Merge branch 'devel' into next

fs/lockd/clntlock.c

@@ -14,6 +14,7 @@
 #include <linux/sunrpc/svc.h>
 #include <linux/lockd/lockd.h>
 #include <linux/smp_lock.h>
+#include <linux/kthread.h>
 
 #define NLMDBG_FACILITY		NLMDBG_CLIENT
 
@@ -60,7 +61,7 @@ struct nlm_host *nlmclnt_init(const struct nlmclnt_initdata *nlm_init)
 
 	host = nlmclnt_lookup_host(nlm_init->address, nlm_init->addrlen,
 				   nlm_init->protocol, nlm_version,
-				   nlm_init->hostname);
+				   nlm_init->hostname, nlm_init->noresvport);
 	if (host == NULL) {
 		lockd_down();
 		return ERR_PTR(-ENOLCK);
@@ -191,11 +192,15 @@ __be32 nlmclnt_grant(const struct sockaddr *addr, const struct nlm_lock *lock)
 void
 nlmclnt_recovery(struct nlm_host *host)
 {
+	struct task_struct *task;
+
 	if (!host->h_reclaiming++) {
 		nlm_get_host(host);
-		__module_get(THIS_MODULE);
-		if (kernel_thread(reclaimer, host, CLONE_FS | CLONE_FILES) < 0)
-			module_put(THIS_MODULE);
+		task = kthread_run(reclaimer, host, "%s-reclaim", host->h_name);
+		if (IS_ERR(task))
+			printk(KERN_ERR "lockd: unable to spawn reclaimer "
+				"thread. Locks for %s won't be reclaimed! "
+				"(%ld)\n", host->h_name, PTR_ERR(task));
 	}
 }
 
@@ -207,7 +212,6 @@ reclaimer(void *ptr)
 	struct file_lock *fl, *next;
 	u32 nsmstate;
 
-	daemonize("%s-reclaim", host->h_name);
 	allow_signal(SIGKILL);
 
 	down_write(&host->h_rwsem);
@@ -233,7 +237,12 @@ reclaimer(void *ptr)
 	list_for_each_entry_safe(fl, next, &host->h_reclaim, fl_u.nfs_fl.list) {
 		list_del_init(&fl->fl_u.nfs_fl.list);
 
-		/* Why are we leaking memory here? --okir */
+		/*
+		 * sending this thread a SIGKILL will result in any unreclaimed
+		 * locks being removed from the h_granted list. This means that
+		 * the kernel will not attempt to reclaim them again if a new
+		 * reclaimer thread is spawned for this host.
+		 */
 		if (signalled())
 			continue;
 		if (nlmclnt_reclaim(host, fl) != 0)
@@ -261,5 +270,5 @@ reclaimer(void *ptr)
 	nlm_release_host(host);
 	lockd_down();
 	unlock_kernel();
-	module_put_and_exit(0);
+	return 0;
 }

fs/lockd/host.c

@@ -48,6 +48,7 @@ struct nlm_lookup_host_info {
 	const size_t		hostname_len;	/* it's length */
 	const struct sockaddr	*src_sap;	/* our address (optional) */
 	const size_t		src_len;	/* it's length */
+	const int		noresvport;	/* use non-priv port */
 };
 
 /*
@@ -222,6 +223,7 @@ static struct nlm_host *nlm_lookup_host(struct nlm_lookup_host_info *ni)
 	host->h_nsmstate   = 0;			/* real NSM state */
 	host->h_nsmhandle  = nsm;
 	host->h_server	   = ni->server;
+	host->h_noresvport = ni->noresvport;
 	hlist_add_head(&host->h_hash, chain);
 	INIT_LIST_HEAD(&host->h_lockowners);
 	spin_lock_init(&host->h_lock);
@@ -272,6 +274,7 @@ nlm_destroy_host(struct nlm_host *host)
  * @protocol: transport protocol to use
  * @version: NLM protocol version
  * @hostname: '\0'-terminated hostname of server
+ * @noresvport: 1 if non-privileged port should be used
  *
  * Returns an nlm_host structure that matches the passed-in
  * [server address, transport protocol, NLM version, server hostname].
@@ -281,7 +284,9 @@ nlm_destroy_host(struct nlm_host *host)
 struct nlm_host *nlmclnt_lookup_host(const struct sockaddr *sap,
 				     const size_t salen,
 				     const unsigned short protocol,
-				     const u32 version, const char *hostname)
+				     const u32 version,
+				     const char *hostname,
+				     int noresvport)
 {
 	const struct sockaddr source = {
 		.sa_family	= AF_UNSPEC,
@@ -296,6 +301,7 @@ struct nlm_host *nlmclnt_lookup_host(const struct sockaddr *sap,
 		.hostname_len	= strlen(hostname),
 		.src_sap	= &source,
 		.src_len	= sizeof(source),
+		.noresvport	= noresvport,
 	};
 
