Merge tag 'nfs-for-4.6-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs

Pull NFS client updates from Trond Myklebust:
 "Highlights include:

  Features:
   - Add support for multiple NFSv4.1 callbacks in flight
   - Initial patchset for RPC multipath support
   - Adapt RPC/RDMA to use the new completion queue API

  Bugfixes and cleanups:
   - nfs4: nfs4_ff_layout_prepare_ds should return NULL if connection failed
   - Cleanups to remove nfs_inode_dio_wait and nfs4_file_fsync
   - Fix RPC/RDMA credit accounting
   - Properly handle RDMA_ERROR replies
   - xprtrdma: Do not wait if ib_post_send() fails
   - xprtrdma: Segment head and tail XDR buffers on page boundaries
   - xprtrdma cleanups for dprintk, physical_op_map and unused macros"

* tag 'nfs-for-4.6-1' of git://git.linux-nfs.org/projects/trondmy/linux-nfs: (35 commits)
  nfs/blocklayout: make sure making a aligned read request
  nfs4: nfs4_ff_layout_prepare_ds should return NULL if connection failed
  nfs: remove nfs_inode_dio_wait
  nfs: remove nfs4_file_fsync
  xprtrdma: Use new CQ API for RPC-over-RDMA client send CQs
  xprtrdma: Use an anonymous union in struct rpcrdma_mw
  xprtrdma: Use new CQ API for RPC-over-RDMA client receive CQs
  xprtrdma: Serialize credit accounting again
  xprtrdma: Properly handle RDMA_ERROR replies
  rpcrdma: Add RPCRDMA_HDRLEN_ERR
  xprtrdma: Do not wait if ib_post_send() fails
  xprtrdma: Segment head and tail XDR buffers on page boundaries
  xprtrdma: Clean up dprintk format string containing a newline
  xprtrdma: Clean up physical_op_map()
  xprtrdma: Clean up unused RPCRDMA_INLINE_PAD_THRESH macro
  NFS add callback_ops to nfs4_proc_bind_conn_to_session_callback
  pnfs/NFSv4.1: Add multipath capabilities to pNFS flexfiles servers over NFSv3
  SUNRPC: Allow addition of new transports to a struct rpc_clnt
  NFSv4.1: nfs4_proc_bind_conn_to_session must iterate over all connections
  SUNRPC: Make NFS swap work with multipath
  ...

13 files changed, 187 insertions, 106 deletions

diff --git a/fs/nfs/blocklayout/blocklayout.c b/fs/nfs/blocklayout/blocklayout.c
index ddd0138..8bc870e 100644
--- a/fs/nfs/blocklayout/blocklayout.c
+++ b/fs/nfs/blocklayout/blocklayout.c
@@ -743,7 +743,7 @@ bl_set_layoutdriver(struct nfs_server *server, const struct nfs_fh *fh)
 
 static bool
 is_aligned_req(struct nfs_pageio_descriptor *pgio,
-		struct nfs_page *req, unsigned int alignment)
+		struct nfs_page *req, unsigned int alignment, bool is_write)
 {
 	/*
 	 * Always accept buffered writes, higher layers take care of the
@@ -758,7 +758,8 @@ is_aligned_req(struct nfs_pageio_descriptor *pgio,
 	if (IS_ALIGNED(req->wb_bytes, alignment))
 		return true;
 
