Merge branch 'nfs-for-2.6.39' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6

* 'nfs-for-2.6.39' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6: (54 commits)
  RPC: killing RPC tasks races fixed
  xprt: remove redundant check
  SUNRPC: Convert struct rpc_xprt to use atomic_t counters
  SUNRPC: Ensure we always run the tk_callback before tk_action
  sunrpc: fix printk format warning
  xprt: remove redundant null check
  nfs: BKL is no longer needed, so remove the include
  NFS: Fix a warning in fs/nfs/idmap.c
  Cleanup: Factor out some cut-and-paste code.
  cleanup: save 60 lines/100 bytes by combining two mostly duplicate functions.
  NFS: account direct-io into task io accounting
  gss:krb5 only include enctype numbers in gm_upcall_enctypes
  RPCRDMA: Fix FRMR registration/invalidate handling.
  RPCRDMA: Fix to XDR page base interpretation in marshalling logic.
  NFSv4: Send unmapped uid/gids to the server when using auth_sys
  NFSv4: Propagate the error NFS4ERR_BADOWNER to nfs4_do_setattr
  NFSv4: cleanup idmapper functions to take an nfs_server argument
  NFSv4: Send unmapped uid/gids to the server if the idmapper fails
  NFSv4: If the server sends us a numeric uid/gid then accept it
  NFSv4.1: reject zero layout with zeroed stripe unit
  ...

Documentation/filesystems/nfs/pnfs.txt

@@ -46,3 +46,10 @@ data server cache
 file driver devices refer to data servers, which are kept in a module
 level cache.  Its reference is held over the lifetime of the deviceid
 pointing to it.
+
+lseg
+----
+lseg maintains an extra reference corresponding to the NFS_LSEG_VALID
+bit which holds it in the pnfs_layout_hdr's list.  When the final lseg
+is removed from the pnfs_layout_hdr's list, the NFS_LAYOUT_DESTROYED
+bit is set, preventing any new lsegs from being added.

Documentation/kernel-parameters.txt

@@ -1580,6 +1580,14 @@ bytes respectively. Such letter suffixes can also be entirely omitted.
 			of returning the full 64-bit number.
 			The default is to return 64-bit inode numbers.
 
+	nfs.nfs4_disable_idmapping=
+			[NFSv4] When set, this option disables the NFSv4
+			idmapper on the client, but only if the mount
+			is using the 'sec=sys' security flavour. This may
+			make migration from legacy NFSv2/v3 systems easier
+			provided that the server has the appropriate support.
+			The default is to always enable NFSv4 idmapping.
+
 	nmi_debug=	[KNL,AVR32,SH] Specify one or more actions to take
 			when a NMI is triggered.
 			Format: [state][,regs][,debounce][,die]

fs/nfs/callback_proc.c

@@ -188,10 +188,10 @@ static u32 initiate_bulk_draining(struct nfs_client *clp,
 			rv = NFS4ERR_DELAY;
 		list_del_init(&lo->plh_bulk_recall);
 		spin_unlock(&ino->i_lock);
+		pnfs_free_lseg_list(&free_me_list);
 		put_layout_hdr(lo);
 		iput(ino);
 	}
-	pnfs_free_lseg_list(&free_me_list);
 	return rv;
 }
 

fs/nfs/client.c

@@ -81,6 +81,11 @@ retry:
 }
 #endif /* CONFIG_NFS_V4 */
 
+/*
+ * Turn off NFSv4 uid/gid mapping when using AUTH_SYS
+ */
+static int nfs4_disable_idmapping = 0;
+
 /*
  * RPC cruft for NFS
  */
@@ -481,7 +486,12 @@ static struct nfs_client *nfs_match_client(const struct nfs_client_initdata *dat
  * Look up a client by IP address and protocol version
  * - creates a new record if one doesn't yet exist
  */
-static struct nfs_client *nfs_get_client(const struct nfs_client_initdata *cl_init)
+static struct nfs_client *
+nfs_get_client(const struct nfs_client_initdata *cl_init,
+	       const struct rpc_timeout *timeparms,
+	       const char *ip_addr,
+	       rpc_authflavor_t authflavour,
+	       int noresvport)
 {
 	struct nfs_client *clp, *new = NULL;
 	int error;
@@ -512,6 +522,13 @@ install_client:
 	clp = new;
 	list_add(&clp->cl_share_link, &nfs_client_list);
 	spin_unlock(&nfs_client_lock);
+
+	error = cl_init->rpc_ops->init_client(clp, timeparms, ip_addr,
+					      authflavour, noresvport);
+	if (error < 0) {
+		nfs_put_client(clp);
+		return ERR_PTR(error);
+	}
 	dprintk("--> nfs_get_client() = %p [new]\n", clp);
 	return clp;
 
@@ -767,9 +784,9 @@ static int nfs_init_server_rpcclient(struct nfs_server *server,
 /*
  * Initialise an NFS2 or NFS3 client
  */
-static int nfs_init_client(struct nfs_client *clp,
-			   const struct rpc_timeout *timeparms,
-			   const struct nfs_parsed_mount_data *data)
+int nfs_init_client(struct nfs_client *clp, const struct rpc_timeout *timeparms,
+		    const char *ip_addr, rpc_authflavor_t authflavour,
+		    int noresvport)
 {
 	int error;
 
@@ -784,7 +801,7 @@ static int nfs_init_client(struct nfs_client *clp,
 	 * - RFC 2623, sec 2.3.2
 	 */
 	error = nfs_create_rpc_client(clp, timeparms, RPC_AUTH_UNIX,
-				      0, data->flags & NFS_MOUNT_NORESVPORT);
+				      0, noresvport);
 	if (error < 0)
 		goto error;
 	nfs_mark_client_ready(clp, NFS_CS_READY);
@@ -820,19 +837,17 @@ static int nfs_init_server(struct nfs_server *server,
 		cl_init.rpc_ops = &nfs_v3_clientops;
 #endif
 
+	nfs_init_timeout_values(&timeparms, data->nfs_server.protocol,
+			data->timeo, data->retrans);
+
 	/* Allocate or find a client reference we can use */
-	clp = nfs_get_client(&cl_init);
+	clp = nfs_get_client(&cl_init, &timeparms, NULL, RPC_AUTH_UNIX,
+			     data->flags & NFS_MOUNT_NORESVPORT);
 	if (IS_ERR(clp)) {
 		dprintk("<-- nfs_init_server() = error %ld\n", PTR_ERR(clp));
 		return PTR_ERR(clp);
 	}
 
-	nfs_init_timeout_values(&timeparms, data->nfs_server.protocol,
-			data->timeo, data->retrans);
-	error = nfs_init_client(clp, &timeparms, data);
-	if (error < 0)
-		goto error;
-
 	server->nfs_client = clp;
 
 	/* Initialise the client representation from the mount data */
@@ -1009,14 +1024,19 @@ static void nfs_server_insert_lists(struct nfs_server *server)
 	spin_lock(&nfs_client_lock);
 	list_add_tail_rcu(&server->client_link, &clp->cl_superblocks);
 	list_add_tail(&server->master_link, &nfs_volume_list);
+	clear_bit(NFS_CS_STOP_RENEW, &clp->cl_res_state);
 	spin_unlock(&nfs_client_lock);
 
 }
 
 static void nfs_server_remove_lists(struct nfs_server *server)
 {
+	struct nfs_client *clp = server->nfs_client;
+
 	spin_lock(&nfs_client_lock);
 	list_del_rcu(&server->client_link);
+	if (clp && list_empty(&clp->cl_superblocks))
+		set_bit(NFS_CS_STOP_RENEW, &clp->cl_res_state);
 	list_del(&server->master_link);
 	spin_unlock(&nfs_client_lock);
 
@@ -1307,11 +1327,11 @@ static int nfs4_init_client_minor_version(struct nfs_client *clp)
 /*
  * Initialise an NFS4 client record
  */
-static int nfs4_init_client(struct nfs_client *clp,
-		const struct rpc_timeout *timeparms,
-		const char *ip_addr,
-		rpc_authflavor_t authflavour,
-		int flags)
+int nfs4_init_client(struct nfs_client *clp,
+		     const struct rpc_timeout *timeparms,
+		     const char *ip_addr,
+		     rpc_authflavor_t authflavour,
+		     int noresvport)
 {
 	int error;
 
@@ -1325,7 +1345,7 @@ static int nfs4_init_client(struct nfs_client *clp,
 	clp->rpc_ops = &nfs_v4_clientops;
 
 	error = nfs_create_rpc_client(clp, timeparms, authflavour,
-				      1, flags & NFS_MOUNT_NORESVPORT);
+				      1, noresvport);
 	if (error < 0)
 		goto error;
 	strlcpy(clp->cl_ipaddr, ip_addr, sizeof(clp->cl_ipaddr));
@@ -1378,27 +1398,71 @@ static int nfs4_set_client(struct nfs_server *server,
 	dprintk("--> nfs4_set_client()\n");
 
 	/* Allocate or find a client reference we can use */
-	clp = nfs_get_client(&cl_init);
+	clp = nfs_get_client(&cl_init, timeparms, ip_addr, authflavour,
+			     server->flags & NFS_MOUNT_NORESVPORT);
 	if (IS_ERR(clp)) {
 		error = PTR_ERR(clp);
 		goto error;
 	}
-	error = nfs4_init_client(clp, timeparms, ip_addr, authflavour,
-					server->flags);
-	if (error < 0)
-		goto error_put;
+
+	/*
+	 * Query for the lease time on clientid setup or renewal
+	 *
+	 * Note that this will be set on nfs_clients that were created
+	 * only for the DS role and did not set this bit, but now will
+	 * serve a dual role.
+	 */
+	set_bit(NFS_CS_CHECK_LEASE_TIME, &clp->cl_res_state);
 
 	server->nfs_client = clp;
 	dprintk("<-- nfs4_set_client() = 0 [new %p]\n", clp);
 	return 0;
-
-error_put:
-	nfs_put_client(clp);
 error:
 	dprintk("<-- nfs4_set_client() = xerror %d\n", error);
 	return error;
 }
 
+/*
+ * Set up a pNFS Data Server client.
+ *
+ * Return any existing nfs_client that matches server address,port,version
+ * and minorversion.
+ *
+ * For a new nfs_client, use a soft mount (default), a low retrans and a
+ * low timeout interval so that if a connection is lost, we retry through
+ * the MDS.
+ */
+struct nfs_client *nfs4_set_ds_client(struct nfs_client* mds_clp,
+		const struct sockaddr *ds_addr,
+		int ds_addrlen, int ds_proto)
+{
+	struct nfs_client_initdata cl_init = {
+		.addr = ds_addr,
+		.addrlen = ds_addrlen,
+		.rpc_ops = &nfs_v4_clientops,
+		.proto = ds_proto,
+		.minorversion = mds_clp->cl_minorversion,
+	};
+	struct rpc_timeout ds_timeout = {
+		.to_initval = 15 * HZ,
+		.to_maxval = 15 * HZ,
+		.to_retries = 1,
+		.to_exponential = 1,
+	};
+	struct nfs_client *clp;
+
+	/*
+	 * Set an authflavor equual to the MDS value. Use the MDS nfs_client
+	 * cl_ipaddr so as to use the same EXCHANGE_ID co_ownerid as the MDS
+	 * (section 13.1 RFC 5661).
+	 */
+	clp = nfs_get_client(&cl_init, &ds_timeout, mds_clp->cl_ipaddr,
+			     mds_clp->cl_rpcclient->cl_auth->au_flavor, 0);
+
+	dprintk("<-- %s %p\n", __func__, clp);
+	return clp;
+}
+EXPORT_SYMBOL(nfs4_set_ds_client);
 
 /*
  * Session has been established, and the client marked ready.
@@ -1435,6 +1499,10 @@ static int nfs4_server_common_setup(struct nfs_server *server,
 	BUG_ON(!server->nfs_client->rpc_ops);
 	BUG_ON(!server->nfs_client->rpc_ops->file_inode_ops);
 
+	/* data servers support only a subset of NFSv4.1 */
+	if (is_ds_only_client(server->nfs_client))
+		return -EPROTONOSUPPORT;
+
 	fattr = nfs_alloc_fattr();
 	if (fattr == NULL)
 		return -ENOMEM;
@@ -1504,6 +1572,13 @@ static int nfs4_init_server(struct nfs_server *server,
 	if (error < 0)
 		goto error;
 
+	/*
+	 * Don't use NFS uid/gid mapping if we're using AUTH_SYS or lower
+	 * authentication.
+	 */
+	if (nfs4_disable_idmapping && data->auth_flavors[0] == RPC_AUTH_UNIX)
+		server->caps |= NFS_CAP_UIDGID_NOMAP;
+
 	if (data->rsize)
 		server->rsize = nfs_block_size(data->rsize, NULL);
 	if (data->wsize)
@@ -1921,3 +1996,7 @@ void nfs_fs_proc_exit(void)
 }
 
 #endif /* CONFIG_PROC_FS */
+
+module_param(nfs4_disable_idmapping, bool, 0644);
+MODULE_PARM_DESC(nfs4_disable_idmapping,
+		"Turn off NFSv4 idmapping when using 'sec=sys'");

fs/nfs/direct.c

@@ -45,6 +45,7 @@
 #include <linux/pagemap.h>
 #include <linux/kref.h>
 #include <linux/slab.h>
+#include <linux/task_io_accounting_ops.h>
 
 #include <linux/nfs_fs.h>
 #include <linux/nfs_page.h>
@@ -649,8 +650,7 @@ static void nfs_direct_write_result(struct rpc_task *task, void *calldata)
 {
 	struct nfs_write_data *data = calldata;
 
-	if (nfs_writeback_done(task, data) != 0)
-		return;
+	nfs_writeback_done(task, data);
 }
 
 /*
@@ -938,6 +938,8 @@ ssize_t nfs_file_direct_read(struct kiocb *iocb, const struct iovec *iov,
 	if (retval)
 		goto out;
 
+	task_io_account_read(count);
+
 	retval = nfs_direct_read(iocb, iov, nr_segs, pos);
 	if (retval > 0)
 		iocb->ki_pos = pos + retval;
@@ -999,6 +1001,8 @@ ssize_t nfs_file_direct_write(struct kiocb *iocb, const struct iovec *iov,
 	if (retval)
 		goto out;
 
+	task_io_account_write(count);
+
 	retval = nfs_direct_write(iocb, iov, nr_segs, pos, count);
 
 	if (retval > 0)

fs/nfs/file.c

@@ -387,10 +387,6 @@ static int nfs_write_begin(struct file *file, struct address_space *mapping,
 		file->f_path.dentry->d_name.name,
 		mapping->host->i_ino, len, (long long) pos);
 
-	pnfs_update_layout(mapping->host,
-			   nfs_file_open_context(file),
-			   IOMODE_RW);
-
 start:
 	/*
 	 * Prevent starvation issues if someone is doing a consistency

fs/nfs/idmap.c

@@ -33,16 +33,41 @@
  *  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/types.h>
+#include <linux/string.h>
+#include <linux/kernel.h>
+
+static int nfs_map_string_to_numeric(const char *name, size_t namelen, __u32 *res)
+{
+	unsigned long val;
+	char buf[16];
+
+	if (memchr(name, '@', namelen) != NULL || namelen >= sizeof(buf))
+		return 0;
+	memcpy(buf, name, namelen);
+	buf[namelen] = '\0';
+	if (strict_strtoul(buf, 0, &val) != 0)
+		return 0;
+	*res = val;
+	return 1;
+}
+
+static int nfs_map_numeric_to_string(__u32 id, char *buf, size_t buflen)
+{
+	return snprintf(buf, buflen, "%u", id);
+}
 
 #ifdef CONFIG_NFS_USE_NEW_IDMAPPER
 
 #include <linux/slab.h>
 #include <linux/cred.h>
+#include <linux/sunrpc/sched.h>
+#include <linux/nfs4.h>
+#include <linux/nfs_fs_sb.h>
 #include <linux/nfs_idmap.h>
 #include <linux/keyctl.h>
 #include <linux/key-type.h>
 #include <linux/rcupdate.h>
-#include <linux/kernel.h>
 #include <linux/err.h>
 
 #include <keys/user-type.h>
@@ -219,23 +244,39 @@ static int nfs_idmap_lookup_id(const char *name, size_t namelen,
 	return ret;
 }
 
-int nfs_map_name_to_uid(struct nfs_client *clp, const char *name, size_t namelen, __u32 *uid)
+int nfs_map_name_to_uid(const struct nfs_server *server, const char *name, size_t namelen, __u32 *uid)
 {
+	if (nfs_map_string_to_numeric(name, namelen, uid))
+		return 0;
 	return nfs_idmap_lookup_id(name, namelen, "uid", uid);
 }
 
-int nfs_map_group_to_gid(struct nfs_client *clp, const char *name, size_t namelen, __u32 *gid)
+int nfs_map_group_to_gid(const struct nfs_server *server, const char *name, size_t namelen, __u32 *gid)
 {
+	if (nfs_map_string_to_numeric(name, namelen, gid))
+		return 0;
 	return nfs_idmap_lookup_id(name, namelen, "gid", gid);
 }
 
-int nfs_map_uid_to_name(struct nfs_client *clp, __u32 uid, char *buf, size_t buflen)
+int nfs_map_uid_to_name(const struct nfs_server *server, __u32 uid, char *buf, size_t buflen)
 {
-	return nfs_idmap_lookup_name(uid, "user", buf, buflen);
+	int ret = -EINVAL;
+
+	if (!(server->caps & NFS_CAP_UIDGID_NOMAP))
+		ret = nfs_idmap_lookup_name(uid, "user", buf, buflen);
+	if (ret < 0)
+		ret = nfs_map_numeric_to_string(uid, buf, buflen);
+	return ret;
 }
-int nfs_map_gid_to_group(struct nfs_client *clp, __u32 gid, char *buf, size_t buflen)
+int nfs_map_gid_to_group(const struct nfs_server *server, __u32 gid, char *buf, size_t buflen)
 {
-	return nfs_idmap_lookup_name(gid, "group", buf, buflen);
+	int ret = -EINVAL;
+
+	if (!(server->caps & NFS_CAP_UIDGID_NOMAP))
+		ret = nfs_idmap_lookup_name(gid, "group", buf, buflen);
+	if (ret < 0)
+		ret = nfs_map_numeric_to_string(gid, buf, buflen);
+	return ret;
 }
 
 #else  /* CONFIG_NFS_USE_NEW_IDMAPPER not defined */
@@ -243,7 +284,6 @@ int nfs_map_gid_to_group(struct nfs_client *clp, __u32 gid, char *buf, size_t bu
 #include <linux/module.h>
 #include <linux/mutex.h>
 #include <linux/init.h>
-#include <linux/types.h>
 #include <linux/slab.h>
 #include <linux/socket.h>
 #include <linux/in.h>
@@ -695,31 +735,45 @@ static unsigned int fnvhash32(const void *buf, size_t buflen)
 	return hash;
 }
 
-int nfs_map_name_to_uid(struct nfs_client *clp, const char *name, size_t namelen, __u32 *uid)
+int nfs_map_name_to_uid(const struct nfs_server *server, const char *name, size_t namelen, __u32 *uid)
 {
-	struct idmap *idmap = clp->cl_idmap;
+	struct idmap *idmap = server->nfs_client->cl_idmap;
 
+	if (nfs_map_string_to_numeric(name, namelen, uid))
+		return 0;
 	return nfs_idmap_id(idmap, &idmap->idmap_user_hash, name, namelen, uid);
 }
 
-int nfs_map_group_to_gid(struct nfs_client *clp, const char *name, size_t namelen, __u32 *uid)
+int nfs_map_group_to_gid(const struct nfs_server *server, const char *name, size_t namelen, __u32 *uid)
 {
-	struct idmap *idmap = clp->cl_idmap;
+	struct idmap *idmap = server->nfs_client->cl_idmap;
 
+	if (nfs_map_string_to_numeric(name, namelen, uid))
+		return 0;
 	return nfs_idmap_id(idmap, &idmap->idmap_group_hash, name, namelen, uid);
 }
 
-int nfs_map_uid_to_name(struct nfs_client *clp, __u32 uid, char *buf, size_t buflen)
+int nfs_map_uid_to_name(const struct nfs_server *server, __u32 uid, char *buf, size_t buflen)
 {
-	struct idmap *idmap = clp->cl_idmap;
+	struct idmap *idmap = server->nfs_client->cl_idmap;
+	int ret = -EINVAL;
 
-	return nfs_idmap_name(idmap, &idmap->idmap_user_hash, uid, buf);
+	if (!(server->caps & NFS_CAP_UIDGID_NOMAP))
+		ret = nfs_idmap_name(idmap, &idmap->idmap_user_hash, uid, buf);
+	if (ret < 0)
+		ret = nfs_map_numeric_to_string(uid, buf, buflen);
+	return ret;
 }
-int nfs_map_gid_to_group(struct nfs_client *clp, __u32 uid, char *buf, size_t buflen)
+int nfs_map_gid_to_group(const struct nfs_server *server, __u32 uid, char *buf, size_t buflen)
 {
-	struct idmap *idmap = clp->cl_idmap;
+	struct idmap *idmap = server->nfs_client->cl_idmap;
+	int ret = -EINVAL;
 
-	return nfs_idmap_name(idmap, &idmap->idmap_group_hash, uid, buf);
+	if (!(server->caps & NFS_CAP_UIDGID_NOMAP))
+		ret = nfs_idmap_name(idmap, &idmap->idmap_group_hash, uid, buf);
+	if (ret < 0)
+		ret = nfs_map_numeric_to_string(uid, buf, buflen);
+	return ret;
 }
 
 #endif /* CONFIG_NFS_USE_NEW_IDMAPPER */

fs/nfs/internal.h

@@ -148,6 +148,9 @@ extern struct nfs_server *nfs_clone_server(struct nfs_server *,
 					   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);
+extern struct nfs_client *nfs4_set_ds_client(struct nfs_client* mds_clp,
+					     const struct sockaddr *ds_addr,
+					     int ds_addrlen, int ds_proto);
 #ifdef CONFIG_PROC_FS
 extern int __init nfs_fs_proc_init(void);
 extern void nfs_fs_proc_exit(void);
@@ -213,8 +216,14 @@ extern const u32 nfs41_maxwrite_overhead;
 extern struct rpc_procinfo nfs4_procedures[];
 #endif
 