 	dprintk("lockd: %s(host='%s', vers=%u, proto=%s)\n", __func__,
@@ -417,6 +423,8 @@ nlm_bind_host(struct nlm_host *host)
 		 */
 		if (!host->h_server)
 			args.flags |= RPC_CLNT_CREATE_HARDRTRY;
+		if (host->h_noresvport)
+			args.flags |= RPC_CLNT_CREATE_NONPRIVPORT;
 
 		clnt = rpc_create(&args);
 		if (!IS_ERR(clnt))

fs/lockd/svc.c

@@ -45,7 +45,7 @@
 static struct svc_program	nlmsvc_program;
 
 struct nlmsvc_binding *		nlmsvc_ops;
-EXPORT_SYMBOL(nlmsvc_ops);
+EXPORT_SYMBOL_GPL(nlmsvc_ops);
 
 static DEFINE_MUTEX(nlmsvc_mutex);
 static unsigned int		nlmsvc_users;
@@ -300,7 +300,7 @@ int lockd_up(void)
 	mutex_unlock(&nlmsvc_mutex);
 	return error;
 }
-EXPORT_SYMBOL(lockd_up);
+EXPORT_SYMBOL_GPL(lockd_up);
 
 /*
  * Decrement the user count and bring down lockd if we're the last.
@@ -329,7 +329,7 @@ lockd_down(void)
 out:
 	mutex_unlock(&nlmsvc_mutex);
 }
-EXPORT_SYMBOL(lockd_down);
+EXPORT_SYMBOL_GPL(lockd_down);
 
 #ifdef CONFIG_SYSCTL
 

fs/nfs/callback.c

@@ -16,6 +16,7 @@
 #include <linux/mutex.h>
 #include <linux/freezer.h>
 #include <linux/kthread.h>
+#include <linux/sunrpc/svcauth_gss.h>
 
 #include <net/inet_sock.h>
 
@@ -182,10 +183,34 @@ void nfs_callback_down(void)
 	mutex_unlock(&nfs_callback_mutex);
 }
 
+static int check_gss_callback_principal(struct nfs_client *clp,
+					struct svc_rqst *rqstp)
+{
+	struct rpc_clnt *r = clp->cl_rpcclient;
+	char *p = svc_gss_principal(rqstp);
+
+	/*
+	 * It might just be a normal user principal, in which case
+	 * userspace won't bother to tell us the name at all.
+	 */
+	if (p == NULL)
+		return SVC_DENIED;
+
+	/* Expect a GSS_C_NT_HOSTBASED_NAME like "nfs@serverhostname" */
+
+	if (memcmp(p, "nfs@", 4) != 0)
+		return SVC_DENIED;
+	p += 4;
+	if (strcmp(p, r->cl_server) != 0)
+		return SVC_DENIED;
+	return SVC_OK;
+}
+
 static int nfs_callback_authenticate(struct svc_rqst *rqstp)
 {
 	struct nfs_client *clp;
 	RPC_IFDEBUG(char buf[RPC_MAX_ADDRBUFLEN]);
+	int ret = SVC_OK;
 
 	/* Don't talk to strangers */
 	clp = nfs_find_client(svc_addr(rqstp), 4);
@@ -194,21 +219,22 @@ static int nfs_callback_authenticate(struct svc_rqst *rqstp)
 
 	dprintk("%s: %s NFSv4 callback!\n", __func__,
 			svc_print_addr(rqstp, buf, sizeof(buf)));
-	nfs_put_client(clp);
 
 	switch (rqstp->rq_authop->flavour) {
 		case RPC_AUTH_NULL:
 			if (rqstp->rq_proc != CB_NULL)
-				return SVC_DENIED;
+				ret = SVC_DENIED;
 			break;
 		case RPC_AUTH_UNIX:
 			break;
 		case RPC_AUTH_GSS:
-			/* FIXME: RPCSEC_GSS handling? */
+			ret = check_gss_callback_principal(clp, rqstp);
+			break;
 		default:
-			return SVC_DENIED;
+			ret = SVC_DENIED;
 	}
-	return SVC_OK;
+	nfs_put_client(clp);
+	return ret;
 }
 