-	if (req_offset(req) + req->wb_bytes == i_size_read(pgio->pg_inode)) {
+	if (is_write &&
+	    (req_offset(req) + req->wb_bytes == i_size_read(pgio->pg_inode))) {
 		/*
 		 * If the write goes up to the inode size, just write
 		 * the full page.  Data past the inode size is
@@ -775,7 +776,7 @@ is_aligned_req(struct nfs_pageio_descriptor *pgio,
 static void
 bl_pg_init_read(struct nfs_pageio_descriptor *pgio, struct nfs_page *req)
 {
-	if (!is_aligned_req(pgio, req, SECTOR_SIZE)) {
+	if (!is_aligned_req(pgio, req, SECTOR_SIZE, false)) {
 		nfs_pageio_reset_read_mds(pgio);
 		return;
 	}
@@ -791,7 +792,7 @@ static size_t
 bl_pg_test_read(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev,
 		struct nfs_page *req)
 {
-	if (!is_aligned_req(pgio, req, SECTOR_SIZE))
+	if (!is_aligned_req(pgio, req, SECTOR_SIZE, false))
 		return 0;
 	return pnfs_generic_pg_test(pgio, prev, req);
 }
@@ -824,7 +825,7 @@ bl_pg_init_write(struct nfs_pageio_descriptor *pgio, struct nfs_page *req)
 {
 	u64 wb_size;
 
-	if (!is_aligned_req(pgio, req, PAGE_SIZE)) {
+	if (!is_aligned_req(pgio, req, PAGE_SIZE, true)) {
 		nfs_pageio_reset_write_mds(pgio);
 		return;
 	}
@@ -846,7 +847,7 @@ static size_t
 bl_pg_test_write(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev,
 		 struct nfs_page *req)
 {
-	if (!is_aligned_req(pgio, req, PAGE_SIZE))
+	if (!is_aligned_req(pgio, req, PAGE_SIZE, true))
 		return 0;
 	return pnfs_generic_pg_test(pgio, prev, req);
 }
diff --git a/fs/nfs/callback.h b/fs/nfs/callback.h
index ff8195b..5fe1cec 100644
--- a/fs/nfs/callback.h
+++ b/fs/nfs/callback.h
@@ -37,10 +37,11 @@ enum nfs4_callback_opnum {
 	OP_CB_ILLEGAL = 10044,
 };
 
+struct nfs4_slot;
 struct cb_process_state {
 	__be32			drc_status;
 	struct nfs_client	*clp;
-	u32			slotid;
+	struct nfs4_slot	*slot;
 	u32			minorversion;
 	struct net		*net;
 };
diff --git a/fs/nfs/callback_proc.c b/fs/nfs/callback_proc.c
index f0939d0..618ced3 100644
--- a/fs/nfs/callback_proc.c
+++ b/fs/nfs/callback_proc.c
@@ -354,47 +354,38 @@ out:
  * a single outstanding callback request at a time.
  */
 static __be32
-validate_seqid(struct nfs4_slot_table *tbl, struct cb_sequenceargs * args)
+validate_seqid(const struct nfs4_slot_table *tbl, const struct nfs4_slot *slot,
+		const struct cb_sequenceargs * args)
 {
-	struct nfs4_slot *slot;
-
-	dprintk("%s enter. slotid %u seqid %u\n",
-		__func__, args->csa_slotid, args->csa_sequenceid);
+	dprintk("%s enter. slotid %u seqid %u, slot table seqid: %u\n",
+		__func__, args->csa_slotid, args->csa_sequenceid, slot->seq_nr);
 
-	if (args->csa_slotid >= NFS41_BC_MAX_CALLBACKS)
+	if (args->csa_slotid > tbl->server_highest_slotid)
 		return htonl(NFS4ERR_BADSLOT);
 
-	slot = tbl->slots + args->csa_slotid;
-	dprintk("%s slot table seqid: %u\n", __func__, slot->seq_nr);
-
-	/* Normal */
-	if (likely(args->csa_sequenceid == slot->seq_nr + 1))
-		goto out_ok;
-
 	/* Replay */
 	if (args->csa_sequenceid == slot->seq_nr) {
 		dprintk("%s seqid %u is a replay\n",
 			__func__, args->csa_sequenceid);
+		if (nfs4_test_locked_slot(tbl, slot->slot_nr))
+			return htonl(NFS4ERR_DELAY);
 		/* Signal process_op to set this error on next op */
 		if (args->csa_cachethis == 0)
 			return htonl(NFS4ERR_RETRY_UNCACHED_REP);
 