+extern int nfs4_init_ds_session(struct nfs_client *clp);
+
 /* proc.c */
 void nfs_close_context(struct nfs_open_context *ctx, int is_sync);
+extern int nfs_init_client(struct nfs_client *clp,
+			   const struct rpc_timeout *timeparms,
+			   const char *ip_addr, rpc_authflavor_t authflavour,
+			   int noresvport);
 
 /* dir.c */
 extern int nfs_access_cache_shrinker(struct shrinker *shrink,
@@ -262,9 +271,15 @@ extern int nfs4_get_rootfh(struct nfs_server *server, struct nfs_fh *mntfh);
 #endif
 
 /* read.c */
+extern int nfs_initiate_read(struct nfs_read_data *data, struct rpc_clnt *clnt,
+			     const struct rpc_call_ops *call_ops);
 extern void nfs_read_prepare(struct rpc_task *task, void *calldata);
 
 /* write.c */
+extern int nfs_initiate_write(struct nfs_write_data *data,
+			      struct rpc_clnt *clnt,
+			      const struct rpc_call_ops *call_ops,
+			      int how);
 extern void nfs_write_prepare(struct rpc_task *task, void *calldata);
 #ifdef CONFIG_MIGRATION
 extern int nfs_migrate_page(struct address_space *,
@@ -274,6 +289,13 @@ extern int nfs_migrate_page(struct address_space *,
 #endif
 
 /* nfs4proc.c */
+extern void nfs4_reset_read(struct rpc_task *task, struct nfs_read_data *data);
+extern int nfs4_init_client(struct nfs_client *clp,
+			    const struct rpc_timeout *timeparms,
+			    const char *ip_addr,
+			    rpc_authflavor_t authflavour,
+			    int noresvport);
+extern void nfs4_reset_write(struct rpc_task *task, struct nfs_write_data *data);
 extern int _nfs4_call_sync(struct nfs_server *server,
 			   struct rpc_message *msg,
 			   struct nfs4_sequence_args *args,

fs/nfs/nfs3proc.c

@@ -885,4 +885,5 @@ const struct nfs_rpc_ops nfs_v3_clientops = {
 	.lock		= nfs3_proc_lock,
 	.clear_acl_cache = nfs3_forget_cached_acls,
 	.close_context	= nfs_close_context,
+	.init_client	= nfs_init_client,
 };

fs/nfs/nfs4_fs.h

@@ -252,6 +252,9 @@ static inline struct nfs4_session *nfs4_get_session(const struct nfs_server *ser
 extern int nfs4_setup_sequence(const struct nfs_server *server,
 		struct nfs4_sequence_args *args, struct nfs4_sequence_res *res,
 		int cache_reply, struct rpc_task *task);
+extern int nfs41_setup_sequence(struct nfs4_session *session,
+		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 *);
@@ -259,6 +262,19 @@ extern int nfs4_proc_destroy_session(struct nfs4_session *);
 extern int nfs4_init_session(struct nfs_server *server);
 extern int nfs4_proc_get_lease_time(struct nfs_client *clp,
 		struct nfs_fsinfo *fsinfo);
+
+static inline bool
+is_ds_only_client(struct nfs_client *clp)
+{
+	return (clp->cl_exchange_flags & EXCHGID4_FLAG_MASK_PNFS) ==
+		EXCHGID4_FLAG_USE_PNFS_DS;
+}
+
+static inline bool
+is_ds_client(struct nfs_client *clp)
+{
+	return clp->cl_exchange_flags & EXCHGID4_FLAG_USE_PNFS_DS;
+}
 #else /* CONFIG_NFS_v4_1 */
 static inline struct nfs4_session *nfs4_get_session(const struct nfs_server *server)
 {
@@ -276,6 +292,18 @@ static inline int nfs4_init_session(struct nfs_server *server)
 {
 	return 0;
 }
+
+static inline bool
+is_ds_only_client(struct nfs_client *clp)
+{
+	return false;
+}
+
+static inline bool
+is_ds_client(struct nfs_client *clp)
+{
+	return false;
+}
 #endif /* CONFIG_NFS_V4_1 */
 
 extern const struct nfs4_minor_version_ops *nfs_v4_minor_ops[];

fs/nfs/nfs4filelayout.c

@@ -40,32 +40,309 @@ MODULE_LICENSE("GPL");
 MODULE_AUTHOR("Dean Hildebrand <dhildebz@umich.edu>");
 MODULE_DESCRIPTION("The NFSv4 file layout driver");
 
-static int
-filelayout_set_layoutdriver(struct nfs_server *nfss)
+#define FILELAYOUT_POLL_RETRY_MAX     (15*HZ)
+
+static loff_t
+filelayout_get_dense_offset(struct nfs4_filelayout_segment *flseg,
+			    loff_t offset)
 {
-	int status = pnfs_alloc_init_deviceid_cache(nfss->nfs_client,
-						nfs4_fl_free_deviceid_callback);
-	if (status) {
-		printk(KERN_WARNING "%s: deviceid cache could not be "
-			"initialized\n", __func__);
-		return status;
+	u32 stripe_width = flseg->stripe_unit * flseg->dsaddr->stripe_count;
+	u64 tmp;
+
+	offset -= flseg->pattern_offset;
+	tmp = offset;
+	do_div(tmp, stripe_width);
+
+	return tmp * flseg->stripe_unit + do_div(offset, flseg->stripe_unit);
+}
+
+/* This function is used by the layout driver to calculate the
+ * offset of the file on the dserver based on whether the
+ * layout type is STRIPE_DENSE or STRIPE_SPARSE
+ */
+static loff_t
+filelayout_get_dserver_offset(struct pnfs_layout_segment *lseg, loff_t offset)
+{
+	struct nfs4_filelayout_segment *flseg = FILELAYOUT_LSEG(lseg);
+
+	switch (flseg->stripe_type) {
+	case STRIPE_SPARSE:
+		return offset;
+
+	case STRIPE_DENSE:
+		return filelayout_get_dense_offset(flseg, offset);
 	}
-	dprintk("%s: deviceid cache has been initialized successfully\n",
-		__func__);
+
+	BUG();
+}
+
+/* For data server errors we don't recover from */
+static void
+filelayout_set_lo_fail(struct pnfs_layout_segment *lseg)
+{
+	if (lseg->pls_range.iomode == IOMODE_RW) {
+		dprintk("%s Setting layout IOMODE_RW fail bit\n", __func__);
+		set_bit(lo_fail_bit(IOMODE_RW), &lseg->pls_layout->plh_flags);
+	} else {
+		dprintk("%s Setting layout IOMODE_READ fail bit\n", __func__);
+		set_bit(lo_fail_bit(IOMODE_READ), &lseg->pls_layout->plh_flags);
+	}
+}
+
+static int filelayout_async_handle_error(struct rpc_task *task,
+					 struct nfs4_state *state,
+					 struct nfs_client *clp,
+					 int *reset)
+{
+	if (task->tk_status >= 0)
+		return 0;
+
+	*reset = 0;
+
+	switch (task->tk_status) {
+	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:
+		dprintk("%s ERROR %d, Reset session. Exchangeid "
+			"flags 0x%x\n", __func__, task->tk_status,
+			clp->cl_exchange_flags);
+		nfs4_schedule_session_recovery(clp->cl_session);
+		break;
+	case -NFS4ERR_DELAY:
+	case -NFS4ERR_GRACE:
+	case -EKEYEXPIRED:
+		rpc_delay(task, FILELAYOUT_POLL_RETRY_MAX);
+		break;
+	default:
+		dprintk("%s DS error. Retry through MDS %d\n", __func__,
+			task->tk_status);
+		*reset = 1;
+		break;
+	}
+	task->tk_status = 0;
+	return -EAGAIN;
+}
+
+/* NFS_PROTO call done callback routines */
+
+static int filelayout_read_done_cb(struct rpc_task *task,
+				struct nfs_read_data *data)
+{
+	struct nfs_client *clp = data->ds_clp;
+	int reset = 0;
+
+	dprintk("%s DS read\n", __func__);
+
+	if (filelayout_async_handle_error(task, data->args.context->state,
+					  data->ds_clp, &reset) == -EAGAIN) {
+		dprintk("%s calling restart ds_clp %p ds_clp->cl_session %p\n",
+			__func__, data->ds_clp, data->ds_clp->cl_session);
+		if (reset) {
+			filelayout_set_lo_fail(data->lseg);
+			nfs4_reset_read(task, data);
+			clp = NFS_SERVER(data->inode)->nfs_client;
+		}
+		nfs_restart_rpc(task, clp);
+		return -EAGAIN;
+	}
+
 	return 0;
 }
 
-/* Clear out the layout by destroying its device list */
-static int
-filelayout_clear_layoutdriver(struct nfs_server *nfss)
+/*
+ * Call ops for the async read/write cases
+ * In the case of dense layouts, the offset needs to be reset to its
+ * original value.
+ */
+static void filelayout_read_prepare(struct rpc_task *task, void *data)
 {
-	dprintk("--> %s\n", __func__);
+	struct nfs_read_data *rdata = (struct nfs_read_data *)data;
+
+	rdata->read_done_cb = filelayout_read_done_cb;
+
+	if (nfs41_setup_sequence(rdata->ds_clp->cl_session,
+				&rdata->args.seq_args, &rdata->res.seq_res,
+				0, task))
+		return;
+
+	rpc_call_start(task);
+}
+
+static void filelayout_read_call_done(struct rpc_task *task, void *data)
+{
+	struct nfs_read_data *rdata = (struct nfs_read_data *)data;
+
+	dprintk("--> %s task->tk_status %d\n", __func__, task->tk_status);
+
+	/* Note this may cause RPC to be resent */
+	rdata->mds_ops->rpc_call_done(task, data);
+}
+
+static void filelayout_read_release(void *data)
+{
+	struct nfs_read_data *rdata = (struct nfs_read_data *)data;
+
+	rdata->mds_ops->rpc_release(data);
+}
+
+static int filelayout_write_done_cb(struct rpc_task *task,
+				struct nfs_write_data *data)
+{
+	int reset = 0;
+
+	if (filelayout_async_handle_error(task, data->args.context->state,
+					  data->ds_clp, &reset) == -EAGAIN) {
+		struct nfs_client *clp;
+
+		dprintk("%s calling restart ds_clp %p ds_clp->cl_session %p\n",
+			__func__, data->ds_clp, data->ds_clp->cl_session);
+		if (reset) {
+			filelayout_set_lo_fail(data->lseg);
+			nfs4_reset_write(task, data);
+			clp = NFS_SERVER(data->inode)->nfs_client;
+		} else
+			clp = data->ds_clp;
+		nfs_restart_rpc(task, clp);
+		return -EAGAIN;
+	}
 
-	if (nfss->nfs_client->cl_devid_cache)
-		pnfs_put_deviceid_cache(nfss->nfs_client);
 	return 0;
 }
 
+static void filelayout_write_prepare(struct rpc_task *task, void *data)
+{
+	struct nfs_write_data *wdata = (struct nfs_write_data *)data;
+
+	if (nfs41_setup_sequence(wdata->ds_clp->cl_session,
+				&wdata->args.seq_args, &wdata->res.seq_res,
+				0, task))
+		return;
+
+	rpc_call_start(task);
+}
+
+static void filelayout_write_call_done(struct rpc_task *task, void *data)
+{
+	struct nfs_write_data *wdata = (struct nfs_write_data *)data;
+
+	/* Note this may cause RPC to be resent */
+	wdata->mds_ops->rpc_call_done(task, data);
+}
+
+static void filelayout_write_release(void *data)
+{
+	struct nfs_write_data *wdata = (struct nfs_write_data *)data;
+
+	wdata->mds_ops->rpc_release(data);
+}
+
+struct rpc_call_ops filelayout_read_call_ops = {
+	.rpc_call_prepare = filelayout_read_prepare,
+	.rpc_call_done = filelayout_read_call_done,
+	.rpc_release = filelayout_read_release,
+};
+
+struct rpc_call_ops filelayout_write_call_ops = {
+	.rpc_call_prepare = filelayout_write_prepare,
+	.rpc_call_done = filelayout_write_call_done,
+	.rpc_release = filelayout_write_release,
+};
+
+static enum pnfs_try_status
+filelayout_read_pagelist(struct nfs_read_data *data)
+{
+	struct pnfs_layout_segment *lseg = data->lseg;
+	struct nfs4_pnfs_ds *ds;
+	loff_t offset = data->args.offset;
+	u32 j, idx;
+	struct nfs_fh *fh;
+	int status;
+
+	dprintk("--> %s ino %lu pgbase %u req %Zu@%llu\n",
+		__func__, data->inode->i_ino,
+		data->args.pgbase, (size_t)data->args.count, offset);
+
+	/* Retrieve the correct rpc_client for the byte range */
+	j = nfs4_fl_calc_j_index(lseg, offset);
+	idx = nfs4_fl_calc_ds_index(lseg, j);
+	ds = nfs4_fl_prepare_ds(lseg, idx);
+	if (!ds) {
+		/* Either layout fh index faulty, or ds connect failed */
+		set_bit(lo_fail_bit(IOMODE_RW), &lseg->pls_layout->plh_flags);
+		set_bit(lo_fail_bit(IOMODE_READ), &lseg->pls_layout->plh_flags);
+		return PNFS_NOT_ATTEMPTED;
+	}
+	dprintk("%s USE DS:ip %x %hu\n", __func__,
+		ntohl(ds->ds_ip_addr), ntohs(ds->ds_port));
+
+	/* No multipath support. Use first DS */
+	data->ds_clp = ds->ds_clp;
+	fh = nfs4_fl_select_ds_fh(lseg, j);
+	if (fh)
+		data->args.fh = fh;
+
+	data->args.offset = filelayout_get_dserver_offset(lseg, offset);
+	data->mds_offset = offset;
+
+	/* Perform an asynchronous read to ds */
+	status = nfs_initiate_read(data, ds->ds_clp->cl_rpcclient,
+				   &filelayout_read_call_ops);
+	BUG_ON(status != 0);
+	return PNFS_ATTEMPTED;
+}
+
+/* Perform async writes. */
+static enum pnfs_try_status
+filelayout_write_pagelist(struct nfs_write_data *data, int sync)
+{
+	struct pnfs_layout_segment *lseg = data->lseg;
+	struct nfs4_pnfs_ds *ds;
+	loff_t offset = data->args.offset;
+	u32 j, idx;
+	struct nfs_fh *fh;
+	int status;
+
+	/* Retrieve the correct rpc_client for the byte range */
+	j = nfs4_fl_calc_j_index(lseg, offset);
+	idx = nfs4_fl_calc_ds_index(lseg, j);
+	ds = nfs4_fl_prepare_ds(lseg, idx);
+	if (!ds) {
+		printk(KERN_ERR "%s: prepare_ds failed, use MDS\n", __func__);
+		set_bit(lo_fail_bit(IOMODE_RW), &lseg->pls_layout->plh_flags);
+		set_bit(lo_fail_bit(IOMODE_READ), &lseg->pls_layout->plh_flags);
+		return PNFS_NOT_ATTEMPTED;
+	}
+	dprintk("%s ino %lu sync %d req %Zu@%llu DS:%x:%hu\n", __func__,
+		data->inode->i_ino, sync, (size_t) data->args.count, offset,
+		ntohl(ds->ds_ip_addr), ntohs(ds->ds_port));
+
+	/* We can't handle commit to ds yet */
+	if (!FILELAYOUT_LSEG(lseg)->commit_through_mds)
+		data->args.stable = NFS_FILE_SYNC;
+
+	data->write_done_cb = filelayout_write_done_cb;
+	data->ds_clp = ds->ds_clp;
+	fh = nfs4_fl_select_ds_fh(lseg, j);
+	if (fh)
+		data->args.fh = fh;
+	/*
+	 * Get the file offset on the dserver. Set the write offset to
+	 * this offset and save the original offset.
+	 */
+	data->args.offset = filelayout_get_dserver_offset(lseg, offset);
+	data->mds_offset = offset;
+
+	/* Perform an asynchronous write */
+	status = nfs_initiate_write(data, ds->ds_clp->cl_rpcclient,
+				    &filelayout_write_call_ops, sync);
+	BUG_ON(status != 0);
+	return PNFS_ATTEMPTED;
+}
+
 /*
  * filelayout_check_layout()
  *
@@ -92,14 +369,14 @@ filelayout_check_layout(struct pnfs_layout_hdr *lo,
 		goto out;
 	}
 
-	if (fl->stripe_unit % PAGE_SIZE) {
-		dprintk("%s Stripe unit (%u) not page aligned\n",
+	if (!fl->stripe_unit || fl->stripe_unit % PAGE_SIZE) {
+		dprintk("%s Invalid stripe unit (%u)\n",
 			__func__, fl->stripe_unit);
 		goto out;
 	}
 
 	/* find and reference the deviceid */
-	dsaddr = nfs4_fl_find_get_deviceid(nfss->nfs_client, id);
+	dsaddr = nfs4_fl_find_get_deviceid(id);
 	if (dsaddr == NULL) {
 		dsaddr = get_device_info(lo->plh_inode, id);
 		if (dsaddr == NULL)
@@ -134,7 +411,7 @@ out:
 	dprintk("--> %s returns %d\n", __func__, status);
 	return status;
 out_put:
-	pnfs_put_deviceid(nfss->nfs_client->cl_devid_cache, &dsaddr->deviceid);
+	nfs4_fl_put_deviceid(dsaddr);
 	goto out;
 }
 
@@ -243,23 +520,47 @@ filelayout_alloc_lseg(struct pnfs_layout_hdr *layoutid,
 static void
 filelayout_free_lseg(struct pnfs_layout_segment *lseg)
 {
-	struct nfs_server *nfss = NFS_SERVER(lseg->pls_layout->plh_inode);
 	struct nfs4_filelayout_segment *fl = FILELAYOUT_LSEG(lseg);
 
 	dprintk("--> %s\n", __func__);
-	pnfs_put_deviceid(nfss->nfs_client->cl_devid_cache,
-			  &fl->dsaddr->deviceid);
+	nfs4_fl_put_deviceid(fl->dsaddr);
 	_filelayout_free_lseg(fl);
 }
 
+/*
+ * filelayout_pg_test(). Called by nfs_can_coalesce_requests()
+ *
+ * return 1 :  coalesce page
+ * return 0 :  don't coalesce page
+ */
+int
+filelayout_pg_test(struct nfs_pageio_descriptor *pgio, struct nfs_page *prev,
+		   struct nfs_page *req)
+{
+	u64 p_stripe, r_stripe;
+	u32 stripe_unit;
+
+	if (!pgio->pg_lseg)
+		return 1;
+	p_stripe = (u64)prev->wb_index << PAGE_CACHE_SHIFT;
+	r_stripe = (u64)req->wb_index << PAGE_CACHE_SHIFT;
+	stripe_unit = FILELAYOUT_LSEG(pgio->pg_lseg)->stripe_unit;
+
+	do_div(p_stripe, stripe_unit);
+	do_div(r_stripe, stripe_unit);
+
+	return (p_stripe == r_stripe);
+}
+
 static struct pnfs_layoutdriver_type filelayout_type = {
-	.id = LAYOUT_NFSV4_1_FILES,
-	.name = "LAYOUT_NFSV4_1_FILES",
-	.owner = THIS_MODULE,
-	.set_layoutdriver = filelayout_set_layoutdriver,
-	.clear_layoutdriver = filelayout_clear_layoutdriver,
-	.alloc_lseg              = filelayout_alloc_lseg,
-	.free_lseg               = filelayout_free_lseg,
+	.id			= LAYOUT_NFSV4_1_FILES,
+	.name			= "LAYOUT_NFSV4_1_FILES",
+	.owner			= THIS_MODULE,
+	.alloc_lseg		= filelayout_alloc_lseg,
+	.free_lseg		= filelayout_free_lseg,
+	.pg_test		= filelayout_pg_test,
+	.read_pagelist		= filelayout_read_pagelist,
+	.write_pagelist		= filelayout_write_pagelist,
 };
 
 static int __init nfs4filelayout_init(void)

fs/nfs/nfs4filelayout.h

@@ -55,8 +55,14 @@ struct nfs4_pnfs_ds {
 	atomic_t		ds_count;
 };
 
+/* nfs4_file_layout_dsaddr flags */
+#define NFS4_DEVICE_ID_NEG_ENTRY	0x00000001
+
 struct nfs4_file_layout_dsaddr {
-	struct pnfs_deviceid_node	deviceid;
+	struct hlist_node		node;
+	struct nfs4_deviceid		deviceid;
+	atomic_t			ref;
+	unsigned long			flags;
 	u32				stripe_count;
 	u8				*stripe_indices;
 	u32				ds_num;
@@ -83,11 +89,18 @@ FILELAYOUT_LSEG(struct pnfs_layout_segment *lseg)
 			    generic_hdr);
 }
 