 /*

fs/nfs/client.c

@@ -143,7 +143,6 @@ static struct nfs_client *nfs_alloc_client(const struct nfs_client_initdata *cl_
 	clp->cl_proto = cl_init->proto;
 
 #ifdef CONFIG_NFS_V4
-	init_rwsem(&clp->cl_sem);
 	INIT_LIST_HEAD(&clp->cl_delegations);
 	spin_lock_init(&clp->cl_lock);
 	INIT_DELAYED_WORK(&clp->cl_renewd, nfs4_renew_state);
@@ -224,31 +223,54 @@ void nfs_put_client(struct nfs_client *clp)
 	}
 }
 
-static int nfs_sockaddr_match_ipaddr4(const struct sockaddr_in *sa1,
-				 const struct sockaddr_in *sa2)
+#if defined(CONFIG_IPV6) || defined(CONFIG_IPV6_MODULE)
+static const struct in6_addr *nfs_map_ipv4_addr(const struct sockaddr *sa, struct in6_addr *addr_mapped)
 {
-	return sa1->sin_addr.s_addr == sa2->sin_addr.s_addr;
+	switch (sa->sa_family) {
+		default:
+			return NULL;
+		case AF_INET6:
+			return &((const struct sockaddr_in6 *)sa)->sin6_addr;
+			break;
+		case AF_INET:
+			ipv6_addr_set_v4mapped(((const struct sockaddr_in *)sa)->sin_addr.s_addr,
+					addr_mapped);
+			return addr_mapped;
+	}
 }
 
-static int nfs_sockaddr_match_ipaddr6(const struct sockaddr_in6 *sa1,
-				 const struct sockaddr_in6 *sa2)
+static int nfs_sockaddr_match_ipaddr(const struct sockaddr *sa1,
+		const struct sockaddr *sa2)
+{
+	const struct in6_addr *addr1;
+	const struct in6_addr *addr2;
+	struct in6_addr addr1_mapped;
+	struct in6_addr addr2_mapped;
+
+	addr1 = nfs_map_ipv4_addr(sa1, &addr1_mapped);
+	if (likely(addr1 != NULL)) {
+		addr2 = nfs_map_ipv4_addr(sa2, &addr2_mapped);
+		if (likely(addr2 != NULL))
+			return ipv6_addr_equal(addr1, addr2);
+	}
+	return 0;
+}
+#else
+static int nfs_sockaddr_match_ipaddr4(const struct sockaddr_in *sa1,
+				 const struct sockaddr_in *sa2)
 {
-	return ipv6_addr_equal(&sa1->sin6_addr, &sa2->sin6_addr);
+	return sa1->sin_addr.s_addr == sa2->sin_addr.s_addr;
 }
 
 static int nfs_sockaddr_match_ipaddr(const struct sockaddr *sa1,
 				 const struct sockaddr *sa2)
 {
-	switch (sa1->sa_family) {
-	case AF_INET:
-		return nfs_sockaddr_match_ipaddr4((const struct sockaddr_in *)sa1,
-				(const struct sockaddr_in *)sa2);
-	case AF_INET6:
-		return nfs_sockaddr_match_ipaddr6((const struct sockaddr_in6 *)sa1,
-				(const struct sockaddr_in6 *)sa2);
-	}
-	BUG();
+	if (unlikely(sa1->sa_family != AF_INET || sa2->sa_family != AF_INET))
+		return 0;
+	return nfs_sockaddr_match_ipaddr4((const struct sockaddr_in *)sa1,
+			(const struct sockaddr_in *)sa2);
 }
+#endif
 
 /*
  * Find a client by IP address and protocol version
@@ -270,8 +292,6 @@ struct nfs_client *nfs_find_client(const struct sockaddr *addr, u32 nfsversion)
 		if (clp->rpc_ops->version != nfsversion)
 			continue;
 