-		/* The ca_maxresponsesize_cached is 0 with no DRC */
-		else if (args->csa_cachethis == 1)
-			return htonl(NFS4ERR_REP_TOO_BIG_TO_CACHE);
+		/* Liar! We never allowed you to set csa_cachethis != 0 */
+		return htonl(NFS4ERR_SEQ_FALSE_RETRY);
 	}
 
 	/* Wraparound */
-	if (args->csa_sequenceid == 1 && (slot->seq_nr + 1) == 0) {
-		slot->seq_nr = 1;
-		goto out_ok;
-	}
+	if (unlikely(slot->seq_nr == 0xFFFFFFFFU)) {
+		if (args->csa_sequenceid == 1)
+			return htonl(NFS4_OK);
+	} else if (likely(args->csa_sequenceid == slot->seq_nr + 1))
+		return htonl(NFS4_OK);
 
 	/* Misordered request */
 	return htonl(NFS4ERR_SEQ_MISORDERED);
-out_ok:
-	tbl->highest_used_slotid = args->csa_slotid;
-	return htonl(NFS4_OK);
 }
 
 /*
@@ -473,6 +464,12 @@ __be32 nfs4_callback_sequence(struct cb_sequenceargs *args,
 	tbl = &clp->cl_session->bc_slot_table;
 	slot = tbl->slots + args->csa_slotid;
 
+	/* Set up res before grabbing the spinlock */
+	memcpy(&res->csr_sessionid, &args->csa_sessionid,
+	       sizeof(res->csr_sessionid));
+	res->csr_sequenceid = args->csa_sequenceid;
+	res->csr_slotid = args->csa_slotid;
+
 	spin_lock(&tbl->slot_tbl_lock);
 	/* state manager is resetting the session */
 	if (test_bit(NFS4_SLOT_TBL_DRAINING, &tbl->slot_tbl_state)) {
@@ -485,18 +482,26 @@ __be32 nfs4_callback_sequence(struct cb_sequenceargs *args,
 		goto out_unlock;
 	}
 
-	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;
+	status = htonl(NFS4ERR_BADSLOT);
+	slot = nfs4_lookup_slot(tbl, args->csa_slotid);
+	if (IS_ERR(slot))
+		goto out_unlock;
+
+	res->csr_highestslotid = tbl->server_highest_slotid;
+	res->csr_target_highestslotid = tbl->target_highest_slotid;
 
-	status = validate_seqid(tbl, args);
+	status = validate_seqid(tbl, slot, args);
 	if (status)
 		goto out_unlock;
+	if (!nfs4_try_to_lock_slot(tbl, slot)) {
+		status = htonl(NFS4ERR_DELAY);
+		goto out_unlock;
+	}
+	cps->slot = slot;
 
-	cps->slotid = args->csa_slotid;
+	/* The ca_maxresponsesize_cached is 0 with no DRC */
+	if (args->csa_cachethis != 0)
+		return htonl(NFS4ERR_REP_TOO_BIG_TO_CACHE);
 
 	/*
 	 * Check for pending referring calls.  If a match is found, a
@@ -513,7 +518,7 @@ __be32 nfs4_callback_sequence(struct cb_sequenceargs *args,
 	 * If CB_SEQUENCE returns an error, then the state of the slot
 	 * (sequence ID, cached reply) MUST NOT change.
 	 */
-	slot->seq_nr++;
+	slot->seq_nr = args->csa_sequenceid;
 out_unlock:
 	spin_unlock(&tbl->slot_tbl_lock);
 
diff --git a/fs/nfs/callback_xdr.c b/fs/nfs/callback_xdr.c
index 646cdac..976c906 100644
--- a/fs/nfs/callback_xdr.c
+++ b/fs/nfs/callback_xdr.c
@@ -752,7 +752,8 @@ preprocess_nfs41_op(int nop, unsigned int op_nr, struct callback_op **op)
 	return htonl(NFS_OK);
 }
 