-extern void nfs4_fl_free_deviceid_callback(struct pnfs_deviceid_node *);
+extern struct nfs_fh *
+nfs4_fl_select_ds_fh(struct pnfs_layout_segment *lseg, u32 j);
+
 extern void print_ds(struct nfs4_pnfs_ds *ds);
 extern void print_deviceid(struct nfs4_deviceid *dev_id);
+u32 nfs4_fl_calc_j_index(struct pnfs_layout_segment *lseg, loff_t offset);
+u32 nfs4_fl_calc_ds_index(struct pnfs_layout_segment *lseg, u32 j);
+struct nfs4_pnfs_ds *nfs4_fl_prepare_ds(struct pnfs_layout_segment *lseg,
+					u32 ds_idx);
 extern struct nfs4_file_layout_dsaddr *
-nfs4_fl_find_get_deviceid(struct nfs_client *, struct nfs4_deviceid *dev_id);
+nfs4_fl_find_get_deviceid(struct nfs4_deviceid *dev_id);
+extern void nfs4_fl_put_deviceid(struct nfs4_file_layout_dsaddr *dsaddr);
 struct nfs4_file_layout_dsaddr *
 get_device_info(struct inode *inode, struct nfs4_deviceid *dev_id);
 

fs/nfs/nfs4filelayoutdev.c

@@ -36,6 +36,30 @@
 
 #define NFSDBG_FACILITY		NFSDBG_PNFS_LD
 
+/*
+ * Device ID RCU cache. A device ID is unique per client ID and layout type.
+ */
+#define NFS4_FL_DEVICE_ID_HASH_BITS	5
+#define NFS4_FL_DEVICE_ID_HASH_SIZE	(1 << NFS4_FL_DEVICE_ID_HASH_BITS)
+#define NFS4_FL_DEVICE_ID_HASH_MASK	(NFS4_FL_DEVICE_ID_HASH_SIZE - 1)
+
+static inline u32
+nfs4_fl_deviceid_hash(struct nfs4_deviceid *id)
+{
+	unsigned char *cptr = (unsigned char *)id->data;
+	unsigned int nbytes = NFS4_DEVICEID4_SIZE;
+	u32 x = 0;
+
+	while (nbytes--) {
+		x *= 37;
+		x += *cptr++;
+	}
+	return x & NFS4_FL_DEVICE_ID_HASH_MASK;
+}
+
+static struct hlist_head filelayout_deviceid_cache[NFS4_FL_DEVICE_ID_HASH_SIZE];
+static DEFINE_SPINLOCK(filelayout_deviceid_lock);
+
 /*
  * Data server cache
  *
@@ -104,6 +128,67 @@ _data_server_lookup_locked(u32 ip_addr, u32 port)
 	return NULL;
 }
 
+/*
+ * Create an rpc connection to the nfs4_pnfs_ds data server
+ * Currently only support IPv4
+ */
+static int
+nfs4_ds_connect(struct nfs_server *mds_srv, struct nfs4_pnfs_ds *ds)
+{
+	struct nfs_client *clp;
+	struct sockaddr_in sin;
+	int status = 0;
+
+	dprintk("--> %s ip:port %x:%hu au_flavor %d\n", __func__,
+		ntohl(ds->ds_ip_addr), ntohs(ds->ds_port),
+		mds_srv->nfs_client->cl_rpcclient->cl_auth->au_flavor);
+
+	sin.sin_family = AF_INET;
+	sin.sin_addr.s_addr = ds->ds_ip_addr;
+	sin.sin_port = ds->ds_port;
+
+	clp = nfs4_set_ds_client(mds_srv->nfs_client, (struct sockaddr *)&sin,
+				 sizeof(sin), IPPROTO_TCP);
+	if (IS_ERR(clp)) {
+		status = PTR_ERR(clp);
+		goto out;
+	}
+
+	if ((clp->cl_exchange_flags & EXCHGID4_FLAG_MASK_PNFS) != 0) {
+		if (!is_ds_client(clp)) {
+			status = -ENODEV;
+			goto out_put;
+		}
+		ds->ds_clp = clp;
+		dprintk("%s [existing] ip=%x, port=%hu\n", __func__,
+			ntohl(ds->ds_ip_addr), ntohs(ds->ds_port));
+		goto out;
+	}
+
+	/*
+	 * Do not set NFS_CS_CHECK_LEASE_TIME instead set the DS lease to
+	 * be equal to the MDS lease. Renewal is scheduled in create_session.
+	 */
+	spin_lock(&mds_srv->nfs_client->cl_lock);
+	clp->cl_lease_time = mds_srv->nfs_client->cl_lease_time;
+	spin_unlock(&mds_srv->nfs_client->cl_lock);
+	clp->cl_last_renewal = jiffies;
+
+	/* New nfs_client */
+	status = nfs4_init_ds_session(clp);
+	if (status)
+		goto out_put;
+
+	ds->ds_clp = clp;
+	dprintk("%s [new] ip=%x, port=%hu\n", __func__, ntohl(ds->ds_ip_addr),
+		ntohs(ds->ds_port));
+out:
+	return status;
+out_put:
+	nfs_put_client(clp);
+	goto out;
+}
+
 static void
 destroy_ds(struct nfs4_pnfs_ds *ds)
 {
@@ -122,7 +207,7 @@ nfs4_fl_free_deviceid(struct nfs4_file_layout_dsaddr *dsaddr)
 	struct nfs4_pnfs_ds *ds;
 	int i;
 
-	print_deviceid(&dsaddr->deviceid.de_id);
+	print_deviceid(&dsaddr->deviceid);
 
 	for (i = 0; i < dsaddr->ds_num; i++) {
 		ds = dsaddr->ds_list[i];
@@ -139,15 +224,6 @@ nfs4_fl_free_deviceid(struct nfs4_file_layout_dsaddr *dsaddr)
 	kfree(dsaddr);
 }
 
-void
-nfs4_fl_free_deviceid_callback(struct pnfs_deviceid_node *device)
-{
-	struct nfs4_file_layout_dsaddr *dsaddr =
-		container_of(device, struct nfs4_file_layout_dsaddr, deviceid);
-
-	nfs4_fl_free_deviceid(dsaddr);
-}
-
 static struct nfs4_pnfs_ds *
 nfs4_pnfs_ds_add(struct inode *inode, u32 ip_addr, u32 port)
 {
@@ -300,7 +376,7 @@ decode_device(struct inode *ino, struct pnfs_device *pdev)
 	dsaddr->stripe_count = cnt;
 	dsaddr->ds_num = num;
 
-	memcpy(&dsaddr->deviceid.de_id, &pdev->dev_id, sizeof(pdev->dev_id));
+	memcpy(&dsaddr->deviceid, &pdev->dev_id, sizeof(pdev->dev_id));
 
 	/* Go back an read stripe indices */
 	p = indicesp;
@@ -350,28 +426,37 @@ out_err:
 }
 
 /*
- * Decode the opaque device specified in 'dev'
- * and add it to the list of available devices.
- * If the deviceid is already cached, nfs4_add_deviceid will return
- * a pointer to the cached struct and throw away the new.
+ * Decode the opaque device specified in 'dev' and add it to the cache of
+ * available devices.
  */
-static struct nfs4_file_layout_dsaddr*
+static struct nfs4_file_layout_dsaddr *
 decode_and_add_device(struct inode *inode, struct pnfs_device *dev)
 {
-	struct nfs4_file_layout_dsaddr *dsaddr;
-	struct pnfs_deviceid_node *d;
+	struct nfs4_file_layout_dsaddr *d, *new;
+	long hash;
 
-	dsaddr = decode_device(inode, dev);
-	if (!dsaddr) {
+	new = decode_device(inode, dev);
+	if (!new) {
 		printk(KERN_WARNING "%s: Could not decode or add device\n",
 			__func__);
 		return NULL;
 	}
 
-	d = pnfs_add_deviceid(NFS_SERVER(inode)->nfs_client->cl_devid_cache,
-			      &dsaddr->deviceid);
+	spin_lock(&filelayout_deviceid_lock);
+	d = nfs4_fl_find_get_deviceid(&new->deviceid);
+	if (d) {
+		spin_unlock(&filelayout_deviceid_lock);
+		nfs4_fl_free_deviceid(new);
+		return d;
+	}
+
+	INIT_HLIST_NODE(&new->node);
+	atomic_set(&new->ref, 1);
+	hash = nfs4_fl_deviceid_hash(&new->deviceid);
+	hlist_add_head_rcu(&new->node, &filelayout_deviceid_cache[hash]);
+	spin_unlock(&filelayout_deviceid_lock);
 
-	return container_of(d, struct nfs4_file_layout_dsaddr, deviceid);
+	return new;
 }
 
 /*
@@ -446,12 +531,123 @@ out_free:
 	return dsaddr;
 }
 
+void
+nfs4_fl_put_deviceid(struct nfs4_file_layout_dsaddr *dsaddr)
+{
+	if (atomic_dec_and_lock(&dsaddr->ref, &filelayout_deviceid_lock)) {
+		hlist_del_rcu(&dsaddr->node);
+		spin_unlock(&filelayout_deviceid_lock);
+
+		synchronize_rcu();
+		nfs4_fl_free_deviceid(dsaddr);
+	}
+}
+
 struct nfs4_file_layout_dsaddr *
-nfs4_fl_find_get_deviceid(struct nfs_client *clp, struct nfs4_deviceid *id)
+nfs4_fl_find_get_deviceid(struct nfs4_deviceid *id)
+{
+	struct nfs4_file_layout_dsaddr *d;
+	struct hlist_node *n;
+	long hash = nfs4_fl_deviceid_hash(id);
+
+
+	rcu_read_lock();
+	hlist_for_each_entry_rcu(d, n, &filelayout_deviceid_cache[hash], node) {
+		if (!memcmp(&d->deviceid, id, sizeof(*id))) {
+			if (!atomic_inc_not_zero(&d->ref))
+				goto fail;
+			rcu_read_unlock();
+			return d;
+		}
+	}
+fail:
+	rcu_read_unlock();
+	return NULL;
+}
+
+/*
+ * Want res = (offset - layout->pattern_offset)/ layout->stripe_unit
+ * Then: ((res + fsi) % dsaddr->stripe_count)
+ */
+u32
+nfs4_fl_calc_j_index(struct pnfs_layout_segment *lseg, loff_t offset)
+{
+	struct nfs4_filelayout_segment *flseg = FILELAYOUT_LSEG(lseg);
+	u64 tmp;
+
+	tmp = offset - flseg->pattern_offset;
+	do_div(tmp, flseg->stripe_unit);
+	tmp += flseg->first_stripe_index;
+	return do_div(tmp, flseg->dsaddr->stripe_count);
+}
+
+u32
+nfs4_fl_calc_ds_index(struct pnfs_layout_segment *lseg, u32 j)
+{
+	return FILELAYOUT_LSEG(lseg)->dsaddr->stripe_indices[j];
+}
+
+struct nfs_fh *
+nfs4_fl_select_ds_fh(struct pnfs_layout_segment *lseg, u32 j)
+{
+	struct nfs4_filelayout_segment *flseg = FILELAYOUT_LSEG(lseg);
+	u32 i;
+
+	if (flseg->stripe_type == STRIPE_SPARSE) {
+		if (flseg->num_fh == 1)
+			i = 0;
+		else if (flseg->num_fh == 0)
+			/* Use the MDS OPEN fh set in nfs_read_rpcsetup */
+			return NULL;
+		else
+			i = nfs4_fl_calc_ds_index(lseg, j);
+	} else
+		i = j;
+	return flseg->fh_array[i];
+}
+
+static void
+filelayout_mark_devid_negative(struct nfs4_file_layout_dsaddr *dsaddr,
+			       int err, u32 ds_addr)
+{
+	u32 *p = (u32 *)&dsaddr->deviceid;
+
+	printk(KERN_ERR "NFS: data server %x connection error %d."
+		" Deviceid [%x%x%x%x] marked out of use.\n",
+		ds_addr, err, p[0], p[1], p[2], p[3]);
+
+	spin_lock(&filelayout_deviceid_lock);
+	dsaddr->flags |= NFS4_DEVICE_ID_NEG_ENTRY;
+	spin_unlock(&filelayout_deviceid_lock);
+}
+
+struct nfs4_pnfs_ds *
+nfs4_fl_prepare_ds(struct pnfs_layout_segment *lseg, u32 ds_idx)
 {
-	struct pnfs_deviceid_node *d;
+	struct nfs4_file_layout_dsaddr *dsaddr = FILELAYOUT_LSEG(lseg)->dsaddr;
+	struct nfs4_pnfs_ds *ds = dsaddr->ds_list[ds_idx];
 
-	d = pnfs_find_get_deviceid(clp->cl_devid_cache, id);
-	return (d == NULL) ? NULL :
-		container_of(d, struct nfs4_file_layout_dsaddr, deviceid);
+	if (ds == NULL) {
+		printk(KERN_ERR "%s: No data server for offset index %d\n",
+			__func__, ds_idx);
+		return NULL;
+	}
+
+	if (!ds->ds_clp) {
+		struct nfs_server *s = NFS_SERVER(lseg->pls_layout->plh_inode);
+		int err;
+
+		if (dsaddr->flags & NFS4_DEVICE_ID_NEG_ENTRY) {
+			/* Already tried to connect, don't try again */
+			dprintk("%s Deviceid marked out of use\n", __func__);
+			return NULL;
+		}
+		err = nfs4_ds_connect(s, ds);
+		if (err) {
+			filelayout_mark_devid_negative(dsaddr, err,
+						       ntohl(ds->ds_ip_addr));
+			return NULL;
+		}
+	}
+	return ds;
 }

fs/nfs/nfs4proc.c

@@ -85,6 +85,9 @@ static int nfs4_map_errors(int err)
 	switch (err) {
 	case -NFS4ERR_RESOURCE:
 		return -EREMOTEIO;
+	case -NFS4ERR_BADOWNER:
+	case -NFS4ERR_BADNAME:
+		return -EINVAL;
 	default:
 		dprintk("%s could not handle NFSv4 error %d\n",
 				__func__, -err);
@@ -241,7 +244,7 @@ static int nfs4_delay(struct rpc_clnt *clnt, long *timeout)
 /* This is the error handling routine for processes that are allowed
  * to sleep.
  */
-static int nfs4_handle_exception(const struct nfs_server *server, int errorcode, struct nfs4_exception *exception)
+static int nfs4_handle_exception(struct nfs_server *server, int errorcode, struct nfs4_exception *exception)
 {
 	struct nfs_client *clp = server->nfs_client;
 	struct nfs4_state *state = exception->state;
@@ -293,6 +296,19 @@ static int nfs4_handle_exception(const struct nfs_server *server, int errorcode,
 				break;
 		case -NFS4ERR_OLD_STATEID:
 			exception->retry = 1;
+			break;
+		case -NFS4ERR_BADOWNER:
+			/* The following works around a Linux server bug! */
+		case -NFS4ERR_BADNAME:
+			if (server->caps & NFS_CAP_UIDGID_NOMAP) {
+				server->caps &= ~NFS_CAP_UIDGID_NOMAP;
+				exception->retry = 1;
+				printk(KERN_WARNING "NFS: v4 server %s "
+						"does not accept raw "
+						"uid/gids. "
+						"Reenabling the idmapper.\n",
+						server->nfs_client->cl_hostname);
+			}
 	}
 	/* We failed to handle the error */
 	return nfs4_map_errors(ret);
@@ -505,7 +521,7 @@ out:
 	return ret_id;
 }
 
-static int nfs41_setup_sequence(struct nfs4_session *session,
+int nfs41_setup_sequence(struct nfs4_session *session,
 				struct nfs4_sequence_args *args,
 				struct nfs4_sequence_res *res,
 				int cache_reply,
@@ -571,6 +587,7 @@ static int nfs41_setup_sequence(struct nfs4_session *session,
 	res->sr_status = 1;
 	return 0;
 }
+EXPORT_SYMBOL_GPL(nfs41_setup_sequence);
 
 int nfs4_setup_sequence(const struct nfs_server *server,
 			struct nfs4_sequence_args *args,
@@ -1573,9 +1590,8 @@ static int _nfs4_proc_open(struct nfs4_opendata *data)
 	return 0;
 }
 
-static int nfs4_recover_expired_lease(struct nfs_server *server)
+static int nfs4_client_recover_expired_lease(struct nfs_client *clp)
 {
-	struct nfs_client *clp = server->nfs_client;
 	unsigned int loop;
 	int ret;
 
@@ -1592,6 +1608,11 @@ static int nfs4_recover_expired_lease(struct nfs_server *server)
 	return ret;
 }
 
+static int nfs4_recover_expired_lease(struct nfs_server *server)
+{
+	return nfs4_client_recover_expired_lease(server->nfs_client);
+}
+
 /*
  * OPEN_EXPIRED:
  * 	reclaim state on the server after a network partition.
@@ -3069,15 +3090,10 @@ static int nfs4_proc_pathconf(struct nfs_server *server, struct nfs_fh *fhandle,
 	return err;
 }
 
-static int nfs4_read_done(struct rpc_task *task, struct nfs_read_data *data)
+static int nfs4_read_done_cb(struct rpc_task *task, struct nfs_read_data *data)
 {
 	struct nfs_server *server = NFS_SERVER(data->inode);
 
-	dprintk("--> %s\n", __func__);
-
-	if (!nfs4_sequence_done(task, &data->res.seq_res))
-		return -EAGAIN;
-
 	if (nfs4_async_handle_error(task, server, data->args.context->state) == -EAGAIN) {
 		nfs_restart_rpc(task, server->nfs_client);
 		return -EAGAIN;
@@ -3089,19 +3105,44 @@ static int nfs4_read_done(struct rpc_task *task, struct nfs_read_data *data)
 	return 0;
 }
 
+static int nfs4_read_done(struct rpc_task *task, struct nfs_read_data *data)
+{
+
+	dprintk("--> %s\n", __func__);
+
+	if (!nfs4_sequence_done(task, &data->res.seq_res))
+		return -EAGAIN;
+
+	return data->read_done_cb(task, data);
+}
+
 static void nfs4_proc_read_setup(struct nfs_read_data *data, struct rpc_message *msg)
 {
 	data->timestamp   = jiffies;
+	data->read_done_cb = nfs4_read_done_cb;
 	msg->rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_READ];
 }
 
-static int nfs4_write_done(struct rpc_task *task, struct nfs_write_data *data)
+/* Reset the the nfs_read_data to send the read to the MDS. */
+void nfs4_reset_read(struct rpc_task *task, struct nfs_read_data *data)
+{
+	dprintk("%s Reset task for i/o through\n", __func__);
+	put_lseg(data->lseg);
+	data->lseg = NULL;
+	/* offsets will differ in the dense stripe case */
+	data->args.offset = data->mds_offset;
+	data->ds_clp = NULL;
+	data->args.fh     = NFS_FH(data->inode);
+	data->read_done_cb = nfs4_read_done_cb;
+	task->tk_ops = data->mds_ops;
+	rpc_task_reset_client(task, NFS_CLIENT(data->inode));
+}
+EXPORT_SYMBOL_GPL(nfs4_reset_read);
+
+static int nfs4_write_done_cb(struct rpc_task *task, struct nfs_write_data *data)
 {
 	struct inode *inode = data->inode;
 	
-	if (!nfs4_sequence_done(task, &data->res.seq_res))
-		return -EAGAIN;
-
 	if (nfs4_async_handle_error(task, NFS_SERVER(inode), data->args.context->state) == -EAGAIN) {
 		nfs_restart_rpc(task, NFS_SERVER(inode)->nfs_client);
 		return -EAGAIN;
@@ -3113,11 +3154,41 @@ static int nfs4_write_done(struct rpc_task *task, struct nfs_write_data *data)
 	return 0;
 }
 
+static int nfs4_write_done(struct rpc_task *task, struct nfs_write_data *data)
+{
+	if (!nfs4_sequence_done(task, &data->res.seq_res))
+		return -EAGAIN;
+	return data->write_done_cb(task, data);
+}
+
+/* Reset the the nfs_write_data to send the write to the MDS. */
+void nfs4_reset_write(struct rpc_task *task, struct nfs_write_data *data)
+{
+	dprintk("%s Reset task for i/o through\n", __func__);
+	put_lseg(data->lseg);
+	data->lseg          = NULL;
+	data->ds_clp        = NULL;
+	data->write_done_cb = nfs4_write_done_cb;
+	data->args.fh       = NFS_FH(data->inode);
+	data->args.bitmask  = data->res.server->cache_consistency_bitmask;
+	data->args.offset   = data->mds_offset;
+	data->res.fattr     = &data->fattr;
+	task->tk_ops        = data->mds_ops;
+	rpc_task_reset_client(task, NFS_CLIENT(data->inode));
+}
+EXPORT_SYMBOL_GPL(nfs4_reset_write);
+
 static void nfs4_proc_write_setup(struct nfs_write_data *data, struct rpc_message *msg)
 {
 	struct nfs_server *server = NFS_SERVER(data->inode);
 
-	data->args.bitmask = server->cache_consistency_bitmask;
+	if (data->lseg) {
+		data->args.bitmask = NULL;
+		data->res.fattr = NULL;
+	} else
+		data->args.bitmask = server->cache_consistency_bitmask;
+	if (!data->write_done_cb)
+		data->write_done_cb = nfs4_write_done_cb;
 	data->res.server = server;
 	data->timestamp   = jiffies;
 
@@ -5118,6 +5189,27 @@ int nfs4_init_session(struct nfs_server *server)
 	return ret;
 }
 
+int nfs4_init_ds_session(struct nfs_client *clp)
+{
+	struct nfs4_session *session = clp->cl_session;
+	int ret;
+
+	if (!test_and_clear_bit(NFS4_SESSION_INITING, &session->session_state))
+		return 0;
+
+	ret = nfs4_client_recover_expired_lease(clp);
+	if (!ret)
+		/* Test for the DS role */
+		if (!is_ds_client(clp))
+			ret = -ENODEV;
+	if (!ret)
+		ret = nfs4_check_client_ready(clp);
+	return ret;
+
+}
+EXPORT_SYMBOL_GPL(nfs4_init_ds_session);
+
+
 /*
  * Renew the cl_session lease.
  */
@@ -5648,6 +5740,7 @@ const struct nfs_rpc_ops nfs_v4_clientops = {
 	.clear_acl_cache = nfs4_zap_acl_attr,
 	.close_context  = nfs4_close_context,
 	.open_context	= nfs4_atomic_open,
+	.init_client	= nfs4_init_client,
 };
 
 static const struct xattr_handler nfs4_xattr_nfs4_acl_handler = {

fs/nfs/nfs4renewd.c

@@ -64,12 +64,8 @@ nfs4_renew_state(struct work_struct *work)
 	ops = clp->cl_mvops->state_renewal_ops;
 	dprintk("%s: start\n", __func__);
 
