf_free		= rpc_free,
 	.send_request		= xs_tcp_send_request,
 	.set_retrans_timeout	= xprt_set_retrans_timeout_def,
+#if defined(CONFIG_NFS_V4_1)
+	.release_request	= bc_release_request,
+#endif /* CONFIG_NFS_V4_1 */
 	.close			= xs_tcp_close,
 	.destroy		= xs_destroy,
 	.print_stats		= xs_tcp_print_stats,