-static void nfs4_callback_free_slot(struct nfs4_session *session)
+static void nfs4_callback_free_slot(struct nfs4_session *session,
+		struct nfs4_slot *slot)
 {
 	struct nfs4_slot_table *tbl = &session->bc_slot_table;
 
@@ -761,15 +762,17 @@ static void nfs4_callback_free_slot(struct nfs4_session *session)
 	 * Let the state manager know callback processing done.
 	 * A single slot, so highest used slotid is either 0 or -1
 	 */
-	tbl->highest_used_slotid = NFS4_NO_SLOT;
+	nfs4_free_slot(tbl, slot);
 	nfs4_slot_tbl_drain_complete(tbl);
 	spin_unlock(&tbl->slot_tbl_lock);
 }
 
 static void nfs4_cb_free_slot(struct cb_process_state *cps)
 {
-	if (cps->slotid != NFS4_NO_SLOT)
-		nfs4_callback_free_slot(cps->clp->cl_session);
+	if (cps->slot) {
+		nfs4_callback_free_slot(cps->clp->cl_session, cps->slot);
+		cps->slot = NULL;
+	}
 }
 
 #else /* CONFIG_NFS_V4_1 */
@@ -893,7 +896,6 @@ static __be32 nfs4_callback_compound(struct svc_rqst *rqstp, void *argp, void *r
 	struct cb_process_state cps = {
 		.drc_status = 0,
 		.clp = NULL,
-		.slotid = NFS4_NO_SLOT,
 		.net = SVC_NET(rqstp),
 	};
 	unsigned int nops = 0;
diff --git a/fs/nfs/file.c b/fs/nfs/file.c
index 748bb81..89bf093 100644
--- a/fs/nfs/file.c
+++ b/fs/nfs/file.c
@@ -233,7 +233,7 @@ EXPORT_SYMBOL_GPL(nfs_file_mmap);
  * nfs_file_write() that a write error occurred, and hence cause it to
  * fall back to doing a synchronous write.
  */
-int
+static int
 nfs_file_fsync_commit(struct file *file, loff_t start, loff_t end, int datasync)
 {
 	struct nfs_open_context *ctx = nfs_file_open_context(file);
@@ -263,9 +263,8 @@ nfs_file_fsync_commit(struct file *file, loff_t start, loff_t end, int datasync)
 out:
 	return ret;
 }
-EXPORT_SYMBOL_GPL(nfs_file_fsync_commit);
 
-static int
+int
 nfs_file_fsync(struct file *file, loff_t start, loff_t end, int datasync)
 {
 	int ret;
@@ -273,13 +272,15 @@ nfs_file_fsync(struct file *file, loff_t start, loff_t end, int datasync)
 
 	trace_nfs_fsync_enter(inode);
 
-	nfs_inode_dio_wait(inode);
+	inode_dio_wait(inode);
 	do {
 		ret = filemap_write_and_wait_range(inode->i_mapping, start, end);
 		if (ret != 0)
 			break;
 		inode_lock(inode);
 		ret = nfs_file_fsync_commit(file, start, end, datasync);
+		if (!ret)
+			ret = pnfs_sync_inode(inode, !!datasync);
 		inode_unlock(inode);
 		/*
 		 * If nfs_file_fsync_commit detected a server reboot, then
@@ -293,6 +294,7 @@ nfs_file_fsync(struct file *file, loff_t start, loff_t end, int datasync)
 	trace_nfs_fsync_exit(inode, ret);
 	return ret;
 }
+EXPORT_SYMBOL_GPL(nfs_file_fsync);
 