-	rcu_read_lock();
-	if (list_empty(&clp->cl_superblocks)) {
-		rcu_read_unlock();
+	if (test_bit(NFS_CS_STOP_RENEW, &clp->cl_res_state))
 		goto out;
-	}
-	rcu_read_unlock();
 
 	spin_lock(&clp->cl_lock);
 	lease = clp->cl_lease_time;

fs/nfs/nfs4state.c

@@ -153,6 +153,11 @@ static int nfs41_setup_state_renewal(struct nfs_client *clp)
 	int status;
 	struct nfs_fsinfo fsinfo;
 
+	if (!test_bit(NFS_CS_CHECK_LEASE_TIME, &clp->cl_res_state)) {
+		nfs4_schedule_state_renewal(clp);
+		return 0;
+	}
+
 	status = nfs4_proc_get_lease_time(clp, &fsinfo);
 	if (status == 0) {
 		/* Update lease time and schedule renewal */
@@ -1448,6 +1453,7 @@ void nfs4_schedule_session_recovery(struct nfs4_session *session)
 {
 	nfs4_schedule_lease_recovery(session->clp);
 }
+EXPORT_SYMBOL_GPL(nfs4_schedule_session_recovery);
 
 void nfs41_handle_recall_slot(struct nfs_client *clp)
 {

fs/nfs/nfs4xdr.c

@@ -844,7 +844,7 @@ static void encode_attrs(struct xdr_stream *xdr, const struct iattr *iap, const
 	if (iap->ia_valid & ATTR_MODE)
 		len += 4;
 	if (iap->ia_valid & ATTR_UID) {
-		owner_namelen = nfs_map_uid_to_name(server->nfs_client, iap->ia_uid, owner_name, IDMAP_NAMESZ);
+		owner_namelen = nfs_map_uid_to_name(server, iap->ia_uid, owner_name, IDMAP_NAMESZ);
 		if (owner_namelen < 0) {
 			dprintk("nfs: couldn't resolve uid %d to string\n",
 					iap->ia_uid);
@@ -856,7 +856,7 @@ static void encode_attrs(struct xdr_stream *xdr, const struct iattr *iap, const
 		len += 4 + (XDR_QUADLEN(owner_namelen) << 2);
 	}
 	if (iap->ia_valid & ATTR_GID) {
-		owner_grouplen = nfs_map_gid_to_group(server->nfs_client, iap->ia_gid, owner_group, IDMAP_NAMESZ);
+		owner_grouplen = nfs_map_gid_to_group(server, iap->ia_gid, owner_group, IDMAP_NAMESZ);
 		if (owner_grouplen < 0) {
 			dprintk("nfs: couldn't resolve gid %d to string\n",
 					iap->ia_gid);
@@ -1384,7 +1384,7 @@ static void encode_putrootfh(struct xdr_stream *xdr, struct compound_hdr *hdr)
 	hdr->replen += decode_putrootfh_maxsz;
 }
 
-static void encode_stateid(struct xdr_stream *xdr, const struct nfs_open_context *ctx, const struct nfs_lock_context *l_ctx)
+static void encode_stateid(struct xdr_stream *xdr, const struct nfs_open_context *ctx, const struct nfs_lock_context *l_ctx, int zero_seqid)
 {
 	nfs4_stateid stateid;
 	__be32 *p;
@@ -1392,6 +1392,8 @@ static void encode_stateid(struct xdr_stream *xdr, const struct nfs_open_context
 	p = reserve_space(xdr, NFS4_STATEID_SIZE);
 	if (ctx->state != NULL) {
 		nfs4_copy_stateid(&stateid, ctx->state, l_ctx->lockowner, l_ctx->pid);
+		if (zero_seqid)
+			stateid.stateid.seqid = 0;
 		xdr_encode_opaque_fixed(p, stateid.data, NFS4_STATEID_SIZE);
 	} else
 		xdr_encode_opaque_fixed(p, zero_stateid.data, NFS4_STATEID_SIZE);
@@ -1404,7 +1406,8 @@ static void encode_read(struct xdr_stream *xdr, const struct nfs_readargs *args,
 	p = reserve_space(xdr, 4);
 	*p = cpu_to_be32(OP_READ);
 
-	encode_stateid(xdr, args->context, args->lock_context);
+	encode_stateid(xdr, args->context, args->lock_context,
+		       hdr->minorversion);
 
 	p = reserve_space(xdr, 12);
 	p = xdr_encode_hyper(p, args->offset);
@@ -1592,7 +1595,8 @@ static void encode_write(struct xdr_stream *xdr, const struct nfs_writeargs *arg
 	p = reserve_space(xdr, 4);
 	*p = cpu_to_be32(OP_WRITE);
 
-	encode_stateid(xdr, args->context, args->lock_context);
+	encode_stateid(xdr, args->context, args->lock_context,
+		       hdr->minorversion);
 
 	p = reserve_space(xdr, 16);
 	p = xdr_encode_hyper(p, args->offset);
@@ -2271,7 +2275,8 @@ static void nfs4_xdr_enc_write(struct rpc_rqst *req, struct xdr_stream *xdr,
 	encode_putfh(xdr, args->fh, &hdr);
 	encode_write(xdr, args, &hdr);
 	req->rq_snd_buf.flags |= XDRBUF_WRITE;
-	encode_getfattr(xdr, args->bitmask, &hdr);
+	if (args->bitmask)
+		encode_getfattr(xdr, args->bitmask, &hdr);
 	encode_nops(&hdr);
 }
 
@@ -3382,7 +3387,7 @@ out_overflow:
 }
 
 static int decode_attr_owner(struct xdr_stream *xdr, uint32_t *bitmap,
-		struct nfs_client *clp, uint32_t *uid, int may_sleep)
+		const struct nfs_server *server, uint32_t *uid, int may_sleep)
 {
 	uint32_t len;
 	__be32 *p;
@@ -3402,7 +3407,7 @@ static int decode_attr_owner(struct xdr_stream *xdr, uint32_t *bitmap,
 		if (!may_sleep) {
 			/* do nothing */
 		} else if (len < XDR_MAX_NETOBJ) {
-			if (nfs_map_name_to_uid(clp, (char *)p, len, uid) == 0)
+			if (nfs_map_name_to_uid(server, (char *)p, len, uid) == 0)
 				ret = NFS_ATTR_FATTR_OWNER;
 			else
 				dprintk("%s: nfs_map_name_to_uid failed!\n",
@@ -3420,7 +3425,7 @@ out_overflow:
 }
 
 static int decode_attr_group(struct xdr_stream *xdr, uint32_t *bitmap,
-		struct nfs_client *clp, uint32_t *gid, int may_sleep)
+		const struct nfs_server *server, uint32_t *gid, int may_sleep)
 {
 	uint32_t len;
 	__be32 *p;
@@ -3440,7 +3445,7 @@ static int decode_attr_group(struct xdr_stream *xdr, uint32_t *bitmap,
 		if (!may_sleep) {
 			/* do nothing */
 		} else if (len < XDR_MAX_NETOBJ) {
-			if (nfs_map_group_to_gid(clp, (char *)p, len, gid) == 0)
+			if (nfs_map_group_to_gid(server, (char *)p, len, gid) == 0)
 				ret = NFS_ATTR_FATTR_GROUP;
 			else
 				dprintk("%s: nfs_map_group_to_gid failed!\n",
@@ -3939,14 +3944,12 @@ static int decode_getfattr_attrs(struct xdr_stream *xdr, uint32_t *bitmap,
 		goto xdr_error;
 	fattr->valid |= status;
 
-	status = decode_attr_owner(xdr, bitmap, server->nfs_client,
-			&fattr->uid, may_sleep);
+	status = decode_attr_owner(xdr, bitmap, server, &fattr->uid, may_sleep);
 	if (status < 0)
 		goto xdr_error;
 	fattr->valid |= status;
 
-	status = decode_attr_group(xdr, bitmap, server->nfs_client,
-			&fattr->gid, may_sleep);
+	status = decode_attr_group(xdr, bitmap, server, &fattr->gid, may_sleep);
 	if (status < 0)
 		goto xdr_error;
 	fattr->valid |= status;
@@ -5690,8 +5693,9 @@ static int nfs4_xdr_dec_write(struct rpc_rqst *rqstp, struct xdr_stream *xdr,
 	status = decode_write(xdr, res);
 	if (status)
 		goto out;
-	decode_getfattr(xdr, res->fattr, res->server,
-			!RPC_IS_ASYNC(rqstp->rq_task));
+	if (res->fattr)
+		decode_getfattr(xdr, res->fattr, res->server,
+				!RPC_IS_ASYNC(rqstp->rq_task));
 	if (!status)
 		status = res->count;
 out:
@@ -6167,8 +6171,6 @@ static struct {
 	{ NFS4ERR_DQUOT,	-EDQUOT		},
 	{ NFS4ERR_STALE,	-ESTALE		},
 	{ NFS4ERR_BADHANDLE,	-EBADHANDLE	},
-	{ NFS4ERR_BADOWNER,	-EINVAL		},
-	{ NFS4ERR_BADNAME,	-EINVAL		},
 	{ NFS4ERR_BAD_COOKIE,	-EBADCOOKIE	},
 	{ NFS4ERR_NOTSUPP,	-ENOTSUPP	},
 	{ NFS4ERR_TOOSMALL,	-ETOOSMALL	},

fs/nfs/pagelist.c

@@ -20,6 +20,7 @@
 #include <linux/nfs_mount.h>
 
 #include "internal.h"
+#include "pnfs.h"
 
 static struct kmem_cache *nfs_page_cachep;
 
@@ -213,7 +214,7 @@ nfs_wait_on_request(struct nfs_page *req)
  */
 void nfs_pageio_init(struct nfs_pageio_descriptor *desc,
 		     struct inode *inode,
-		     int (*doio)(struct inode *, struct list_head *, unsigned int, size_t, int),
+		     int (*doio)(struct nfs_pageio_descriptor *),
 		     size_t bsize,
 		     int io_flags)
 {
@@ -226,6 +227,7 @@ void nfs_pageio_init(struct nfs_pageio_descriptor *desc,
 	desc->pg_doio = doio;
 	desc->pg_ioflags = io_flags;
 	desc->pg_error = 0;
+	desc->pg_lseg = NULL;
 }
 
 /**
@@ -240,7 +242,8 @@ void nfs_pageio_init(struct nfs_pageio_descriptor *desc,
  * Return 'true' if this is the case, else return 'false'.
  */
 static int nfs_can_coalesce_requests(struct nfs_page *prev,
-				     struct nfs_page *req)
+				     struct nfs_page *req,
+				     struct nfs_pageio_descriptor *pgio)
 {
 	if (req->wb_context->cred != prev->wb_context->cred)
 		return 0;
@@ -254,6 +257,12 @@ static int nfs_can_coalesce_requests(struct nfs_page *prev,
 		return 0;
 	if (prev->wb_pgbase + prev->wb_bytes != PAGE_CACHE_SIZE)
 		return 0;
+	/*
+	 * Non-whole file layouts need to check that req is inside of
+	 * pgio->pg_lseg.
+	 */
+	if (pgio->pg_test && !pgio->pg_test(pgio, prev, req))
+		return 0;
 	return 1;
 }
 
@@ -286,7 +295,7 @@ static int nfs_pageio_do_add_request(struct nfs_pageio_descriptor *desc,
 		if (newlen > desc->pg_bsize)
 			return 0;
 		prev = nfs_list_entry(desc->pg_list.prev);
-		if (!nfs_can_coalesce_requests(prev, req))
+		if (!nfs_can_coalesce_requests(prev, req, desc))
 			return 0;
 	} else
 		desc->pg_base = req->wb_pgbase;
@@ -302,12 +311,7 @@ static int nfs_pageio_do_add_request(struct nfs_pageio_descriptor *desc,
 static void nfs_pageio_doio(struct nfs_pageio_descriptor *desc)
 {
 	if (!list_empty(&desc->pg_list)) {
-		int error = desc->pg_doio(desc->pg_inode,
-					  &desc->pg_list,
-					  nfs_page_array_len(desc->pg_base,
-							     desc->pg_count),
-					  desc->pg_count,
-					  desc->pg_ioflags);
+		int error = desc->pg_doio(desc);
 		if (error < 0)
 			desc->pg_error = error;
 		else

fs/nfs/pnfs.c

@@ -30,6 +30,7 @@
 #include <linux/nfs_fs.h>
 #include "internal.h"
 #include "pnfs.h"
+#include "iostat.h"
 
 #define NFSDBG_FACILITY		NFSDBG_PNFS
 
@@ -74,10 +75,8 @@ find_pnfs_driver(u32 id)
 void
 unset_pnfs_layoutdriver(struct nfs_server *nfss)
 {
-	if (nfss->pnfs_curr_ld) {
-		nfss->pnfs_curr_ld->clear_layoutdriver(nfss);
+	if (nfss->pnfs_curr_ld)
 		module_put(nfss->pnfs_curr_ld->owner);
-	}
 	nfss->pnfs_curr_ld = NULL;
 }
 
@@ -115,13 +114,7 @@ set_pnfs_layoutdriver(struct nfs_server *server, u32 id)
 		goto out_no_driver;
 	}
 	server->pnfs_curr_ld = ld_type;
-	if (ld_type->set_layoutdriver(server)) {
-		printk(KERN_ERR
-		       "%s: Error initializing mount point for layout driver %u.\n",
-		       __func__, id);
-		module_put(ld_type->owner);
-		goto out_no_driver;
-	}
+
 	dprintk("%s: pNFS module for %u set\n", __func__, id);
 	return;
 
@@ -230,37 +223,41 @@ static void free_lseg(struct pnfs_layout_segment *lseg)
 	put_layout_hdr(NFS_I(ino)->layout);
 }
 
-/* The use of tmp_list is necessary because pnfs_curr_ld->free_lseg
- * could sleep, so must be called outside of the lock.
- * Returns 1 if object was removed, otherwise return 0.
- */
-static int
-put_lseg_locked(struct pnfs_layout_segment *lseg,
-		struct list_head *tmp_list)
+static void
+put_lseg_common(struct pnfs_layout_segment *lseg)
+{
+	struct inode *inode = lseg->pls_layout->plh_inode;
+
+	BUG_ON(test_bit(NFS_LSEG_VALID, &lseg->pls_flags));
+	list_del_init(&lseg->pls_list);
+	if (list_empty(&lseg->pls_layout->plh_segs)) {
+		set_bit(NFS_LAYOUT_DESTROYED, &lseg->pls_layout->plh_flags);
+		/* Matched by initial refcount set in alloc_init_layout_hdr */
+		put_layout_hdr_locked(lseg->pls_layout);
+	}
+	rpc_wake_up(&NFS_SERVER(inode)->roc_rpcwaitq);
+}
+
+void
+put_lseg(struct pnfs_layout_segment *lseg)
 {
+	struct inode *inode;
+
+	if (!lseg)
+		return;
+
 	dprintk("%s: lseg %p ref %d valid %d\n", __func__, lseg,
 		atomic_read(&lseg->pls_refcount),
 		test_bit(NFS_LSEG_VALID, &lseg->pls_flags));
-	if (atomic_dec_and_test(&lseg->pls_refcount)) {
-		struct inode *ino = lseg->pls_layout->plh_inode;
+	inode = lseg->pls_layout->plh_inode;
+	if (atomic_dec_and_lock(&lseg->pls_refcount, &inode->i_lock)) {
+		LIST_HEAD(free_me);
 
-		BUG_ON(test_bit(NFS_LSEG_VALID, &lseg->pls_flags));
-		list_del(&lseg->pls_list);
-		if (list_empty(&lseg->pls_layout->plh_segs)) {
-			struct nfs_client *clp;
-
-			clp = NFS_SERVER(ino)->nfs_client;
-			spin_lock(&clp->cl_lock);
-			/* List does not take a reference, so no need for put here */
-			list_del_init(&lseg->pls_layout->plh_layouts);
-			spin_unlock(&clp->cl_lock);
-			clear_bit(NFS_LAYOUT_BULK_RECALL, &lseg->pls_layout->plh_flags);
-		}
-		rpc_wake_up(&NFS_SERVER(ino)->roc_rpcwaitq);
-		list_add(&lseg->pls_list, tmp_list);
-		return 1;
+		put_lseg_common(lseg);
+		list_add(&lseg->pls_list, &free_me);
+		spin_unlock(&inode->i_lock);
+		pnfs_free_lseg_list(&free_me);
 	}
-	return 0;
 }
 
 static bool
@@ -281,7 +278,13 @@ static int mark_lseg_invalid(struct pnfs_layout_segment *lseg,
 		 * list.  It will now be removed when all
 		 * outstanding io is finished.
 		 */
-		rv = put_lseg_locked(lseg, tmp_list);
+		dprintk("%s: lseg %p ref %d\n", __func__, lseg,
+			atomic_read(&lseg->pls_refcount));
+		if (atomic_dec_and_test(&lseg->pls_refcount)) {
+			put_lseg_common(lseg);
+			list_add(&lseg->pls_list, tmp_list);
+			rv = 1;
+		}
 	}
 	return rv;
 }
@@ -299,6 +302,11 @@ mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo,
 
 	dprintk("%s:Begin lo %p\n", __func__, lo);
 
+	if (list_empty(&lo->plh_segs)) {
+		if (!test_and_set_bit(NFS_LAYOUT_DESTROYED, &lo->plh_flags))
+			put_layout_hdr_locked(lo);
+		return 0;
+	}
 	list_for_each_entry_safe(lseg, next, &lo->plh_segs, pls_list)
 		if (should_free_lseg(lseg->pls_range.iomode, iomode)) {
 			dprintk("%s: freeing lseg %p iomode %d "
@@ -312,11 +320,27 @@ mark_matching_lsegs_invalid(struct pnfs_layout_hdr *lo,
 	return invalid - removed;
 }
 
+/* note free_me must contain lsegs from a single layout_hdr */
 void
 pnfs_free_lseg_list(struct list_head *free_me)
 {
 	struct pnfs_layout_segment *lseg, *tmp;
+	struct pnfs_layout_hdr *lo;
+
+	if (list_empty(free_me))
+		return;
 
+	lo = list_first_entry(free_me, struct pnfs_layout_segment,
+			      pls_list)->pls_layout;
+
+	if (test_bit(NFS_LAYOUT_DESTROYED, &lo->plh_flags)) {
+		struct nfs_client *clp;
+
+		clp = NFS_SERVER(lo->plh_inode)->nfs_client;
+		spin_lock(&clp->cl_lock);
+		list_del_init(&lo->plh_layouts);
+		spin_unlock(&clp->cl_lock);
+	}
 	list_for_each_entry_safe(lseg, tmp, free_me, pls_list) {
 		list_del(&lseg->pls_list);
 		free_lseg(lseg);
@@ -332,10 +356,8 @@ pnfs_destroy_layout(struct nfs_inode *nfsi)
 	spin_lock(&nfsi->vfs_inode.i_lock);
 	lo = nfsi->layout;
 	if (lo) {
-		set_bit(NFS_LAYOUT_DESTROYED, &nfsi->layout->plh_flags);
+		lo->plh_block_lgets++; /* permanently block new LAYOUTGETs */
 		mark_matching_lsegs_invalid(lo, &tmp_list, IOMODE_ANY);
-		/* Matched by refcount set to 1 in alloc_init_layout_hdr */
-		put_layout_hdr_locked(lo);
 	}
 	spin_unlock(&nfsi->vfs_inode.i_lock);
 	pnfs_free_lseg_list(&tmp_list);
@@ -403,6 +425,7 @@ pnfs_layoutgets_blocked(struct pnfs_layout_hdr *lo, nfs4_stateid *stateid,
 	    (int)(lo->plh_barrier - be32_to_cpu(stateid->stateid.seqid)) >= 0)
 		return true;
 	return lo->plh_block_lgets ||
+		test_bit(NFS_LAYOUT_DESTROYED, &lo->plh_flags) ||
 		test_bit(NFS_LAYOUT_BULK_RECALL, &lo->plh_flags) ||
 		(list_empty(&lo->plh_segs) &&
 		 (atomic_read(&lo->plh_outstanding) > lget));
@@ -674,7 +697,7 @@ pnfs_find_lseg(struct pnfs_layout_hdr *lo, u32 iomode)
 	list_for_each_entry(lseg, &lo->plh_segs, pls_list) {
 		if (test_bit(NFS_LSEG_VALID, &lseg->pls_flags) &&
 		    is_matching_lseg(lseg, iomode)) {
-			ret = lseg;
+			ret = get_lseg(lseg);
 			break;
 		}
 		if (cmp_layout(iomode, lseg->pls_range.iomode) > 0)
@@ -699,6 +722,7 @@ pnfs_update_layout(struct inode *ino,
 	struct nfs_client *clp = NFS_SERVER(ino)->nfs_client;
 	struct pnfs_layout_hdr *lo;
 	struct pnfs_layout_segment *lseg = NULL;
+	bool first = false;
 
 	if (!pnfs_enabled_sb(NFS_SERVER(ino)))
 		return NULL;
@@ -715,21 +739,25 @@ pnfs_update_layout(struct inode *ino,
 		dprintk("%s matches recall, use MDS\n", __func__);
 		goto out_unlock;
 	}
-	/* Check to see if the layout for the given range already exists */
-	lseg = pnfs_find_lseg(lo, iomode);
-	if (lseg)
-		goto out_unlock;
 
 	/* if LAYOUTGET already failed once we don't try again */
 	if (test_bit(lo_fail_bit(iomode), &nfsi->layout->plh_flags))
 		goto out_unlock;
 