-		if (addr->sa_family != clap->sa_family)
-			continue;
 		/* Match only the IP address, not the port number */
 		if (!nfs_sockaddr_match_ipaddr(addr, clap))
 			continue;
@@ -305,8 +325,6 @@ struct nfs_client *nfs_find_client_next(struct nfs_client *clp)
 		if (clp->rpc_ops->version != nfsvers)
 			continue;
 
-		if (sap->sa_family != clap->sa_family)
-			continue;
 		/* Match only the IP address, not the port number */
 		if (!nfs_sockaddr_match_ipaddr(sap, clap))
 			continue;
@@ -470,7 +488,7 @@ static void nfs_init_timeout_values(struct rpc_timeout *to, int proto,
 static int nfs_create_rpc_client(struct nfs_client *clp,
 				 const struct rpc_timeout *timeparms,
 				 rpc_authflavor_t flavor,
-				 int flags)
+				 int discrtry, int noresvport)
 {
 	struct rpc_clnt		*clnt = NULL;
 	struct rpc_create_args args = {
@@ -482,9 +500,13 @@ static int nfs_create_rpc_client(struct nfs_client *clp,
 		.program	= &nfs_program,
 		.version	= clp->rpc_ops->version,
 		.authflavor	= flavor,
-		.flags		= flags,
 	};
 
+	if (discrtry)
+		args.flags |= RPC_CLNT_CREATE_DISCRTRY;
+	if (noresvport)
+		args.flags |= RPC_CLNT_CREATE_NONPRIVPORT;
+
 	if (!IS_ERR(clp->cl_rpcclient))
 		return 0;
 
@@ -522,6 +544,8 @@ static int nfs_start_lockd(struct nfs_server *server)
 		.protocol	= server->flags & NFS_MOUNT_TCP ?
 						IPPROTO_TCP : IPPROTO_UDP,
 		.nfs_version	= clp->rpc_ops->version,
+		.noresvport	= server->flags & NFS_MOUNT_NORESVPORT ?
+					1 : 0,
 	};
 
 	if (nlm_init.nfs_version > 3)
@@ -623,7 +647,8 @@ static int nfs_init_client(struct nfs_client *clp,
 	 * Create a client RPC handle for doing FSSTAT with UNIX auth only
 	 * - RFC 2623, sec 2.3.2
 	 */
-	error = nfs_create_rpc_client(clp, timeparms, RPC_AUTH_UNIX, 0);
+	error = nfs_create_rpc_client(clp, timeparms, RPC_AUTH_UNIX,
+				      0, data->flags & NFS_MOUNT_NORESVPORT);
 	if (error < 0)
 		goto error;
 	nfs_mark_client_ready(clp, NFS_CS_READY);
@@ -965,7 +990,8 @@ struct nfs_server *nfs_create_server(const struct nfs_parsed_mount_data *data,
 static int nfs4_init_client(struct nfs_client *clp,
 		const struct rpc_timeout *timeparms,
 		const char *ip_addr,
-		rpc_authflavor_t authflavour)
+		rpc_authflavor_t authflavour,
+		int flags)
 {
 	int error;
 
@@ -979,7 +1005,7 @@ static int nfs4_init_client(struct nfs_client *clp,
 	clp->rpc_ops = &nfs_v4_clientops;
 
 	error = nfs_create_rpc_client(clp, timeparms, authflavour,
-					RPC_CLNT_CREATE_DISCRTRY);
+				      1, flags & NFS_MOUNT_NORESVPORT);
 	if (error < 0)
 		goto error;
 	memcpy(clp->cl_ipaddr, ip_addr, sizeof(clp->cl_ipaddr));
@@ -1030,7 +1056,8 @@ static int nfs4_set_client(struct nfs_server *server,
 		error = PTR_ERR(clp);
 		goto error;
 	}
-	error = nfs4_init_client(clp, timeparms, ip_addr, authflavour);
+	error = nfs4_init_client(clp, timeparms, ip_addr, authflavour,
+					server->flags);
 	if (error < 0)
 		goto error_put;
 
@@ -1059,6 +1086,10 @@ static int nfs4_init_server(struct nfs_server *server,
 	nfs_init_timeout_values(&timeparms, data->nfs_server.protocol,
 			data->timeo, data->retrans);
 