 /*
  * Decide whether a read/modify/write cycle may be more efficient
@@ -368,7 +370,7 @@ start:
 	/*
 	 * Wait for O_DIRECT to complete
 	 */
-	nfs_inode_dio_wait(mapping->host);
+	inode_dio_wait(mapping->host);
 
 	page = grab_cache_page_write_begin(mapping, index, flags);
 	if (!page)
diff --git a/fs/nfs/flexfilelayout/flexfilelayoutdev.c b/fs/nfs/flexfilelayout/flexfilelayoutdev.c
index eb37046..add0e5a7 100644
--- a/fs/nfs/flexfilelayout/flexfilelayoutdev.c
+++ b/fs/nfs/flexfilelayout/flexfilelayoutdev.c
@@ -418,6 +418,8 @@ nfs4_ff_layout_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx,
 				pnfs_error_mark_layout_for_return(ino, lseg);
 		} else
 			pnfs_error_mark_layout_for_return(ino, lseg);
+		ds = NULL;
+		goto out;
 	}
 out_update_creds:
 	if (ff_layout_update_mirror_cred(mirror, ds))
diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c
index 86faecf..33d18c4 100644
--- a/fs/nfs/inode.c
+++ b/fs/nfs/inode.c
@@ -141,7 +141,7 @@ void nfs_evict_inode(struct inode *inode)
 
 int nfs_sync_inode(struct inode *inode)
 {
-	nfs_inode_dio_wait(inode);
+	inode_dio_wait(inode);
 	return nfs_wb_all(inode);
 }
 EXPORT_SYMBOL_GPL(nfs_sync_inode);
diff --git a/fs/nfs/internal.h b/fs/nfs/internal.h
index 9a547aa..565f813 100644
--- a/fs/nfs/internal.h
+++ b/fs/nfs/internal.h
@@ -358,7 +358,7 @@ int nfs_mknod(struct inode *, struct dentry *, umode_t, dev_t);
 int nfs_rename(struct inode *, struct dentry *, struct inode *, struct dentry *);
 
 /* file.c */
-int nfs_file_fsync_commit(struct file *, loff_t, loff_t, int);
+int nfs_file_fsync(struct file *file, loff_t start, loff_t end, int datasync);
 loff_t nfs_file_llseek(struct file *, loff_t, int);
 ssize_t nfs_file_read(struct kiocb *, struct iov_iter *);
 ssize_t nfs_file_splice_read(struct file *, loff_t *, struct pipe_inode_info *,
@@ -515,10 +515,6 @@ extern int nfs_sillyrename(struct inode *dir, struct dentry *dentry);
 /* direct.c */
 void nfs_init_cinfo_from_dreq(struct nfs_commit_info *cinfo,
 			      struct nfs_direct_req *dreq);
-static inline void nfs_inode_dio_wait(struct inode *inode)
-{
-	inode_dio_wait(inode);
-}
 extern ssize_t nfs_dreq_bytes_left(struct nfs_direct_req *dreq);
 
 /* nfs4proc.c */
diff --git a/fs/nfs/nfs4file.c b/fs/nfs/nfs4file.c
index 57ca1c8..22c35ab 100644
--- a/fs/nfs/nfs4file.c
+++ b/fs/nfs/nfs4file.c
@@ -128,37 +128,6 @@ nfs4_file_flush(struct file *file, fl_owner_t id)
 	return vfs_fsync(file, 0);
 }
 
-static int
-nfs4_file_fsync(struct file *file, loff_t start, loff_t end, int datasync)
-{
-	int ret;
-	struct inode *inode = file_inode(file);
-
-	trace_nfs_fsync_enter(inode);
-
-	nfs_inode_dio_wait(inode);
-	do {
-		ret = filemap_write_and_wait_range(inode->i_mapping, start, end);
-		if (ret != 0)
-			break;
-		inode_lock(inode);
-		ret = nfs_file_fsync_commit(file, start, end, datasync);
-		if (!ret)
-			ret = pnfs_sync_inode(inode, !!datasync);
-		inode_unlock(inode);
-		/*
-		 * If nfs_file_fsync_commit detected a server reboot, then
-		 * resend all dirty pages that might have been covered by
-		 * the NFS_CONTEXT_RESEND_WRITES flag
-		 */
-		start = 0;
-		end = LLONG_MAX;
-	} while (ret == -EAGAIN);
-
-	trace_nfs_fsync_exit(inode, ret);
-	return ret;
-}
-
 #ifdef CONFIG_NFS_V4_2
 static loff_t nfs4_file_llseek(struct file *filep, loff_t offset, int whence)
 {
@@ -266,7 +235,7 @@ const struct file_operations nfs4_file_operations = {
 	.open		= nfs4_file_open,
 	.flush		= nfs4_file_flush,
 	.release	= nfs_file_release,
-	.fsync		= nfs4_file_fsync,
+	.fsync		= nfs_file_fsync,
 	.lock		= nfs_lock,
 	.flock		= nfs_flock,
 	.splice_read	= nfs_file_splice_read,
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index 400a70b..327b8c3 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -6783,13 +6783,26 @@ nfs41_same_server_scope(struct nfs41_server_scope *a,
 	return false;
 }
 