+	/* Check to see if the layout for the given range already exists */
+	lseg = pnfs_find_lseg(lo, iomode);
+	if (lseg)
+		goto out_unlock;
+
 	if (pnfs_layoutgets_blocked(lo, NULL, 0))
 		goto out_unlock;
 	atomic_inc(&lo->plh_outstanding);
 
 	get_layout_hdr(lo);
-	if (list_empty(&lo->plh_segs)) {
+	if (list_empty(&lo->plh_segs))
+		first = true;
+	spin_unlock(&ino->i_lock);
+	if (first) {
 		/* The lo must be on the clp list if there is any
 		 * chance of a CB_LAYOUTRECALL(FILE) coming in.
 		 */
@@ -738,24 +766,18 @@ pnfs_update_layout(struct inode *ino,
 		list_add_tail(&lo->plh_layouts, &clp->cl_layouts);
 		spin_unlock(&clp->cl_lock);
 	}
-	spin_unlock(&ino->i_lock);
 
 	lseg = send_layoutget(lo, ctx, iomode);
-	if (!lseg) {
-		spin_lock(&ino->i_lock);
-		if (list_empty(&lo->plh_segs)) {
-			spin_lock(&clp->cl_lock);
-			list_del_init(&lo->plh_layouts);
-			spin_unlock(&clp->cl_lock);
-			clear_bit(NFS_LAYOUT_BULK_RECALL, &lo->plh_flags);
-		}
-		spin_unlock(&ino->i_lock);
+	if (!lseg && first) {
+		spin_lock(&clp->cl_lock);
+		list_del_init(&lo->plh_layouts);
+		spin_unlock(&clp->cl_lock);
 	}
 	atomic_dec(&lo->plh_outstanding);
 	put_layout_hdr(lo);
 out:
 	dprintk("%s end, state 0x%lx lseg %p\n", __func__,
-		nfsi->layout->plh_flags, lseg);
+		nfsi->layout ? nfsi->layout->plh_flags : -1, lseg);
 	return lseg;
 out_unlock:
 	spin_unlock(&ino->i_lock);
@@ -808,7 +830,7 @@ pnfs_layout_process(struct nfs4_layoutget *lgp)
 	}
 	init_lseg(lo, lseg);
 	lseg->pls_range = res->range;
-	*lgp->lsegpp = lseg;
+	*lgp->lsegpp = get_lseg(lseg);
 	pnfs_insert_layout(lo, lseg);
 
 	if (res->return_on_close) {
@@ -829,137 +851,97 @@ out_forget_reply:
 	goto out;
 }
 
-/*
- * Device ID cache. Currently supports one layout type per struct nfs_client.
- * Add layout type to the lookup key to expand to support multiple types.
- */
-int
-pnfs_alloc_init_deviceid_cache(struct nfs_client *clp,
-			 void (*free_callback)(struct pnfs_deviceid_node *))
+static int pnfs_read_pg_test(struct nfs_pageio_descriptor *pgio,
+			     struct nfs_page *prev,
+			     struct nfs_page *req)
 {
-	struct pnfs_deviceid_cache *c;
-
-	c = kzalloc(sizeof(struct pnfs_deviceid_cache), GFP_KERNEL);
-	if (!c)
-		return -ENOMEM;
-	spin_lock(&clp->cl_lock);
-	if (clp->cl_devid_cache != NULL) {
-		atomic_inc(&clp->cl_devid_cache->dc_ref);
-		dprintk("%s [kref [%d]]\n", __func__,
-			atomic_read(&clp->cl_devid_cache->dc_ref));
-		kfree(c);
-	} else {
-		/* kzalloc initializes hlists */
-		spin_lock_init(&c->dc_lock);
-		atomic_set(&c->dc_ref, 1);
-		c->dc_free_callback = free_callback;
-		clp->cl_devid_cache = c;
-		dprintk("%s [new]\n", __func__);
+	if (pgio->pg_count == prev->wb_bytes) {
+		/* This is first coelesce call for a series of nfs_pages */
+		pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
+						   prev->wb_context,
+						   IOMODE_READ);
 	}
-	spin_unlock(&clp->cl_lock);
-	return 0;
+	return NFS_SERVER(pgio->pg_inode)->pnfs_curr_ld->pg_test(pgio, prev, req);
 }
-EXPORT_SYMBOL_GPL(pnfs_alloc_init_deviceid_cache);
 
-/*
- * Called from pnfs_layoutdriver_type->free_lseg
- * last layout segment reference frees deviceid
- */
 void
-pnfs_put_deviceid(struct pnfs_deviceid_cache *c,
-		  struct pnfs_deviceid_node *devid)
+pnfs_pageio_init_read(struct nfs_pageio_descriptor *pgio, struct inode *inode)
 {
-	struct nfs4_deviceid *id = &devid->de_id;
-	struct pnfs_deviceid_node *d;
-	struct hlist_node *n;
-	long h = nfs4_deviceid_hash(id);
+	struct pnfs_layoutdriver_type *ld;
 
-	dprintk("%s [%d]\n", __func__, atomic_read(&devid->de_ref));
-	if (!atomic_dec_and_lock(&devid->de_ref, &c->dc_lock))
-		return;
+	ld = NFS_SERVER(inode)->pnfs_curr_ld;
+	pgio->pg_test = (ld && ld->pg_test) ? pnfs_read_pg_test : NULL;
+}
 
-	hlist_for_each_entry_rcu(d, n, &c->dc_deviceids[h], de_node)
-		if (!memcmp(&d->de_id, id, sizeof(*id))) {
-			hlist_del_rcu(&d->de_node);
-			spin_unlock(&c->dc_lock);
-			synchronize_rcu();
-			c->dc_free_callback(devid);
-			return;
-		}
-	spin_unlock(&c->dc_lock);
-	/* Why wasn't it found in  the list? */
-	BUG();
-}
-EXPORT_SYMBOL_GPL(pnfs_put_deviceid);
-
-/* Find and reference a deviceid */
-struct pnfs_deviceid_node *
-pnfs_find_get_deviceid(struct pnfs_deviceid_cache *c, struct nfs4_deviceid *id)
-{
-	struct pnfs_deviceid_node *d;
-	struct hlist_node *n;
-	long hash = nfs4_deviceid_hash(id);
-
-	dprintk("--> %s hash %ld\n", __func__, hash);
-	rcu_read_lock();
-	hlist_for_each_entry_rcu(d, n, &c->dc_deviceids[hash], de_node) {
-		if (!memcmp(&d->de_id, id, sizeof(*id))) {
-			if (!atomic_inc_not_zero(&d->de_ref)) {
-				goto fail;
-			} else {
-				rcu_read_unlock();
-				return d;
-			}
-		}
+static int pnfs_write_pg_test(struct nfs_pageio_descriptor *pgio,
+			      struct nfs_page *prev,
+			      struct nfs_page *req)
+{
+	if (pgio->pg_count == prev->wb_bytes) {
+		/* This is first coelesce call for a series of nfs_pages */
+		pgio->pg_lseg = pnfs_update_layout(pgio->pg_inode,
+						   prev->wb_context,
+						   IOMODE_RW);
 	}
-fail:
-	rcu_read_unlock();
-	return NULL;
+	return NFS_SERVER(pgio->pg_inode)->pnfs_curr_ld->pg_test(pgio, prev, req);
+}
+
+void
+pnfs_pageio_init_write(struct nfs_pageio_descriptor *pgio, struct inode *inode)
+{
+	struct pnfs_layoutdriver_type *ld;
+
+	ld = NFS_SERVER(inode)->pnfs_curr_ld;
+	pgio->pg_test = (ld && ld->pg_test) ? pnfs_write_pg_test : NULL;
+}
+
+enum pnfs_try_status
+pnfs_try_to_write_data(struct nfs_write_data *wdata,
+			const struct rpc_call_ops *call_ops, int how)
+{
+	struct inode *inode = wdata->inode;
+	enum pnfs_try_status trypnfs;
+	struct nfs_server *nfss = NFS_SERVER(inode);
+
+	wdata->mds_ops = call_ops;
+
+	dprintk("%s: Writing ino:%lu %u@%llu (how %d)\n", __func__,
+		inode->i_ino, wdata->args.count, wdata->args.offset, how);
+
+	trypnfs = nfss->pnfs_curr_ld->write_pagelist(wdata, how);
+	if (trypnfs == PNFS_NOT_ATTEMPTED) {
+		put_lseg(wdata->lseg);
+		wdata->lseg = NULL;
+	} else
+		nfs_inc_stats(inode, NFSIOS_PNFS_WRITE);
+
+	dprintk("%s End (trypnfs:%d)\n", __func__, trypnfs);
+	return trypnfs;
 }
-EXPORT_SYMBOL_GPL(pnfs_find_get_deviceid);
 
 /*
- * Add a deviceid to the cache.
- * GETDEVICEINFOs for same deviceid can race. If deviceid is found, discard new
+ * Call the appropriate parallel I/O subsystem read function.
  */
-struct pnfs_deviceid_node *
-pnfs_add_deviceid(struct pnfs_deviceid_cache *c, struct pnfs_deviceid_node *new)
-{
-	struct pnfs_deviceid_node *d;
-	long hash = nfs4_deviceid_hash(&new->de_id);
-
-	dprintk("--> %s hash %ld\n", __func__, hash);
-	spin_lock(&c->dc_lock);
-	d = pnfs_find_get_deviceid(c, &new->de_id);
-	if (d) {
-		spin_unlock(&c->dc_lock);
-		dprintk("%s [discard]\n", __func__);
-		c->dc_free_callback(new);
-		return d;
-	}
-	INIT_HLIST_NODE(&new->de_node);
-	atomic_set(&new->de_ref, 1);
-	hlist_add_head_rcu(&new->de_node, &c->dc_deviceids[hash]);
-	spin_unlock(&c->dc_lock);
-	dprintk("%s [new]\n", __func__);
-	return new;
-}
-EXPORT_SYMBOL_GPL(pnfs_add_deviceid);
-
-void
-pnfs_put_deviceid_cache(struct nfs_client *clp)
+enum pnfs_try_status
+pnfs_try_to_read_data(struct nfs_read_data *rdata,
+		       const struct rpc_call_ops *call_ops)
 {
-	struct pnfs_deviceid_cache *local = clp->cl_devid_cache;
+	struct inode *inode = rdata->inode;
+	struct nfs_server *nfss = NFS_SERVER(inode);
+	enum pnfs_try_status trypnfs;
 
-	dprintk("--> %s ({%d})\n", __func__, atomic_read(&local->dc_ref));
-	if (atomic_dec_and_lock(&local->dc_ref, &clp->cl_lock)) {
-		int i;
-		/* Verify cache is empty */
-		for (i = 0; i < NFS4_DEVICE_ID_HASH_SIZE; i++)
-			BUG_ON(!hlist_empty(&local->dc_deviceids[i]));
-		clp->cl_devid_cache = NULL;
-		spin_unlock(&clp->cl_lock);
-		kfree(local);
+	rdata->mds_ops = call_ops;
+
+	dprintk("%s: Reading ino:%lu %u@%llu\n",
+		__func__, inode->i_ino, rdata->args.count, rdata->args.offset);
+
+	trypnfs = nfss->pnfs_curr_ld->read_pagelist(rdata);
+	if (trypnfs == PNFS_NOT_ATTEMPTED) {
+		put_lseg(rdata->lseg);
+		rdata->lseg = NULL;
+	} else {
+		nfs_inc_stats(inode, NFSIOS_PNFS_READ);
 	}
+	dprintk("%s End (trypnfs:%d)\n", __func__, trypnfs);
+	return trypnfs;
 }
-EXPORT_SYMBOL_GPL(pnfs_put_deviceid_cache);

fs/nfs/pnfs.h

@@ -30,6 +30,8 @@
 #ifndef FS_NFS_PNFS_H
 #define FS_NFS_PNFS_H
 
+#include <linux/nfs_page.h>
+
 enum {
 	NFS_LSEG_VALID = 0,	/* cleared when lseg is recalled/returned */
 	NFS_LSEG_ROC,		/* roc bit received from server */
@@ -43,6 +45,11 @@ struct pnfs_layout_segment {
 	struct pnfs_layout_hdr *pls_layout;
 };
 
+enum pnfs_try_status {
+	PNFS_ATTEMPTED     = 0,
+	PNFS_NOT_ATTEMPTED = 1,
+};
+
 #ifdef CONFIG_NFS_V4_1
 
 #define LAYOUT_NFSV4_1_MODULE_PREFIX "nfs-layouttype4"
@@ -61,10 +68,18 @@ struct pnfs_layoutdriver_type {
 	const u32 id;
 	const char *name;
 	struct module *owner;
-	int (*set_layoutdriver) (struct nfs_server *);
-	int (*clear_layoutdriver) (struct nfs_server *);
 	struct pnfs_layout_segment * (*alloc_lseg) (struct pnfs_layout_hdr *layoutid, struct nfs4_layoutget_res *lgr);
 	void (*free_lseg) (struct pnfs_layout_segment *lseg);
+
+	/* test for nfs page cache coalescing */
+	int (*pg_test)(struct nfs_pageio_descriptor *, struct nfs_page *, struct nfs_page *);
+
+	/*
+	 * Return PNFS_ATTEMPTED to indicate the layout code has attempted
+	 * I/O, else return PNFS_NOT_ATTEMPTED to fall back to normal NFS
+	 */
+	enum pnfs_try_status (*read_pagelist) (struct nfs_read_data *nfs_data);
+	enum pnfs_try_status (*write_pagelist) (struct nfs_write_data *nfs_data, int how);
 };
 
 struct pnfs_layout_hdr {
@@ -90,52 +105,6 @@ struct pnfs_device {
 	unsigned int  pglen;
 };
 
-/*
- * Device ID RCU cache. A device ID is unique per client ID and layout type.
- */
-#define NFS4_DEVICE_ID_HASH_BITS	5
-#define NFS4_DEVICE_ID_HASH_SIZE	(1 << NFS4_DEVICE_ID_HASH_BITS)
-#define NFS4_DEVICE_ID_HASH_MASK	(NFS4_DEVICE_ID_HASH_SIZE - 1)
-
-static inline u32
-nfs4_deviceid_hash(struct nfs4_deviceid *id)
-{
-	unsigned char *cptr = (unsigned char *)id->data;
-	unsigned int nbytes = NFS4_DEVICEID4_SIZE;
-	u32 x = 0;
-
-	while (nbytes--) {
-		x *= 37;
-		x += *cptr++;
-	}
-	return x & NFS4_DEVICE_ID_HASH_MASK;
-}
-
-struct pnfs_deviceid_node {
-	struct hlist_node	de_node;
-	struct nfs4_deviceid	de_id;
-	atomic_t		de_ref;
-};
-
-struct pnfs_deviceid_cache {
-	spinlock_t		dc_lock;
-	atomic_t		dc_ref;
-	void			(*dc_free_callback)(struct pnfs_deviceid_node *);
-	struct hlist_head	dc_deviceids[NFS4_DEVICE_ID_HASH_SIZE];
-};
-
-extern int pnfs_alloc_init_deviceid_cache(struct nfs_client *,
-			void (*free_callback)(struct pnfs_deviceid_node *));
-extern void pnfs_put_deviceid_cache(struct nfs_client *);
-extern struct pnfs_deviceid_node *pnfs_find_get_deviceid(
-				struct pnfs_deviceid_cache *,
-				struct nfs4_deviceid *);
-extern struct pnfs_deviceid_node *pnfs_add_deviceid(
-				struct pnfs_deviceid_cache *,
-				struct pnfs_deviceid_node *);
-extern void pnfs_put_deviceid(struct pnfs_deviceid_cache *c,
-			      struct pnfs_deviceid_node *devid);
-
 extern int pnfs_register_layoutdriver(struct pnfs_layoutdriver_type *);
 extern void pnfs_unregister_layoutdriver(struct pnfs_layoutdriver_type *);
 
@@ -146,11 +115,18 @@ extern int nfs4_proc_layoutget(struct nfs4_layoutget *lgp);
 
 /* pnfs.c */
 void get_layout_hdr(struct pnfs_layout_hdr *lo);
+void put_lseg(struct pnfs_layout_segment *lseg);
 struct pnfs_layout_segment *
 pnfs_update_layout(struct inode *ino, struct nfs_open_context *ctx,
 		   enum pnfs_iomode access_type);
 void set_pnfs_layoutdriver(struct nfs_server *, u32 id);
 void unset_pnfs_layoutdriver(struct nfs_server *);
+enum pnfs_try_status pnfs_try_to_write_data(struct nfs_write_data *,
+					     const struct rpc_call_ops *, int);
+enum pnfs_try_status pnfs_try_to_read_data(struct nfs_read_data *,
+					    const struct rpc_call_ops *);
+void pnfs_pageio_init_read(struct nfs_pageio_descriptor *, struct inode *);
+void pnfs_pageio_init_write(struct nfs_pageio_descriptor *, struct inode *);
 int pnfs_layout_process(struct nfs4_layoutget *lgp);
 void pnfs_free_lseg_list(struct list_head *tmp_list);
 void pnfs_destroy_layout(struct nfs_inode *);
@@ -177,6 +153,16 @@ static inline int lo_fail_bit(u32 iomode)
 			 NFS_LAYOUT_RW_FAILED : NFS_LAYOUT_RO_FAILED;
 }
 
+static inline struct pnfs_layout_segment *
+get_lseg(struct pnfs_layout_segment *lseg)
+{
+	if (lseg) {
+		atomic_inc(&lseg->pls_refcount);
+		smp_mb__after_atomic_inc();
+	}
+	return lseg;
+}
+
 /* Return true if a layout driver is being used for this mountpoint */
 static inline int pnfs_enabled_sb(struct nfs_server *nfss)
 {
@@ -193,6 +179,16 @@ static inline void pnfs_destroy_layout(struct nfs_inode *nfsi)
 {
 }
 
+static inline struct pnfs_layout_segment *
+get_lseg(struct pnfs_layout_segment *lseg)
+{
+	return NULL;
+}
+
+static inline void put_lseg(struct pnfs_layout_segment *lseg)
+{
+}
+
 static inline struct pnfs_layout_segment *
 pnfs_update_layout(struct inode *ino, struct nfs_open_context *ctx,
 		   enum pnfs_iomode access_type)
@@ -200,6 +196,20 @@ pnfs_update_layout(struct inode *ino, struct nfs_open_context *ctx,
 	return NULL;
 }
 
+static inline enum pnfs_try_status
+pnfs_try_to_read_data(struct nfs_read_data *data,
+		      const struct rpc_call_ops *call_ops)
+{
+	return PNFS_NOT_ATTEMPTED;
+}
+
+static inline enum pnfs_try_status
+pnfs_try_to_write_data(struct nfs_write_data *data,
+		       const struct rpc_call_ops *call_ops, int how)
+{
+	return PNFS_NOT_ATTEMPTED;
+}
+
 static inline bool
 pnfs_roc(struct inode *ino)
 {
@@ -230,6 +240,18 @@ static inline void unset_pnfs_layoutdriver(struct nfs_server *s)
 {
 }
 
+static inline void
+pnfs_pageio_init_read(struct nfs_pageio_descriptor *pgio, struct inode *ino)
+{
+	pgio->pg_test = NULL;
+}
+
+static inline void
+pnfs_pageio_init_write(struct nfs_pageio_descriptor *pgio, struct inode *ino)
+{
+	pgio->pg_test = NULL;
+}
+
 #endif /* CONFIG_NFS_V4_1 */
 
 #endif /* FS_NFS_PNFS_H */

fs/nfs/proc.c

@@ -741,4 +741,5 @@ const struct nfs_rpc_ops nfs_v2_clientops = {
 	.lock		= nfs_proc_lock,
 	.lock_check_bounds = nfs_lock_check_bounds,
 	.close_context	= nfs_close_context,
+	.init_client	= nfs_init_client,
 };

fs/nfs/read.c

@@ -18,19 +18,20 @@
 #include <linux/sunrpc/clnt.h>
 #include <linux/nfs_fs.h>
 #include <linux/nfs_page.h>
+#include <linux/module.h>
 
 #include <asm/system.h>
+#include "pnfs.h"
 
 #include "nfs4_fs.h"
 #include "internal.h"
 #include "iostat.h"
 #include "fscache.h"
-#include "pnfs.h"
 
 #define NFSDBG_FACILITY		NFSDBG_PAGECACHE
 
-static int nfs_pagein_multi(struct inode *, struct list_head *, unsigned int, size_t, int);
-static int nfs_pagein_one(struct inode *, struct list_head *, unsigned int, size_t, int);
+static int nfs_pagein_multi(struct nfs_pageio_descriptor *desc);
+static int nfs_pagein_one(struct nfs_pageio_descriptor *desc);
 static const struct rpc_call_ops nfs_read_partial_ops;
 static const struct rpc_call_ops nfs_read_full_ops;
 
@@ -69,6 +70,7 @@ void nfs_readdata_free(struct nfs_read_data *p)
 
 static void nfs_readdata_release(struct nfs_read_data *rdata)
 {
+	put_lseg(rdata->lseg);
 	put_nfs_open_context(rdata->args.context);
 	nfs_readdata_free(rdata);
 }
@@ -114,14 +116,13 @@ static void nfs_readpage_truncate_uninitialised_page(struct nfs_read_data *data)
 int nfs_readpage_async(struct nfs_open_context *ctx, struct inode *inode,
 		       struct page *page)
 {
-	LIST_HEAD(one_request);
 	struct nfs_page	*new;
 	unsigned int len;
+	struct nfs_pageio_descriptor pgio;
 
 	len = nfs_page_length(page);
 	if (len == 0)
 		return nfs_return_empty_page(page);
-	pnfs_update_layout(inode, ctx, IOMODE_READ);
 	new = nfs_create_request(ctx, inode, page, 0, len);
 	if (IS_ERR(new)) {
 		unlock_page(page);
@@ -130,11 +131,14 @@ int nfs_readpage_async(struct nfs_open_context *ctx, struct inode *inode,
 	if (len < PAGE_CACHE_SIZE)
 		zero_user_segment(page, len, PAGE_CACHE_SIZE);
 
-	nfs_list_add_request(new, &one_request);
+	nfs_pageio_init(&pgio, inode, NULL, 0, 0);
+	nfs_list_add_request(new, &pgio.pg_list);
+	pgio.pg_count = len;
+
 	if (NFS_SERVER(inode)->rsize < PAGE_CACHE_SIZE)
-		nfs_pagein_multi(inode, &one_request, 1, len, 0);
+		nfs_pagein_multi(&pgio);
 	else
-		nfs_pagein_one(inode, &one_request, 1, len, 0);
+		nfs_pagein_one(&pgio);
 	return 0;
 }
 
@@ -155,24 +159,20 @@ static void nfs_readpage_release(struct nfs_page *req)
 	nfs_release_request(req);
 }
 