+static void
+nfs4_bind_one_conn_to_session_done(struct rpc_task *task, void *calldata)
+{
+}
+
+static const struct rpc_call_ops nfs4_bind_one_conn_to_session_ops = {
+	.rpc_call_done =  &nfs4_bind_one_conn_to_session_done,
+};
+
 /*
- * nfs4_proc_bind_conn_to_session()
+ * nfs4_proc_bind_one_conn_to_session()
  *
  * The 4.1 client currently uses the same TCP connection for the
  * fore and backchannel.
  */
-int nfs4_proc_bind_conn_to_session(struct nfs_client *clp, struct rpc_cred *cred)
+static
+int nfs4_proc_bind_one_conn_to_session(struct rpc_clnt *clnt,
+		struct rpc_xprt *xprt,
+		struct nfs_client *clp,
+		struct rpc_cred *cred)
 {
 	int status;
 	struct nfs41_bind_conn_to_session_args args = {
@@ -6804,6 +6817,14 @@ int nfs4_proc_bind_conn_to_session(struct nfs_client *clp, struct rpc_cred *cred
 		.rpc_resp = &res,
 		.rpc_cred = cred,
 	};
+	struct rpc_task_setup task_setup_data = {
+		.rpc_client = clnt,
+		.rpc_xprt = xprt,
+		.callback_ops = &nfs4_bind_one_conn_to_session_ops,
+		.rpc_message = &msg,
+		.flags = RPC_TASK_TIMEOUT,
+	};
+	struct rpc_task *task;
 
 	dprintk("--> %s\n", __func__);
 
@@ -6811,7 +6832,16 @@ int nfs4_proc_bind_conn_to_session(struct nfs_client *clp, struct rpc_cred *cred
 	if (!(clp->cl_session->flags & SESSION4_BACK_CHAN))
 		args.dir = NFS4_CDFC4_FORE;
 
-	status = rpc_call_sync(clp->cl_rpcclient, &msg, RPC_TASK_TIMEOUT);
+	/* Do not set the backchannel flag unless this is clnt->cl_xprt */
+	if (xprt != rcu_access_pointer(clnt->cl_xprt))
+		args.dir = NFS4_CDFC4_FORE;
+
+	task = rpc_run_task(&task_setup_data);
+	if (!IS_ERR(task)) {
+		status = task->tk_status;
+		rpc_put_task(task);
+	} else
+		status = PTR_ERR(task);
 	trace_nfs4_bind_conn_to_session(clp, status);
 	if (status == 0) {
 		if (memcmp(res.sessionid.data,
@@ -6838,6 +6868,31 @@ out:
 	return status;
 }
 