-/*
- * Set up the NFS read request struct
- */
-static int nfs_read_rpcsetup(struct nfs_page *req, struct nfs_read_data *data,
-		const struct rpc_call_ops *call_ops,
-		unsigned int count, unsigned int offset)
+int nfs_initiate_read(struct nfs_read_data *data, struct rpc_clnt *clnt,
+		      const struct rpc_call_ops *call_ops)
 {
-	struct inode *inode = req->wb_context->path.dentry->d_inode;
+	struct inode *inode = data->inode;
 	int swap_flags = IS_SWAPFILE(inode) ? NFS_RPC_SWAPFLAGS : 0;
 	struct rpc_task *task;
 	struct rpc_message msg = {
 		.rpc_argp = &data->args,
 		.rpc_resp = &data->res,
-		.rpc_cred = req->wb_context->cred,
+		.rpc_cred = data->cred,
 	};
 	struct rpc_task_setup task_setup_data = {
 		.task = &data->task,
-		.rpc_client = NFS_CLIENT(inode),
+		.rpc_client = clnt,
 		.rpc_message = &msg,
 		.callback_ops = call_ops,
 		.callback_data = data,
@@ -180,9 +180,39 @@ static int nfs_read_rpcsetup(struct nfs_page *req, struct nfs_read_data *data,
 		.flags = RPC_TASK_ASYNC | swap_flags,
 	};
 
+	/* Set up the initial task struct. */
+	NFS_PROTO(inode)->read_setup(data, &msg);
+
+	dprintk("NFS: %5u initiated read call (req %s/%lld, %u bytes @ "
+			"offset %llu)\n",
+			data->task.tk_pid,
+			inode->i_sb->s_id,
+			(long long)NFS_FILEID(inode),
+			data->args.count,
+			(unsigned long long)data->args.offset);
+
+	task = rpc_run_task(&task_setup_data);
+	if (IS_ERR(task))
+		return PTR_ERR(task);
+	rpc_put_task(task);
+	return 0;
+}
+EXPORT_SYMBOL_GPL(nfs_initiate_read);
+
+/*
+ * Set up the NFS read request struct
+ */
+static int nfs_read_rpcsetup(struct nfs_page *req, struct nfs_read_data *data,
+		const struct rpc_call_ops *call_ops,
+		unsigned int count, unsigned int offset,
+		struct pnfs_layout_segment *lseg)
+{
+	struct inode *inode = req->wb_context->path.dentry->d_inode;
+
 	data->req	  = req;
 	data->inode	  = inode;
-	data->cred	  = msg.rpc_cred;
+	data->cred	  = req->wb_context->cred;
+	data->lseg	  = get_lseg(lseg);
 
 	data->args.fh     = NFS_FH(inode);
 	data->args.offset = req_offset(req) + offset;
@@ -197,21 +227,11 @@ static int nfs_read_rpcsetup(struct nfs_page *req, struct nfs_read_data *data,
 	data->res.eof     = 0;
 	nfs_fattr_init(&data->fattr);
 
-	/* Set up the initial task struct. */
-	NFS_PROTO(inode)->read_setup(data, &msg);
-
-	dprintk("NFS: %5u initiated read call (req %s/%Ld, %u bytes @ offset %Lu)\n",
-			data->task.tk_pid,
-			inode->i_sb->s_id,
-			(long long)NFS_FILEID(inode),
-			count,
-			(unsigned long long)data->args.offset);
+	if (data->lseg &&
+	    (pnfs_try_to_read_data(data, call_ops) == PNFS_ATTEMPTED))
+		return 0;
 
-	task = rpc_run_task(&task_setup_data);
-	if (IS_ERR(task))
-		return PTR_ERR(task);
-	rpc_put_task(task);
-	return 0;
+	return nfs_initiate_read(data, NFS_CLIENT(inode), call_ops);
 }
 
 static void
@@ -240,20 +260,21 @@ nfs_async_read_error(struct list_head *head)
  * won't see the new data until our attribute cache is updated.  This is more
  * or less conventional NFS client behavior.
  */
-static int nfs_pagein_multi(struct inode *inode, struct list_head *head, unsigned int npages, size_t count, int flags)
+static int nfs_pagein_multi(struct nfs_pageio_descriptor *desc)
 {
-	struct nfs_page *req = nfs_list_entry(head->next);
+	struct nfs_page *req = nfs_list_entry(desc->pg_list.next);
 	struct page *page = req->wb_page;
 	struct nfs_read_data *data;
-	size_t rsize = NFS_SERVER(inode)->rsize, nbytes;
+	size_t rsize = NFS_SERVER(desc->pg_inode)->rsize, nbytes;
 	unsigned int offset;
 	int requests = 0;
 	int ret = 0;
+	struct pnfs_layout_segment *lseg;
 	LIST_HEAD(list);
 
 	nfs_list_remove_request(req);
 
-	nbytes = count;
+	nbytes = desc->pg_count;
 	do {
 		size_t len = min(nbytes,rsize);
 
@@ -266,9 +287,11 @@ static int nfs_pagein_multi(struct inode *inode, struct list_head *head, unsigne
 	} while(nbytes != 0);
 	atomic_set(&req->wb_complete, requests);
 
+	BUG_ON(desc->pg_lseg != NULL);
+	lseg = pnfs_update_layout(desc->pg_inode, req->wb_context, IOMODE_READ);
 	ClearPageError(page);
 	offset = 0;
-	nbytes = count;
+	nbytes = desc->pg_count;
 	do {
 		int ret2;
 
@@ -280,12 +303,14 @@ static int nfs_pagein_multi(struct inode *inode, struct list_head *head, unsigne
 		if (nbytes < rsize)
 			rsize = nbytes;
 		ret2 = nfs_read_rpcsetup(req, data, &nfs_read_partial_ops,
-				  rsize, offset);
+					 rsize, offset, lseg);
 		if (ret == 0)
 			ret = ret2;
 		offset += rsize;
 		nbytes -= rsize;
 	} while (nbytes != 0);
+	put_lseg(lseg);
+	desc->pg_lseg = NULL;
 
 	return ret;
 
@@ -300,16 +325,21 @@ out_bad:
 	return -ENOMEM;
 }
 
-static int nfs_pagein_one(struct inode *inode, struct list_head *head, unsigned int npages, size_t count, int flags)
+static int nfs_pagein_one(struct nfs_pageio_descriptor *desc)
 {
 	struct nfs_page		*req;
 	struct page		**pages;
 	struct nfs_read_data	*data;
+	struct list_head *head = &desc->pg_list;
+	struct pnfs_layout_segment *lseg = desc->pg_lseg;
 	int ret = -ENOMEM;
 
-	data = nfs_readdata_alloc(npages);
-	if (!data)
-		goto out_bad;
+	data = nfs_readdata_alloc(nfs_page_array_len(desc->pg_base,
+						     desc->pg_count));
+	if (!data) {
+		nfs_async_read_error(head);
+		goto out;
+	}
 
 	pages = data->pagevec;
 	while (!list_empty(head)) {
@@ -320,10 +350,14 @@ static int nfs_pagein_one(struct inode *inode, struct list_head *head, unsigned
 		*pages++ = req->wb_page;
 	}
 	req = nfs_list_entry(data->pages.next);
+	if ((!lseg) && list_is_singular(&data->pages))
+		lseg = pnfs_update_layout(desc->pg_inode, req->wb_context, IOMODE_READ);
 
-	return nfs_read_rpcsetup(req, data, &nfs_read_full_ops, count, 0);
-out_bad:
-	nfs_async_read_error(head);
+	ret = nfs_read_rpcsetup(req, data, &nfs_read_full_ops, desc->pg_count,
+				0, lseg);
+out:
+	put_lseg(lseg);
+	desc->pg_lseg = NULL;
 	return ret;
 }
 
@@ -366,6 +400,7 @@ static void nfs_readpage_retry(struct rpc_task *task, struct nfs_read_data *data
 		return;
 
 	/* Yes, so retry the read at the end of the data */
+	data->mds_offset += resp->count;
 	argp->offset += resp->count;
 	argp->pgbase += resp->count;
 	argp->count -= resp->count;
@@ -625,7 +660,7 @@ int nfs_readpages(struct file *filp, struct address_space *mapping,
 	if (ret == 0)
 		goto read_complete; /* all pages were read */
 
-	pnfs_update_layout(inode, desc.ctx, IOMODE_READ);
+	pnfs_pageio_init_read(&pgio, inode);
 	if (rsize < PAGE_CACHE_SIZE)
 		nfs_pageio_init(&pgio, inode, nfs_pagein_multi, rsize, 0);
 	else

fs/nfs/super.c

@@ -1008,6 +1008,27 @@ static int nfs_parse_security_flavors(char *value,
 	return 1;
 }
 
+static int nfs_get_option_str(substring_t args[], char **option)
+{
+	kfree(*option);
+	*option = match_strdup(args);
+	return !option;
+}
+
+static int nfs_get_option_ul(substring_t args[], unsigned long *option)
+{
+	int rc;
+	char *string;
+
+	string = match_strdup(args);
+	if (string == NULL)
+		return -ENOMEM;
+	rc = strict_strtoul(string, 10, option);
+	kfree(string);
+
+	return rc;
+}
+
 /*
  * Error-check and convert a string of mount options from user space into
  * a data structure.  The whole mount string is processed; bad options are
@@ -1156,155 +1177,82 @@ static int nfs_parse_mount_options(char *raw,
 		 * options that take numeric values
 		 */
 		case Opt_port:
-			string = match_strdup(args);
-			if (string == NULL)
-				goto out_nomem;
-			rc = strict_strtoul(string, 10, &option);
-			kfree(string);
-			if (rc != 0 || option > USHRT_MAX)
+			if (nfs_get_option_ul(args, &option) ||
+			    option > USHRT_MAX)
 				goto out_invalid_value;
 			mnt->nfs_server.port = option;
 			break;
 		case Opt_rsize:
-			string = match_strdup(args);
-			if (string == NULL)
-				goto out_nomem;
-			rc = strict_strtoul(string, 10, &option);
-			kfree(string);
-			if (rc != 0)
+			if (nfs_get_option_ul(args, &option))
 				goto out_invalid_value;
 			mnt->rsize = option;
 			break;
 		case Opt_wsize:
-			string = match_strdup(args);
-			if (string == NULL)
-				goto out_nomem;
-			rc = strict_strtoul(string, 10, &option);
-			kfree(string);
-			if (rc != 0)
+			if (nfs_get_option_ul(args, &option))
 				goto out_invalid_value;
 			mnt->wsize = option;
 			break;
 		case Opt_bsize:
-			string = match_strdup(args);
-			if (string == NULL)
-				goto out_nomem;
-			rc = strict_strtoul(string, 10, &option);
-			kfree(string);
-			if (rc != 0)
+			if (nfs_get_option_ul(args, &option))
 				goto out_invalid_value;
 			mnt->bsize = option;
 			break;
 		case Opt_timeo:
-			string = match_strdup(args);
-			if (string == NULL)
-				goto out_nomem;
-			rc = strict_strtoul(string, 10, &option);
-			kfree(string);
-			if (rc != 0 || option == 0)
+			if (nfs_get_option_ul(args, &option) || option == 0)
 				goto out_invalid_value;
 			mnt->timeo = option;
 			break;
 		case Opt_retrans:
-			string = match_strdup(args);
-			if (string == NULL)
-				goto out_nomem;
-			rc = strict_strtoul(string, 10, &option);
-			kfree(string);
-			if (rc != 0 || option == 0)
+			if (nfs_get_option_ul(args, &option) || option == 0)
 				goto out_invalid_value;
 			mnt->retrans = option;
 			break;
 		case Opt_acregmin:
-			string = match_strdup(args);
-			if (string == NULL)
-				goto out_nomem;
-			rc = strict_strtoul(string, 10, &option);
-			kfree(string);
-			if (rc != 0)
+			if (nfs_get_option_ul(args, &option))
 				goto out_invalid_value;
 			mnt->acregmin = option;
 			break;
 		case Opt_acregmax:
-			string = match_strdup(args);
-			if (string == NULL)
-				goto out_nomem;
-			rc = strict_strtoul(string, 10, &option);
-			kfree(string);
-			if (rc != 0)
+			if (nfs_get_option_ul(args, &option))
 				goto out_invalid_value;
 			mnt->acregmax = option;
 			break;
 		case Opt_acdirmin:
-			string = match_strdup(args);
-			if (string == NULL)
-				goto out_nomem;
-			rc = strict_strtoul(string, 10, &option);
-			kfree(string);
-			if (rc != 0)
+			if (nfs_get_option_ul(args, &option))
 				goto out_invalid_value;
 			mnt->acdirmin = option;
 			break;
 		case Opt_acdirmax:
-			string = match_strdup(args);
-			if (string == NULL)
-				goto out_nomem;
-			rc = strict_strtoul(string, 10, &option);
-			kfree(string);
-			if (rc != 0)
+			if (nfs_get_option_ul(args, &option))
 				goto out_invalid_value;
 			mnt->acdirmax = option;
 			break;
 		case Opt_actimeo:
-			string = match_strdup(args);
-			if (string == NULL)
-				goto out_nomem;
-			rc = strict_strtoul(string, 10, &option);
-			kfree(string);
-			if (rc != 0)
+			if (nfs_get_option_ul(args, &option))
 				goto out_invalid_value;
 			mnt->acregmin = mnt->acregmax =
 			mnt->acdirmin = mnt->acdirmax = option;
 			break;
 		case Opt_namelen:
-			string = match_strdup(args);
-			if (string == NULL)
-				goto out_nomem;
-			rc = strict_strtoul(string, 10, &option);
-			kfree(string);
-			if (rc != 0)
+			if (nfs_get_option_ul(args, &option))
 				goto out_invalid_value;
 			mnt->namlen = option;
 			break;
 		case Opt_mountport:
-			string = match_strdup(args);
-			if (string == NULL)
-				goto out_nomem;
-			rc = strict_strtoul(string, 10, &option);
-			kfree(string);
-			if (rc != 0 || option > USHRT_MAX)
+			if (nfs_get_option_ul(args, &option) ||
+			    option > USHRT_MAX)
 				goto out_invalid_value;
 			mnt->mount_server.port = option;
 			break;
 		case Opt_mountvers:
-			string = match_strdup(args);
-			if (string == NULL)
-				goto out_nomem;
-			rc = strict_strtoul(string, 10, &option);
-			kfree(string);
-			if (rc != 0 ||
+			if (nfs_get_option_ul(args, &option) ||
 			    option < NFS_MNT_VERSION ||
 			    option > NFS_MNT3_VERSION)
 				goto out_invalid_value;
 			mnt->mount_server.version = option;
 			break;
 		case Opt_nfsvers:
-			string = match_strdup(args);
-			if (string == NULL)
-				goto out_nomem;
-			rc = strict_strtoul(string, 10, &option);
-			kfree(string);
-			if (rc != 0)
+			if (nfs_get_option_ul(args, &option))
 				goto out_invalid_value;
 			switch (option) {
 			case NFS2_VERSION:
@@ -1324,12 +1272,7 @@ static int nfs_parse_mount_options(char *raw,
 			}
 			break;
 		case Opt_minorversion:
-			string = match_strdup(args);
-			if (string == NULL)
-				goto out_nomem;
-			rc = strict_strtoul(string, 10, &option);
-			kfree(string);
-			if (rc != 0)
+			if (nfs_get_option_ul(args, &option))
 				goto out_invalid_value;
 			if (option > NFS4_MAX_MINOR_VERSION)
 				goto out_invalid_value;
@@ -1365,21 +1308,18 @@ static int nfs_parse_mount_options(char *raw,
 			case Opt_xprt_udp:
 				mnt->flags &= ~NFS_MOUNT_TCP;
 				mnt->nfs_server.protocol = XPRT_TRANSPORT_UDP;
-				kfree(string);
 				break;
 			case Opt_xprt_tcp6:
 				protofamily = AF_INET6;
 			case Opt_xprt_tcp:
 				mnt->flags |= NFS_MOUNT_TCP;
 				mnt->nfs_server.protocol = XPRT_TRANSPORT_TCP;
-				kfree(string);
 				break;
 			case Opt_xprt_rdma:
 				/* vector side protocols to TCP */
 				mnt->flags |= NFS_MOUNT_TCP;
 				mnt->nfs_server.protocol = XPRT_TRANSPORT_RDMA;
 				xprt_load_transport(string);
-				kfree(string);
 				break;
 			default:
 				dfprintk(MOUNT, "NFS:   unrecognized "
@@ -1387,6 +1327,7 @@ static int nfs_parse_mount_options(char *raw,
 				kfree(string);
 				return 0;
 			}
+			kfree(string);
 			break;
 		case Opt_mountproto:
 			string = match_strdup(args);
@@ -1429,18 +1370,13 @@ static int nfs_parse_mount_options(char *raw,
 				goto out_invalid_address;
 			break;
 		case Opt_clientaddr:
-			string = match_strdup(args);
-			if (string == NULL)
+			if (nfs_get_option_str(args, &mnt->client_address))
 				goto out_nomem;
-			kfree(mnt->client_address);
-			mnt->client_address = string;
 			break;
 		case Opt_mounthost:
-			string = match_strdup(args);
-			if (string == NULL)
+			if (nfs_get_option_str(args,
+					       &mnt->mount_server.hostname))
 				goto out_nomem;
-			kfree(mnt->mount_server.hostname);
-			mnt->mount_server.hostname = string;
 			break;
 		case Opt_mountaddr:
 			string = match_strdup(args);
@@ -1480,11 +1416,8 @@ static int nfs_parse_mount_options(char *raw,
 			};
 			break;
 		case Opt_fscache_uniq:
-			string = match_strdup(args);
-			if (string == NULL)
+			if (nfs_get_option_str(args, &mnt->fscache_uniq))
 				goto out_nomem;
-			kfree(mnt->fscache_uniq);
-			mnt->fscache_uniq = string;
 			mnt->options |= NFS_OPTION_FSCACHE;
 			break;
 		case Opt_local_lock:
@@ -1694,99 +1627,59 @@ static int nfs_try_mount(struct nfs_parsed_mount_data *args,
 	return nfs_walk_authlist(args, &request);
 }
 
-static int nfs_parse_simple_hostname(const char *dev_name,
-				     char **hostname, size_t maxnamlen,
-				     char **export_path, size_t maxpathlen)
+/*
+ * Split "dev_name" into "hostname:export_path".
+ *
+ * The leftmost colon demarks the split between the server's hostname
+ * and the export path.  If the hostname starts with a left square
+ * bracket, then it may contain colons.
+ *
+ * Note: caller frees hostname and export path, even on error.
+ */
+static int nfs_parse_devname(const char *dev_name,
+			     char **hostname, size_t maxnamlen,
+			     char **export_path, size_t maxpathlen)
 {
 	size_t len;
-	char *colon, *comma;
+	char *end;
 
-	colon = strchr(dev_name, ':');
-	if (colon == NULL)
-		goto out_bad_devname;
-
-	len = colon - dev_name;
-	if (len > maxnamlen)
-		goto out_hostname;
-
-	/* N.B. caller will free nfs_server.hostname in all cases */
-	*hostname = kstrndup(dev_name, len, GFP_KERNEL);
-	if (!*hostname)
-		goto out_nomem;
-
-	/* kill possible hostname list: not supported */
-	comma = strchr(*hostname, ',');
-	if (comma != NULL) {
-		if (comma == *hostname)
+	/* Is the host name protected with square brakcets? */
+	if (*dev_name == '[') {
+		end = strchr(++dev_name, ']');
+		if (end == NULL || end[1] != ':')
 			goto out_bad_devname;
-		*comma = '\0';
-	}
-
-	colon++;
-	len = strlen(colon);
-	if (len > maxpathlen)
-		goto out_path;
-	*export_path = kstrndup(colon, len, GFP_KERNEL);
-	if (!*export_path)
-		goto out_nomem;
-
-	dfprintk(MOUNT, "NFS: MNTPATH: '%s'\n", *export_path);
-	return 0;
-
-out_bad_devname:
-	dfprintk(MOUNT, "NFS: device name not in host:path format\n");
-	return -EINVAL;
 
-out_nomem:
-	dfprintk(MOUNT, "NFS: not enough memory to parse device name\n");
-	return -ENOMEM;
-
-out_hostname:
-	dfprintk(MOUNT, "NFS: server hostname too long\n");
-	return -ENAMETOOLONG;
-
-out_path:
-	dfprintk(MOUNT, "NFS: export pathname too long\n");
-	return -ENAMETOOLONG;
-}
-
-/*
- * Hostname has square brackets around it because it contains one or
- * more colons.  We look for the first closing square bracket, and a
- * colon must follow it.
- */
-static int nfs_parse_protected_hostname(const char *dev_name,
-					char **hostname, size_t maxnamlen,
-					char **export_path, size_t maxpathlen)
-{
-	size_t len;
-	char *start, *end;
+		len = end - dev_name;
+		end++;
+	} else {
+		char *comma;
 
-	start = (char *)(dev_name + 1);
+		end = strchr(dev_name, ':');
+		if (end == NULL)
+			goto out_bad_devname;
+		len = end - dev_name;
 
-	end = strchr(start, ']');
-	if (end == NULL)
-		goto out_bad_devname;
-	if (*(end + 1) != ':')
-		goto out_bad_devname;
+		/* kill possible hostname list: not supported */
+		comma = strchr(dev_name, ',');
+		if (comma != NULL && comma < end)
+			*comma = 0;
+	}
 
-	len = end - start;
 	if (len > maxnamlen)
 		goto out_hostname;
 
 	/* N.B. caller will free nfs_server.hostname in all cases */
-	*hostname = kstrndup(start, len, GFP_KERNEL);
+	*hostname = kstrndup(dev_name, len, GFP_KERNEL);
 	if (*hostname == NULL)
 		goto out_nomem;
-
-	end += 2;
-	len = strlen(end);
+	len = strlen(++end);
 	if (len > maxpathlen)
 		goto out_path;
 	*export_path = kstrndup(end, len, GFP_KERNEL);
 	if (!*export_path)
 		goto out_nomem;
 
+	dfprintk(MOUNT, "NFS: MNTPATH: '%s'\n", *export_path);
 	return 0;
 
 out_bad_devname:
@@ -1806,29 +1699,6 @@ out_path:
 	return -ENAMETOOLONG;
 }
 
-/*
- * Split "dev_name" into "hostname:export_path".
- *
- * The leftmost colon demarks the split between the server's hostname
- * and the export path.  If the hostname starts with a left square
- * bracket, then it may contain colons.
- *
- * Note: caller frees hostname and export path, even on error.
- */
-static int nfs_parse_devname(const char *dev_name,
-			     char **hostname, size_t maxnamlen,
-			     char **export_path, size_t maxpathlen)
-{
-	if (*dev_name == '[')
-		return nfs_parse_protected_hostname(dev_name,
-						    hostname, maxnamlen,
-						    export_path, maxpathlen);
-
-	return nfs_parse_simple_hostname(dev_name,
-					 hostname, maxnamlen,
-					 export_path, maxpathlen);
-}
-
 /*
  * Validate the NFS2/NFS3 mount data
  * - fills in the mount root filehandle

fs/nfs/write.c

@@ -28,6 +28,7 @@
 #include "iostat.h"
 #include "nfs4_fs.h"
 #include "fscache.h"
+#include "pnfs.h"
 
 #define NFSDBG_FACILITY		NFSDBG_PAGECACHE
 
@@ -96,6 +97,7 @@ void nfs_writedata_free(struct nfs_write_data *p)
 
 static void nfs_writedata_release(struct nfs_write_data *wdata)
 {
+	put_lseg(wdata->lseg);
 	put_nfs_open_context(wdata->args.context);
 	nfs_writedata_free(wdata);
 }
@@ -781,25 +783,21 @@ static int flush_task_priority(int how)
 	return RPC_PRIORITY_NORMAL;
 }
 
-/*
- * Set up the argument/result storage required for the RPC call.
- */
-static int nfs_write_rpcsetup(struct nfs_page *req,
-		struct nfs_write_data *data,
-		const struct rpc_call_ops *call_ops,
-		unsigned int count, unsigned int offset,
-		int how)
+int nfs_initiate_write(struct nfs_write_data *data,
+		       struct rpc_clnt *clnt,
+		       const struct rpc_call_ops *call_ops,
+		       int how)
 {
-	struct inode *inode = req->wb_context->path.dentry->d_inode;
+	struct inode *inode = data->inode;
 	int priority = flush_task_priority(how);
 	struct rpc_task *task;
 	struct rpc_message msg = {
 		.rpc_argp = &data->args,
 		.rpc_resp = &data->res,
-		.rpc_cred = req->wb_context->cred,
+		.rpc_cred = data->cred,
 	};
 	struct rpc_task_setup task_setup_data = {
-		.rpc_client = NFS_CLIENT(inode),
+		.rpc_client = clnt,
 		.task = &data->task,
 		.rpc_message = &msg,
 		.callback_ops = call_ops,
@@ -810,12 +808,52 @@ static int nfs_write_rpcsetup(struct nfs_page *req,
 	};
 	int ret = 0;
 
+	/* Set up the initial task struct.  */
+	NFS_PROTO(inode)->write_setup(data, &msg);
+
+	dprintk("NFS: %5u initiated write call "
+		"(req %s/%lld, %u bytes @ offset %llu)\n",
+		data->task.tk_pid,
+		inode->i_sb->s_id,
+		(long long)NFS_FILEID(inode),
+		data->args.count,
+		(unsigned long long)data->args.offset);
+
+	task = rpc_run_task(&task_setup_data);
+	if (IS_ERR(task)) {
+		ret = PTR_ERR(task);
+		goto out;
+	}
+	if (how & FLUSH_SYNC) {
+		ret = rpc_wait_for_completion_task(task);
+		if (ret == 0)
+			ret = task->tk_status;
+	}
+	rpc_put_task(task);
+out:
+	return ret;
+}
+EXPORT_SYMBOL_GPL(nfs_initiate_write);
+
+/*
+ * Set up the argument/result storage required for the RPC call.
+ */
+static int nfs_write_rpcsetup(struct nfs_page *req,
+		struct nfs_write_data *data,
+		const struct rpc_call_ops *call_ops,
+		unsigned int count, unsigned int offset,
+		struct pnfs_layout_segment *lseg,
+		int how)
+{
+	struct inode *inode = req->wb_context->path.dentry->d_inode;
+
 	/* Set up the RPC argument and reply structs
 	 * NB: take care not to mess about with data->commit et al. */
 
 	data->req = req;
 	data->inode = inode = req->wb_context->path.dentry->d_inode;
-	data->cred = msg.rpc_cred;
+	data->cred = req->wb_context->cred;
+	data->lseg = get_lseg(lseg);
 
 	data->args.fh     = NFS_FH(inode);
 	data->args.offset = req_offset(req) + offset;
@@ -836,30 +874,11 @@ static int nfs_write_rpcsetup(struct nfs_page *req,
 	data->res.verf    = &data->verf;
 	nfs_fattr_init(&data->fattr);
 
-	/* Set up the initial task struct.  */
-	NFS_PROTO(inode)->write_setup(data, &msg);
-
-	dprintk("NFS: %5u initiated write call "
-		"(req %s/%lld, %u bytes @ offset %llu)\n",
-		data->task.tk_pid,
-		inode->i_sb->s_id,
-		(long long)NFS_FILEID(inode),
-		count,
-		(unsigned long long)data->args.offset);
+	if (data->lseg &&
+	    (pnfs_try_to_write_data(data, call_ops, how) == PNFS_ATTEMPTED))
+		return 0;
 
-	task = rpc_run_task(&task_setup_data);
-	if (IS_ERR(task)) {
-		ret = PTR_ERR(task);
-		goto out;
-	}
-	if (how & FLUSH_SYNC) {
-		ret = rpc_wait_for_completion_task(task);
-		if (ret == 0)
-			ret = task->tk_status;
-	}
-	rpc_put_task(task);
-out:
-	return ret;
+	return nfs_initiate_write(data, NFS_CLIENT(inode), call_ops, how);
 }
 
 /* If a nfs_flush_* function fails, it should remove reqs from @head and
@@ -879,20 +898,21 @@ static void nfs_redirty_request(struct nfs_page *req)
  * Generate multiple small requests to write out a single
  * contiguous dirty area on one page.
  */
-static int nfs_flush_multi(struct inode *inode, struct list_head *head, unsigned int npages, size_t count, int how)
+static int nfs_flush_multi(struct nfs_pageio_descriptor *desc)
 {
-	struct nfs_page *req = nfs_list_entry(head->next);
+	struct nfs_page *req = nfs_list_entry(desc->pg_list.next);
 	struct page *page = req->wb_page;
 	struct nfs_write_data *data;
-	size_t wsize = NFS_SERVER(inode)->wsize, nbytes;
+	size_t wsize = NFS_SERVER(desc->pg_inode)->wsize, nbytes;
 	unsigned int offset;
 	int requests = 0;
 	int ret = 0;
+	struct pnfs_layout_segment *lseg;
 	LIST_HEAD(list);
 
 	nfs_list_remove_request(req);
 
-	nbytes = count;
+	nbytes = desc->pg_count;
 	do {
 		size_t len = min(nbytes, wsize);
 
@@ -905,9 +925,11 @@ static int nfs_flush_multi(struct inode *inode, struct list_head *head, unsigned
 	} while (nbytes != 0);
 	atomic_set(&req->wb_complete, requests);
 
+	BUG_ON(desc->pg_lseg);
+	lseg = pnfs_update_layout(desc->pg_inode, req->wb_context, IOMODE_RW);
 	ClearPageError(page);
 	offset = 0;
-	nbytes = count;
+	nbytes = desc->pg_count;
 	do {
 		int ret2;
 
@@ -919,13 +941,15 @@ static int nfs_flush_multi(struct inode *inode, struct list_head *head, unsigned
 		if (nbytes < wsize)
 			wsize = nbytes;
 		ret2 = nfs_write_rpcsetup(req, data, &nfs_write_partial_ops,
-				   wsize, offset, how);
+					  wsize, offset, lseg, desc->pg_ioflags);
 		if (ret == 0)
 			ret = ret2;
 		offset += wsize;
 		nbytes -= wsize;
 	} while (nbytes != 0);
 
+	put_lseg(lseg);
+	desc->pg_lseg = NULL;
 	return ret;
 
 out_bad:
@@ -946,16 +970,26 @@ out_bad:
  * This is the case if nfs_updatepage detects a conflicting request
  * that has been written but not committed.
  */
-static int nfs_flush_one(struct inode *inode, struct list_head *head, unsigned int npages, size_t count, int how)
+static int nfs_flush_one(struct nfs_pageio_descriptor *desc)
 {
 	struct nfs_page		*req;
 	struct page		**pages;
 	struct nfs_write_data	*data;
+	struct list_head *head = &desc->pg_list;
+	struct pnfs_layout_segment *lseg = desc->pg_lseg;
+	int ret;
 
-	data = nfs_writedata_alloc(npages);
-	if (!data)
-		goto out_bad;
-
+	data = nfs_writedata_alloc(nfs_page_array_len(desc->pg_base,
+						      desc->pg_count));
+	if (!data) {
+		while (!list_empty(head)) {
+			req = nfs_list_entry(head->next);
+			nfs_list_remove_request(req);
+			nfs_redirty_request(req);
+		}
+		ret = -ENOMEM;
+		goto out;
+	}
 	pages = data->pagevec;
 	while (!list_empty(head)) {
 		req = nfs_list_entry(head->next);
@@ -965,16 +999,15 @@ static int nfs_flush_one(struct inode *inode, struct list_head *head, unsigned i
 		*pages++ = req->wb_page;
 	}
 	req = nfs_list_entry(data->pages.next);
+	if ((!lseg) && list_is_singular(&data->pages))
+		lseg = pnfs_update_layout(desc->pg_inode, req->wb_context, IOMODE_RW);
 
 	/* Set up the argument struct */
-	return nfs_write_rpcsetup(req, data, &nfs_write_full_ops, count, 0, how);
- out_bad:
-	while (!list_empty(head)) {
-		req = nfs_list_entry(head->next);
-		nfs_list_remove_request(req);
-		nfs_redirty_request(req);
-	}
-	return -ENOMEM;
+	ret = nfs_write_rpcsetup(req, data, &nfs_write_full_ops, desc->pg_count, 0, lseg, desc->pg_ioflags);
+out:
+	put_lseg(lseg); /* Cleans any gotten in ->pg_test */
+	desc->pg_lseg = NULL;
+	return ret;
 }
 
 static void nfs_pageio_init_write(struct nfs_pageio_descriptor *pgio,
@@ -982,6 +1015,8 @@ static void nfs_pageio_init_write(struct nfs_pageio_descriptor *pgio,
 {
 	size_t wsize = NFS_SERVER(inode)->wsize;
 
+	pnfs_pageio_init_write(pgio, inode);
+
 	if (wsize < PAGE_CACHE_SIZE)
 		nfs_pageio_init(pgio, inode, nfs_flush_multi, wsize, ioflags);
 	else
@@ -1132,7 +1167,7 @@ static const struct rpc_call_ops nfs_write_full_ops = {
 /*
  * This function is called when the WRITE call is complete.
  */
-int nfs_writeback_done(struct rpc_task *task, struct nfs_write_data *data)
+void nfs_writeback_done(struct rpc_task *task, struct nfs_write_data *data)
 {
 	struct nfs_writeargs	*argp = &data->args;
 	struct nfs_writeres	*resp = &data->res;
@@ -1151,7 +1186,7 @@ int nfs_writeback_done(struct rpc_task *task, struct nfs_write_data *data)
 	 */
 	status = NFS_PROTO(data->inode)->write_done(task, data);
 	if (status != 0)
-		return status;
+		return;
 	nfs_add_stats(data->inode, NFSIOS_SERVERWRITTENBYTES, resp->count);
 
 #if defined(CONFIG_NFS_V3) || defined(CONFIG_NFS_V4)
@@ -1166,6 +1201,7 @@ int nfs_writeback_done(struct rpc_task *task, struct nfs_write_data *data)
 		 */
 		static unsigned long    complain;
 
+		/* Note this will print the MDS for a DS write */
 		if (time_before(complain, jiffies)) {
 			dprintk("NFS:       faulty NFS server %s:"
 				" (committed = %d) != (stable = %d)\n",
@@ -1186,6 +1222,7 @@ int nfs_writeback_done(struct rpc_task *task, struct nfs_write_data *data)
 			/* Was this an NFSv2 write or an NFSv3 stable write? */
 			if (resp->verf->committed != NFS_UNSTABLE) {
 				/* Resend from where the server left off */
+				data->mds_offset += resp->count;
 				argp->offset += resp->count;
 				argp->pgbase += resp->count;
 				argp->count -= resp->count;
@@ -1196,7 +1233,7 @@ int nfs_writeback_done(struct rpc_task *task, struct nfs_write_data *data)
 				argp->stable = NFS_FILE_SYNC;
 			}
 			nfs_restart_rpc(task, server->nfs_client);
-			return -EAGAIN;
+			return;
 		}
 		if (time_before(complain, jiffies)) {
 			printk(KERN_WARNING
@@ -1207,7 +1244,7 @@ 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;
 	}
-	return 0;
+	return;
 }
 
 

include/linux/nfs_fs.h

@@ -501,7 +501,7 @@ extern int  nfs_writepage(struct page *page, struct writeback_control *wbc);
 extern int  nfs_writepages(struct address_space *, struct writeback_control *);
 extern int  nfs_flush_incompatible(struct file *file, struct page *page);
 extern int  nfs_updatepage(struct file *, struct page *, unsigned int, unsigned int);
-extern int nfs_writeback_done(struct rpc_task *, struct nfs_write_data *);
+extern void nfs_writeback_done(struct rpc_task *, struct nfs_write_data *);
 
 /*
  * Try to write back everything synchronously (but check the

include/linux/nfs_fs_sb.h

@@ -30,6 +30,8 @@ struct nfs_client {
 #define NFS_CS_CALLBACK		1		/* - callback started */
 #define NFS_CS_IDMAP		2		/* - idmap started */
 #define NFS_CS_RENEWD		3		/* - renewd started */
+#define NFS_CS_STOP_RENEW	4		/* no more state to renew */
+#define NFS_CS_CHECK_LEASE_TIME	5		/* need to check lease time */
 	struct sockaddr_storage	cl_addr;	/* server identifier */
 	size_t			cl_addrlen;
 	char *			cl_hostname;	/* hostname of server */
@@ -75,7 +77,6 @@ struct nfs_client {
 	u32			cl_exchange_flags;
 	struct nfs4_session	*cl_session; 	/* sharred session */
 	struct list_head	cl_layouts;
-	struct pnfs_deviceid_cache *cl_devid_cache; /* pNFS deviceid cache */
 #endif /* CONFIG_NFS_V4 */
 
 #ifdef CONFIG_NFS_FSCACHE
@@ -176,6 +177,7 @@ struct nfs_server {
 #define NFS_CAP_CTIME		(1U << 12)
 #define NFS_CAP_MTIME		(1U << 13)
 #define NFS_CAP_POSIX_LOCK	(1U << 14)
+#define NFS_CAP_UIDGID_NOMAP	(1U << 15)
 
 
 /* maximum number of slots to use */

include/linux/nfs_idmap.h

@@ -65,6 +65,7 @@ struct idmap_msg {
 
 /* Forward declaration to make this header independent of others */
 struct nfs_client;
+struct nfs_server;
 
 #ifdef CONFIG_NFS_USE_NEW_IDMAPPER
 
@@ -96,10 +97,10 @@ void nfs_idmap_delete(struct nfs_client *);
 
 #endif /* CONFIG_NFS_USE_NEW_IDMAPPER */
 
-int nfs_map_name_to_uid(struct nfs_client *, const char *, size_t, __u32 *);
-int nfs_map_group_to_gid(struct nfs_client *, const char *, size_t, __u32 *);
-int nfs_map_uid_to_name(struct nfs_client *, __u32, char *, size_t);
-int nfs_map_gid_to_group(struct nfs_client *, __u32, char *, size_t);
+int nfs_map_name_to_uid(const struct nfs_server *, const char *, size_t, __u32 *);
+int nfs_map_group_to_gid(const struct nfs_server *, const char *, size_t, __u32 *);
+int nfs_map_uid_to_name(const struct nfs_server *, __u32, char *, size_t);
+int nfs_map_gid_to_group(const struct nfs_server *, __u32, char *, size_t);
 
 extern unsigned int nfs_idmap_cache_timeout;
 #endif /* __KERNEL__ */

include/linux/nfs_iostat.h

@@ -113,6 +113,8 @@ enum nfs_stat_eventcounters {
 	NFSIOS_SHORTREAD,
 	NFSIOS_SHORTWRITE,
 	NFSIOS_DELAY,
+	NFSIOS_PNFS_READ,
+	NFSIOS_PNFS_WRITE,
 	__NFSIOS_COUNTSMAX,
 };
 

include/linux/nfs_page.h

@@ -59,9 +59,11 @@ struct nfs_pageio_descriptor {
 	unsigned int		pg_base;
 
 	struct inode		*pg_inode;
-	int			(*pg_doio)(struct inode *, struct list_head *, unsigned int, size_t, int);
+	int			(*pg_doio)(struct nfs_pageio_descriptor *);
 	int 			pg_ioflags;
 	int			pg_error;
+	struct pnfs_layout_segment *pg_lseg;
+	int			(*pg_test)(struct nfs_pageio_descriptor *, struct nfs_page *, struct nfs_page *);
 };
 
 #define NFS_WBACK_BUSY(req)	(test_bit(PG_BUSY,&(req)->wb_flags))
@@ -79,7 +81,7 @@ extern	int nfs_scan_list(struct nfs_inode *nfsi, struct list_head *dst,
 			  pgoff_t idx_start, unsigned int npages, int tag);
 extern	void nfs_pageio_init(struct nfs_pageio_descriptor *desc,
 			     struct inode *inode,
-			     int (*doio)(struct inode *, struct list_head *, unsigned int, size_t, int),
+			     int (*doio)(struct nfs_pageio_descriptor *desc),
 			     size_t bsize,
 			     int how);
 extern	int nfs_pageio_add_request(struct nfs_pageio_descriptor *,

include/linux/nfs_xdr.h

@@ -1016,9 +1016,12 @@ struct nfs_read_data {
 	unsigned int		npages;	/* Max length of pagevec */
 	struct nfs_readargs args;
 	struct nfs_readres  res;
-#ifdef CONFIG_NFS_V4
 	unsigned long		timestamp;	/* For lease renewal */
-#endif
+	struct pnfs_layout_segment *lseg;
+	struct nfs_client	*ds_clp;	/* pNFS data server */
+	const struct rpc_call_ops *mds_ops;
+	int (*read_done_cb) (struct rpc_task *task, struct nfs_read_data *data);
+	__u64			mds_offset;
 	struct page		*page_array[NFS_PAGEVEC_SIZE];
 };
 
@@ -1035,13 +1038,20 @@ struct nfs_write_data {
 	unsigned int		npages;		/* Max length of pagevec */
 	struct nfs_writeargs	args;		/* argument struct */
 	struct nfs_writeres	res;		/* result struct */
+	struct pnfs_layout_segment *lseg;
+	struct nfs_client	*ds_clp;	/* pNFS data server */
+	const struct rpc_call_ops *mds_ops;
+	int (*write_done_cb) (struct rpc_task *task, struct nfs_write_data *data);
 #ifdef CONFIG_NFS_V4
 	unsigned long		timestamp;	/* For lease renewal */
 #endif
+	__u64			mds_offset;	/* Filelayout dense stripe */
 	struct page		*page_array[NFS_PAGEVEC_SIZE];
 };
 
 struct nfs_access_entry;
+struct nfs_client;
+struct rpc_timeout;
 
 /*
  * RPC procedure vector for NFSv2/NFSv3 demuxing
@@ -1106,6 +1116,8 @@ struct nfs_rpc_ops {
 				struct nfs_open_context *ctx,
 				int open_flags,
 				struct iattr *iattr);
+	int	(*init_client) (struct nfs_client *, const struct rpc_timeout *,
+				const char *, rpc_authflavor_t, int);
 };
 
 /*

include/linux/sunrpc/clnt.h

@@ -129,6 +129,7 @@ struct rpc_create_args {
 struct rpc_clnt *rpc_create(struct rpc_create_args *args);
 struct rpc_clnt	*rpc_bind_new_program(struct rpc_clnt *,
 				struct rpc_program *, u32);
+void rpc_task_reset_client(struct rpc_task *task, struct rpc_clnt *clnt);
 struct rpc_clnt *rpc_clone_client(struct rpc_clnt *);
 void		rpc_shutdown_client(struct rpc_clnt *);
 void		rpc_release_client(struct rpc_clnt *);

include/linux/sunrpc/xprt.h

@@ -12,7 +12,6 @@
 #include <linux/uio.h>
 #include <linux/socket.h>
 #include <linux/in.h>
-#include <linux/kref.h>
 #include <linux/ktime.h>
 #include <linux/sunrpc/sched.h>
 #include <linux/sunrpc/xdr.h>
@@ -146,7 +145,7 @@ enum xprt_transports {
 };
 
 struct rpc_xprt {
-	struct kref		kref;		/* Reference count */
+	atomic_t		count;		/* Reference count */
 	struct rpc_xprt_ops *	ops;		/* transport methods */
 
 	const struct rpc_timeout *timeout;	/* timeout parms */

net/sunrpc/auth_gss/auth_gss.c

@@ -417,7 +417,7 @@ static void gss_encode_v1_msg(struct gss_upcall_msg *gss_msg,
 		gss_msg->msg.len += len;
 	}
 	if (mech->gm_upcall_enctypes) {
-		len = sprintf(p, mech->gm_upcall_enctypes);
+		len = sprintf(p, "enctypes=%s ", mech->gm_upcall_enctypes);
 		p += len;
 		gss_msg->msg.len += len;
 	}

net/sunrpc/auth_gss/gss_krb5_mech.c

@@ -750,7 +750,7 @@ static struct gss_api_mech gss_kerberos_mech = {
 	.gm_ops		= &gss_kerberos_ops,
 	.gm_pf_num	= ARRAY_SIZE(gss_kerberos_pfs),
 	.gm_pfs		= gss_kerberos_pfs,
-	.gm_upcall_enctypes = "enctypes=18,17,16,23,3,1,2 ",
+	.gm_upcall_enctypes = "18,17,16,23,3,1,2",
 };
 
 static int __init init_kerberos_module(void)

net/sunrpc/clnt.c

@@ -436,7 +436,9 @@ void rpc_killall_tasks(struct rpc_clnt *clnt)
 		if (!(rovr->tk_flags & RPC_TASK_KILLED)) {
 			rovr->tk_flags |= RPC_TASK_KILLED;
 			rpc_exit(rovr, -EIO);
-			rpc_wake_up_queued_task(rovr->tk_waitqueue, rovr);
+			if (RPC_IS_QUEUED(rovr))
+				rpc_wake_up_queued_task(rovr->tk_waitqueue,
+							rovr);
 		}
 	}
 	spin_unlock(&clnt->cl_lock);
@@ -597,6 +599,14 @@ void rpc_task_set_client(struct rpc_task *task, struct rpc_clnt *clnt)
 	}
 }
 