+struct rpc_bind_conn_calldata {
+	struct nfs_client *clp;
+	struct rpc_cred *cred;
+};
+
+static int
+nfs4_proc_bind_conn_to_session_callback(struct rpc_clnt *clnt,
+		struct rpc_xprt *xprt,
+		void *calldata)
+{
+	struct rpc_bind_conn_calldata *p = calldata;
+
+	return nfs4_proc_bind_one_conn_to_session(clnt, xprt, p->clp, p->cred);
+}
+
+int nfs4_proc_bind_conn_to_session(struct nfs_client *clp, struct rpc_cred *cred)
+{
+	struct rpc_bind_conn_calldata data = {
+		.clp = clp,
+		.cred = cred,
+	};
+	return rpc_clnt_iterate_for_each_xprt(clp->cl_rpcclient,
+			nfs4_proc_bind_conn_to_session_callback, &data);
+}
+
 /*
  * Minimum set of SP4_MACH_CRED operations from RFC 5661 in the enforce map
  * and operations we'd like to see to enable certain features in the allow map
@@ -7320,7 +7375,7 @@ static void nfs4_init_channel_attrs(struct nfs41_create_session_args *args)
 	args->bc_attrs.max_resp_sz = PAGE_SIZE;
 	args->bc_attrs.max_resp_sz_cached = 0;
 	args->bc_attrs.max_ops = NFS4_MAX_BACK_CHANNEL_OPS;
-	args->bc_attrs.max_reqs = 1;
+	args->bc_attrs.max_reqs = NFS41_BC_MAX_CALLBACKS;
 
 	dprintk("%s: Back Channel : max_rqst_sz=%u max_resp_sz=%u "
 		"max_resp_sz_cached=%u max_ops=%u max_reqs=%u\n",
diff --git a/fs/nfs/nfs4session.c b/fs/nfs/nfs4session.c
index e23366e..332d06e 100644
--- a/fs/nfs/nfs4session.c
+++ b/fs/nfs/nfs4session.c
@@ -135,6 +135,43 @@ static struct nfs4_slot *nfs4_find_or_create_slot(struct nfs4_slot_table  *tbl,
 	return ERR_PTR(-ENOMEM);
 }
 
+static void nfs4_lock_slot(struct nfs4_slot_table *tbl,
+		struct nfs4_slot *slot)
+{
+	u32 slotid = slot->slot_nr;
+
+	__set_bit(slotid, tbl->used_slots);
+	if (slotid > tbl->highest_used_slotid ||
+	    tbl->highest_used_slotid == NFS4_NO_SLOT)
+		tbl->highest_used_slotid = slotid;
+	slot->generation = tbl->generation;
+}
+
+/*
+ * nfs4_try_to_lock_slot - Given a slot try to allocate it
+ *
+ * Note: must be called with the slot_tbl_lock held.
+ */
+bool nfs4_try_to_lock_slot(struct nfs4_slot_table *tbl, struct nfs4_slot *slot)
+{
+	if (nfs4_test_locked_slot(tbl, slot->slot_nr))
+		return false;
+	nfs4_lock_slot(tbl, slot);
+	return true;
+}
+
+/*
+ * nfs4_lookup_slot - Find a slot but don't allocate it
+ *
+ * Note: must be called with the slot_tbl_lock held.
+ */
+struct nfs4_slot *nfs4_lookup_slot(struct nfs4_slot_table *tbl, u32 slotid)
+{
+	if (slotid <= tbl->max_slotid)
+		return nfs4_find_or_create_slot(tbl, slotid, 1, GFP_NOWAIT);
+	return ERR_PTR(-E2BIG);
+}
+
 /*
  * nfs4_alloc_slot - efficiently look for a free slot
  *
@@ -153,18 +190,11 @@ struct nfs4_slot *nfs4_alloc_slot(struct nfs4_slot_table *tbl)
 		__func__, tbl->used_slots[0], tbl->highest_used_slotid,
 		tbl->max_slotid + 1);
 	slotid = find_first_zero_bit(tbl->used_slots, tbl->max_slotid + 1);
-	if (slotid > tbl->max_slotid)
-		goto out;
-	ret = nfs4_find_or_create_slot(tbl, slotid, 1, GFP_NOWAIT);
-	if (IS_ERR(ret))
-		goto out;
-	__set_bit(slotid, tbl->used_slots);
-	if (slotid > tbl->highest_used_slotid ||
-			tbl->highest_used_slotid == NFS4_NO_SLOT)
-		tbl->highest_used_slotid = slotid;
-	ret->generation = tbl->generation;
-
-out:
+	if (slotid <= tbl->max_slotid) {
+		ret = nfs4_find_or_create_slot(tbl, slotid, 1, GFP_NOWAIT);
+		if (!IS_ERR(ret))
+			nfs4_lock_slot(tbl, ret);
+	}
 	dprintk("<-- %s used_slots=%04lx highest_used=%u slotid=%u\n",
 		__func__, tbl->used_slots[0], tbl->highest_used_slotid,
 		!IS_ERR(ret) ? ret->slot_nr : NFS4_NO_SLOT);
diff --git a/fs/nfs/nfs4session.h b/fs/nfs/nfs4session.h
index e3ea2c5..5b51298 100644
--- a/fs/nfs/nfs4session.h
+++ b/fs/nfs/nfs4session.h
@@ -77,6 +77,8 @@ extern int nfs4_setup_slot_table(struct nfs4_slot_table *tbl,
 		unsigned int max_reqs, const char *queue);
 extern void nfs4_shutdown_slot_table(struct nfs4_slot_table *tbl);
 extern struct nfs4_slot *nfs4_alloc_slot(struct nfs4_slot_table *tbl);
+extern struct nfs4_slot *nfs4_lookup_slot(struct nfs4_slot_table *tbl, u32 slotid);
+extern bool nfs4_try_to_lock_slot(struct nfs4_slot_table *tbl, struct nfs4_slot *slot);
 extern void nfs4_free_slot(struct nfs4_slot_table *tbl, struct nfs4_slot *slot);
 extern void nfs4_slot_tbl_drain_complete(struct nfs4_slot_table *tbl);
 bool nfs41_wake_and_assign_slot(struct nfs4_slot_table *tbl,
@@ -88,6 +90,12 @@ static inline bool nfs4_slot_tbl_draining(struct nfs4_slot_table *tbl)
 	return !!test_bit(NFS4_SLOT_TBL_DRAINING, &tbl->slot_tbl_state);
 }
 