+void rpc_task_reset_client(struct rpc_task *task, struct rpc_clnt *clnt)
+{
+	rpc_task_release_client(task);
+	rpc_task_set_client(task, clnt);
+}
+EXPORT_SYMBOL_GPL(rpc_task_reset_client);
+
+
 static void
 rpc_task_set_rpc_message(struct rpc_task *task, const struct rpc_message *msg)
 {
@@ -636,12 +646,6 @@ struct rpc_task *rpc_run_task(const struct rpc_task_setup *task_setup_data)
 	rpc_task_set_client(task, task_setup_data->rpc_client);
 	rpc_task_set_rpc_message(task, task_setup_data->rpc_message);
 
-	if (task->tk_status != 0) {
-		int ret = task->tk_status;
-		rpc_put_task(task);
-		return ERR_PTR(ret);
-	}
-
 	if (task->tk_action == NULL)
 		rpc_call_start(task);
 

net/sunrpc/sched.c

@@ -299,15 +299,8 @@ static void rpc_make_runnable(struct rpc_task *task)
 	if (rpc_test_and_set_running(task))
 		return;
 	if (RPC_IS_ASYNC(task)) {
-		int status;
-
 		INIT_WORK(&task->u.tk_work, rpc_async_schedule);
-		status = queue_work(rpciod_workqueue, &task->u.tk_work);
-		if (status < 0) {
-			printk(KERN_WARNING "RPC: failed to add task to queue: error: %d!\n", status);
-			task->tk_status = status;
-			return;
-		}
+		queue_work(rpciod_workqueue, &task->u.tk_work);
 	} else
 		wake_up_bit(&task->tk_runstate, RPC_TASK_QUEUED);
 }
@@ -637,14 +630,12 @@ static void __rpc_execute(struct rpc_task *task)
 			save_callback = task->tk_callback;
 			task->tk_callback = NULL;
 			save_callback(task);
-		}
-
-		/*
-		 * Perform the next FSM step.
-		 * tk_action may be NULL when the task has been killed
-		 * by someone else.
-		 */
-		if (!RPC_IS_QUEUED(task)) {
+		} else {
+			/*
+			 * Perform the next FSM step.
+			 * tk_action may be NULL when the task has been killed
+			 * by someone else.
+			 */
 			if (task->tk_action == NULL)
 				break;
 			task->tk_action(task);
@@ -843,12 +834,6 @@ struct rpc_task *rpc_new_task(const struct rpc_task_setup *setup_data)
 	}
 
 	rpc_init_task(task, setup_data);
-	if (task->tk_status < 0) {
-		int err = task->tk_status;
-		rpc_put_task(task);
-		return ERR_PTR(err);
-	}
-
 	task->tk_flags |= flags;
 	dprintk("RPC:       allocated task %p\n", task);
 	return task;

net/sunrpc/xprt.c

@@ -202,10 +202,9 @@ int xprt_reserve_xprt(struct rpc_task *task)
 		goto out_sleep;
 	}
 	xprt->snd_task = task;
-	if (req) {
-		req->rq_bytes_sent = 0;
-		req->rq_ntrans++;
-	}
+	req->rq_bytes_sent = 0;
+	req->rq_ntrans++;
+
 	return 1;
 
 out_sleep:
@@ -213,7 +212,7 @@ out_sleep:
 			task->tk_pid, xprt);
 	task->tk_timeout = 0;
 	task->tk_status = -EAGAIN;
-	if (req && req->rq_ntrans)
+	if (req->rq_ntrans)
 		rpc_sleep_on(&xprt->resend, task, NULL);
 	else
 		rpc_sleep_on(&xprt->sending, task, NULL);
@@ -965,7 +964,7 @@ struct rpc_xprt *xprt_alloc(struct net *net, int size, int max_req)
 	xprt = kzalloc(size, GFP_KERNEL);
 	if (xprt == NULL)
 		goto out;
-	kref_init(&xprt->kref);
+	atomic_set(&xprt->count, 1);
 
 	xprt->max_reqs = max_req;
 	xprt->slot = kcalloc(max_req, sizeof(struct rpc_rqst), GFP_KERNEL);
@@ -1145,13 +1144,11 @@ found:
 
 /**
  * xprt_destroy - destroy an RPC transport, killing off all requests.
- * @kref: kref for the transport to destroy
+ * @xprt: transport to destroy
  *
  */
-static void xprt_destroy(struct kref *kref)
+static void xprt_destroy(struct rpc_xprt *xprt)
 {
-	struct rpc_xprt *xprt = container_of(kref, struct rpc_xprt, kref);
-
 	dprintk("RPC:       destroying transport %p\n", xprt);
 	xprt->shutdown = 1;
 	del_timer_sync(&xprt->timer);
@@ -1175,7 +1172,8 @@ static void xprt_destroy(struct kref *kref)
  */
 void xprt_put(struct rpc_xprt *xprt)
 {
-	kref_put(&xprt->kref, xprt_destroy);
+	if (atomic_dec_and_test(&xprt->count))
+		xprt_destroy(xprt);
 }
 
 /**
@@ -1185,6 +1183,7 @@ void xprt_put(struct rpc_xprt *xprt)
  */
 struct rpc_xprt *xprt_get(struct rpc_xprt *xprt)
 {
-	kref_get(&xprt->kref);
-	return xprt;
+	if (atomic_inc_not_zero(&xprt->count))
+		return xprt;
+	return NULL;
 }

net/sunrpc/xprtrdma/rpc_rdma.c

@@ -87,6 +87,8 @@ rpcrdma_convert_iovs(struct xdr_buf *xdrbuf, unsigned int pos,
 	enum rpcrdma_chunktype type, struct rpcrdma_mr_seg *seg, int nsegs)
 {
 	int len, n = 0, p;
+	int page_base;
+	struct page **ppages;
 
 	if (pos == 0 && xdrbuf->head[0].iov_len) {
 		seg[n].mr_page = NULL;
@@ -95,34 +97,32 @@ rpcrdma_convert_iovs(struct xdr_buf *xdrbuf, unsigned int pos,
 		++n;
 	}
 
-	if (xdrbuf->page_len && (xdrbuf->pages[0] != NULL)) {
-		if (n == nsegs)
-			return 0;
-		seg[n].mr_page = xdrbuf->pages[0];
-		seg[n].mr_offset = (void *)(unsigned long) xdrbuf->page_base;
-		seg[n].mr_len = min_t(u32,
-			PAGE_SIZE - xdrbuf->page_base, xdrbuf->page_len);
-		len = xdrbuf->page_len - seg[n].mr_len;
+	len = xdrbuf->page_len;
+	ppages = xdrbuf->pages + (xdrbuf->page_base >> PAGE_SHIFT);
+	page_base = xdrbuf->page_base & ~PAGE_MASK;
+	p = 0;
+	while (len && n < nsegs) {
+		seg[n].mr_page = ppages[p];
+		seg[n].mr_offset = (void *)(unsigned long) page_base;
+		seg[n].mr_len = min_t(u32, PAGE_SIZE - page_base, len);
+		BUG_ON(seg[n].mr_len > PAGE_SIZE);
+		len -= seg[n].mr_len;
 		++n;
-		p = 1;
-		while (len > 0) {
-			if (n == nsegs)
-				return 0;
-			seg[n].mr_page = xdrbuf->pages[p];
-			seg[n].mr_offset = NULL;
-			seg[n].mr_len = min_t(u32, PAGE_SIZE, len);
-			len -= seg[n].mr_len;
-			++n;
-			++p;
-		}
+		++p;
+		page_base = 0;	/* page offset only applies to first page */
 	}
 
+	/* Message overflows the seg array */
+	if (len && n == nsegs)
+		return 0;
+
 	if (xdrbuf->tail[0].iov_len) {
 		/* the rpcrdma protocol allows us to omit any trailing
 		 * xdr pad bytes, saving the server an RDMA operation. */
 		if (xdrbuf->tail[0].iov_len < 4 && xprt_rdma_pad_optimize)
 			return n;
 		if (n == nsegs)
+			/* Tail remains, but we're out of segments */
 			return 0;
 		seg[n].mr_page = NULL;
 		seg[n].mr_offset = xdrbuf->tail[0].iov_base;
@@ -296,6 +296,8 @@ rpcrdma_inline_pullup(struct rpc_rqst *rqst, int pad)
 	int copy_len;
 	unsigned char *srcp, *destp;
 	struct rpcrdma_xprt *r_xprt = rpcx_to_rdmax(rqst->rq_xprt);
+	int page_base;
+	struct page **ppages;
 
 	destp = rqst->rq_svec[0].iov_base;
 	curlen = rqst->rq_svec[0].iov_len;
@@ -324,28 +326,25 @@ rpcrdma_inline_pullup(struct rpc_rqst *rqst, int pad)
 			__func__, destp + copy_len, curlen);
 		rqst->rq_svec[0].iov_len += curlen;
 	}
-
 	r_xprt->rx_stats.pullup_copy_count += copy_len;
-	npages = PAGE_ALIGN(rqst->rq_snd_buf.page_base+copy_len) >> PAGE_SHIFT;
+
+	page_base = rqst->rq_snd_buf.page_base;
+	ppages = rqst->rq_snd_buf.pages + (page_base >> PAGE_SHIFT);
+	page_base &= ~PAGE_MASK;
+	npages = PAGE_ALIGN(page_base+copy_len) >> PAGE_SHIFT;
 	for (i = 0; copy_len && i < npages; i++) {
-		if (i == 0)
-			curlen = PAGE_SIZE - rqst->rq_snd_buf.page_base;
-		else
-			curlen = PAGE_SIZE;
+		curlen = PAGE_SIZE - page_base;
 		if (curlen > copy_len)
 			curlen = copy_len;
 		dprintk("RPC:       %s: page %d destp 0x%p len %d curlen %d\n",
 			__func__, i, destp, copy_len, curlen);
-		srcp = kmap_atomic(rqst->rq_snd_buf.pages[i],
-					KM_SKB_SUNRPC_DATA);
-		if (i == 0)
-			memcpy(destp, srcp+rqst->rq_snd_buf.page_base, curlen);
-		else
-			memcpy(destp, srcp, curlen);
+		srcp = kmap_atomic(ppages[i], KM_SKB_SUNRPC_DATA);
+		memcpy(destp, srcp+page_base, curlen);
 		kunmap_atomic(srcp, KM_SKB_SUNRPC_DATA);
 		rqst->rq_svec[0].iov_len += curlen;
 		destp += curlen;
 		copy_len -= curlen;
+		page_base = 0;
 	}
 	/* header now contains entire send message */
 	return pad;
@@ -606,6 +605,8 @@ rpcrdma_inline_fixup(struct rpc_rqst *rqst, char *srcp, int copy_len, int pad)
 {
 	int i, npages, curlen, olen;
 	char *destp;
+	struct page **ppages;
+	int page_base;
 
 	curlen = rqst->rq_rcv_buf.head[0].iov_len;
 	if (curlen > copy_len) {	/* write chunk header fixup */
@@ -624,32 +625,29 @@ rpcrdma_inline_fixup(struct rpc_rqst *rqst, char *srcp, int copy_len, int pad)
 	olen = copy_len;
 	i = 0;
 	rpcx_to_rdmax(rqst->rq_xprt)->rx_stats.fixup_copy_count += olen;
+	page_base = rqst->rq_rcv_buf.page_base;
+	ppages = rqst->rq_rcv_buf.pages + (page_base >> PAGE_SHIFT);
+	page_base &= ~PAGE_MASK;
+
 	if (copy_len && rqst->rq_rcv_buf.page_len) {
-		npages = PAGE_ALIGN(rqst->rq_rcv_buf.page_base +
+		npages = PAGE_ALIGN(page_base +
 			rqst->rq_rcv_buf.page_len) >> PAGE_SHIFT;
 		for (; i < npages; i++) {
-			if (i == 0)
-				curlen = PAGE_SIZE - rqst->rq_rcv_buf.page_base;
-			else
-				curlen = PAGE_SIZE;
+			curlen = PAGE_SIZE - page_base;
 			if (curlen > copy_len)
 				curlen = copy_len;
 			dprintk("RPC:       %s: page %d"
 				" srcp 0x%p len %d curlen %d\n",
 				__func__, i, srcp, copy_len, curlen);
-			destp = kmap_atomic(rqst->rq_rcv_buf.pages[i],
-						KM_SKB_SUNRPC_DATA);
-			if (i == 0)
-				memcpy(destp + rqst->rq_rcv_buf.page_base,
-						srcp, curlen);
-			else
-				memcpy(destp, srcp, curlen);
-			flush_dcache_page(rqst->rq_rcv_buf.pages[i]);
+			destp = kmap_atomic(ppages[i], KM_SKB_SUNRPC_DATA);
+			memcpy(destp + page_base, srcp, curlen);
+			flush_dcache_page(ppages[i]);
 			kunmap_atomic(destp, KM_SKB_SUNRPC_DATA);
 			srcp += curlen;
 			copy_len -= curlen;
 			if (copy_len == 0)
 				break;
+			page_base = 0;
 		}
 		rqst->rq_rcv_buf.page_len = olen - copy_len;
 	} else

net/sunrpc/xprtrdma/verbs.c

@@ -144,6 +144,7 @@ rpcrdma_cq_async_error_upcall(struct ib_event *event, void *context)
 static inline
 void rpcrdma_event_process(struct ib_wc *wc)
 {
+	struct rpcrdma_mw *frmr;
 	struct rpcrdma_rep *rep =
 			(struct rpcrdma_rep *)(unsigned long) wc->wr_id;
 
@@ -154,15 +155,23 @@ void rpcrdma_event_process(struct ib_wc *wc)
 		return;
 
 	if (IB_WC_SUCCESS != wc->status) {
-		dprintk("RPC:       %s: %s WC status %X, connection lost\n",
-			__func__, (wc->opcode & IB_WC_RECV) ? "recv" : "send",
-			 wc->status);
+		dprintk("RPC:       %s: WC opcode %d status %X, connection lost\n",
+			__func__, wc->opcode, wc->status);
 		rep->rr_len = ~0U;
-		rpcrdma_schedule_tasklet(rep);
+		if (wc->opcode != IB_WC_FAST_REG_MR && wc->opcode != IB_WC_LOCAL_INV)
+			rpcrdma_schedule_tasklet(rep);
 		return;
 	}
 
 	switch (wc->opcode) {
+	case IB_WC_FAST_REG_MR:
+		frmr = (struct rpcrdma_mw *)(unsigned long)wc->wr_id;
+		frmr->r.frmr.state = FRMR_IS_VALID;
+		break;
+	case IB_WC_LOCAL_INV:
+		frmr = (struct rpcrdma_mw *)(unsigned long)wc->wr_id;
+		frmr->r.frmr.state = FRMR_IS_INVALID;
+		break;
 	case IB_WC_RECV:
 		rep->rr_len = wc->byte_len;
 		ib_dma_sync_single_for_cpu(
@@ -1450,6 +1459,12 @@ rpcrdma_map_one(struct rpcrdma_ia *ia, struct rpcrdma_mr_seg *seg, int writing)
 		seg->mr_dma = ib_dma_map_single(ia->ri_id->device,
 				seg->mr_offset,
 				seg->mr_dmalen, seg->mr_dir);
+	if (ib_dma_mapping_error(ia->ri_id->device, seg->mr_dma)) {
+		dprintk("RPC:       %s: mr_dma %llx mr_offset %p mr_dma_len %zu\n",
+			__func__,
+			(unsigned long long)seg->mr_dma,
+			seg->mr_offset, seg->mr_dmalen);
+	}
 }
 
 static void
@@ -1469,7 +1484,8 @@ rpcrdma_register_frmr_external(struct rpcrdma_mr_seg *seg,
 			struct rpcrdma_xprt *r_xprt)
 {
 	struct rpcrdma_mr_seg *seg1 = seg;
-	struct ib_send_wr frmr_wr, *bad_wr;
+	struct ib_send_wr invalidate_wr, frmr_wr, *bad_wr, *post_wr;
+
 	u8 key;
 	int len, pageoff;
 	int i, rc;
@@ -1484,6 +1500,7 @@ rpcrdma_register_frmr_external(struct rpcrdma_mr_seg *seg,
 		rpcrdma_map_one(ia, seg, writing);
 		seg1->mr_chunk.rl_mw->r.frmr.fr_pgl->page_list[i] = seg->mr_dma;
 		len += seg->mr_len;
+		BUG_ON(seg->mr_len > PAGE_SIZE);
 		++seg;
 		++i;
 		/* Check for holes */
@@ -1494,26 +1511,45 @@ rpcrdma_register_frmr_external(struct rpcrdma_mr_seg *seg,
 	dprintk("RPC:       %s: Using frmr %p to map %d segments\n",
 		__func__, seg1->mr_chunk.rl_mw, i);
 
+	if (unlikely(seg1->mr_chunk.rl_mw->r.frmr.state == FRMR_IS_VALID)) {
+		dprintk("RPC:       %s: frmr %x left valid, posting invalidate.\n",
+			__func__,
+			seg1->mr_chunk.rl_mw->r.frmr.fr_mr->rkey);
+		/* Invalidate before using. */
+		memset(&invalidate_wr, 0, sizeof invalidate_wr);
+		invalidate_wr.wr_id = (unsigned long)(void *)seg1->mr_chunk.rl_mw;
+		invalidate_wr.next = &frmr_wr;
+		invalidate_wr.opcode = IB_WR_LOCAL_INV;
+		invalidate_wr.send_flags = IB_SEND_SIGNALED;
+		invalidate_wr.ex.invalidate_rkey =
+			seg1->mr_chunk.rl_mw->r.frmr.fr_mr->rkey;
+		DECR_CQCOUNT(&r_xprt->rx_ep);
+		post_wr = &invalidate_wr;
+	} else
+		post_wr = &frmr_wr;
+
 	/* Bump the key */
 	key = (u8)(seg1->mr_chunk.rl_mw->r.frmr.fr_mr->rkey & 0x000000FF);
 	ib_update_fast_reg_key(seg1->mr_chunk.rl_mw->r.frmr.fr_mr, ++key);
 
 	/* Prepare FRMR WR */
 	memset(&frmr_wr, 0, sizeof frmr_wr);
+	frmr_wr.wr_id = (unsigned long)(void *)seg1->mr_chunk.rl_mw;
 	frmr_wr.opcode = IB_WR_FAST_REG_MR;
-	frmr_wr.send_flags = 0;			/* unsignaled */
+	frmr_wr.send_flags = IB_SEND_SIGNALED;
 	frmr_wr.wr.fast_reg.iova_start = seg1->mr_dma;
 	frmr_wr.wr.fast_reg.page_list = seg1->mr_chunk.rl_mw->r.frmr.fr_pgl;
 	frmr_wr.wr.fast_reg.page_list_len = i;
 	frmr_wr.wr.fast_reg.page_shift = PAGE_SHIFT;
 	frmr_wr.wr.fast_reg.length = i << PAGE_SHIFT;
+	BUG_ON(frmr_wr.wr.fast_reg.length < len);
 	frmr_wr.wr.fast_reg.access_flags = (writing ?
 				IB_ACCESS_REMOTE_WRITE | IB_ACCESS_LOCAL_WRITE :
 				IB_ACCESS_REMOTE_READ);
 	frmr_wr.wr.fast_reg.rkey = seg1->mr_chunk.rl_mw->r.frmr.fr_mr->rkey;
 	DECR_CQCOUNT(&r_xprt->rx_ep);
 
-	rc = ib_post_send(ia->ri_id->qp, &frmr_wr, &bad_wr);
+	rc = ib_post_send(ia->ri_id->qp, post_wr, &bad_wr);
 
 	if (rc) {
 		dprintk("RPC:       %s: failed ib_post_send for register,"
@@ -1542,8 +1578,9 @@ rpcrdma_deregister_frmr_external(struct rpcrdma_mr_seg *seg,
 		rpcrdma_unmap_one(ia, seg++);
 
 	memset(&invalidate_wr, 0, sizeof invalidate_wr);
+	invalidate_wr.wr_id = (unsigned long)(void *)seg1->mr_chunk.rl_mw;
 	invalidate_wr.opcode = IB_WR_LOCAL_INV;
-	invalidate_wr.send_flags = 0;			/* unsignaled */
+	invalidate_wr.send_flags = IB_SEND_SIGNALED;
 	invalidate_wr.ex.invalidate_rkey = seg1->mr_chunk.rl_mw->r.frmr.fr_mr->rkey;
 	DECR_CQCOUNT(&r_xprt->rx_ep);
 

net/sunrpc/xprtrdma/xprt_rdma.h

@@ -164,6 +164,7 @@ struct rpcrdma_mr_seg {		/* chunk descriptors */
 				struct {
 					struct ib_fast_reg_page_list *fr_pgl;
 					struct ib_mr *fr_mr;
+					enum { FRMR_IS_INVALID, FRMR_IS_VALID  } state;
 				} frmr;
 			} r;
 			struct list_head mw_list;