+static inline bool nfs4_test_locked_slot(const struct nfs4_slot_table *tbl,
+		u32 slotid)
+{
+	return !!test_bit(slotid, tbl->used_slots);
+}
+
 #if defined(CONFIG_NFS_V4_1)
 extern void nfs41_set_target_slotid(struct nfs4_slot_table *tbl,
 		u32 target_highest_slotid);
diff --git a/fs/nfs/pnfs_nfs.c b/fs/nfs/pnfs_nfs.c
index 81ac648..4aaed89 100644
--- a/fs/nfs/pnfs_nfs.c
+++ b/fs/nfs/pnfs_nfs.c
@@ -606,12 +606,22 @@ static int _nfs4_pnfs_v3_ds_connect(struct nfs_server *mds_srv,
 		dprintk("%s: DS %s: trying address %s\n",
 			__func__, ds->ds_remotestr, da->da_remotestr);
 
-		clp = get_v3_ds_connect(mds_srv->nfs_client,
+		if (!IS_ERR(clp)) {
+			struct xprt_create xprt_args = {
+				.ident = XPRT_TRANSPORT_TCP,
+				.net = clp->cl_net,
+				.dstaddr = (struct sockaddr *)&da->da_addr,
+				.addrlen = da->da_addrlen,
+				.servername = clp->cl_hostname,
+			};
+			/* Add this address as an alias */
+			rpc_clnt_add_xprt(clp->cl_rpcclient, &xprt_args,
+					rpc_clnt_test_and_add_xprt, NULL);
+		} else
+			clp = get_v3_ds_connect(mds_srv->nfs_client,
 					(struct sockaddr *)&da->da_addr,
 					da->da_addrlen, IPPROTO_TCP,
 					timeo, retrans, au_flavor);
-		if (!IS_ERR(clp))
-			break;
 	}
 
 	if (IS_ERR(clp)) {