+	/* Initialise the client representation from the mount data */
+	server->flags = data->flags;
+	server->caps |= NFS_CAP_ATOMIC_OPEN;
+
 	/* Get a client record */
 	error = nfs4_set_client(server,
 			data->nfs_server.hostname,
@@ -1071,10 +1102,6 @@ static int nfs4_init_server(struct nfs_server *server,
 	if (error < 0)
 		goto error;
 
-	/* Initialise the client representation from the mount data */
-	server->flags = data->flags;
-	server->caps |= NFS_CAP_ATOMIC_OPEN;
-
 	if (data->rsize)
 		server->rsize = nfs_block_size(data->rsize, NULL);
 	if (data->wsize)
@@ -1177,6 +1204,10 @@ struct nfs_server *nfs4_create_referral_server(struct nfs_clone_mount *data,
 	parent_server = NFS_SB(data->sb);
 	parent_client = parent_server->nfs_client;
 
+	/* Initialise the client representation from the parent server */
+	nfs_server_copy_userdata(server, parent_server);
+	server->caps |= NFS_CAP_ATOMIC_OPEN;
+
 	/* Get a client representation.
 	 * Note: NFSv4 always uses TCP, */
 	error = nfs4_set_client(server, data->hostname,
@@ -1189,10 +1220,6 @@ struct nfs_server *nfs4_create_referral_server(struct nfs_clone_mount *data,
 	if (error < 0)
 		goto error;
 
-	/* Initialise the client representation from the parent server */
-	nfs_server_copy_userdata(server, parent_server);
-	server->caps |= NFS_CAP_ATOMIC_OPEN;
-
 	error = nfs_init_server_rpcclient(server, parent_server->client->cl_timeout, data->authflavor);
 	if (error < 0)
 		goto error;

fs/nfs/delegation.c

@@ -43,6 +43,27 @@ static void nfs_free_delegation(struct nfs_delegation *delegation)
 		put_rpccred(cred);
 }
 
+void nfs_mark_delegation_referenced(struct nfs_delegation *delegation)
+{
+	set_bit(NFS_DELEGATION_REFERENCED, &delegation->flags);
+}
+
+int nfs_have_delegation(struct inode *inode, fmode_t flags)
+{
+	struct nfs_delegation *delegation;
+	int ret = 0;
+
+	flags &= FMODE_READ|FMODE_WRITE;
+	rcu_read_lock();
+	delegation = rcu_dereference(NFS_I(inode)->delegation);
+	if (delegation != NULL && (delegation->type & flags) == flags) {
+		nfs_mark_delegation_referenced(delegation);
+		ret = 1;
+	}
+	rcu_read_unlock();
+	return ret;
+}
+
 static int nfs_delegation_claim_locks(struct nfs_open_context *ctx, struct nfs4_state *state)
 {
 	struct inode *inode = state->inode;
@@ -119,7 +140,7 @@ void nfs_inode_reclaim_delegation(struct inode *inode, struct rpc_cred *cred, st
 	delegation->maxsize = res->maxsize;
 	oldcred = delegation->cred;
 	delegation->cred = get_rpccred(cred);
-	delegation->flags &= ~NFS_DELEGATION_NEED_RECLAIM;
+	clear_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags);
 	NFS_I(inode)->delegation_state = delegation->type;
 	smp_wmb();
 	put_rpccred(oldcred);
@@ -134,19 +155,35 @@ static int nfs_do_return_delegation(struct inode *inode, struct nfs_delegation *
 	return res;
 }
 
+static struct inode *nfs_delegation_grab_inode(struct nfs_delegation *delegation)
+{
+	struct inode *inode = NULL;
+
+	spin_lock(&delegation->lock);
+	if (delegation->inode != NULL)
+		inode = igrab(delegation->inode);
+	spin_unlock(&delegation->lock);
+	return inode;
+}
+
 static struct nfs_delegation *nfs_detach_delegation_locked(struct nfs_inode *nfsi, const nfs4_stateid *stateid)
 {
 	struct nfs_delegation *delegation = rcu_dereference(nfsi->delegation);
 
 	if (delegation == NULL)
 		goto nomatch;
+	spin_lock(&delegation->lock);
 	if (stateid != NULL && memcmp(delegation->stateid.data, stateid->data,
 				sizeof(delegation->stateid.data)) != 0)
-		goto nomatch;
+		goto nomatch_unlock;
 	list_del_rcu(&delegation->super_list);
+	delegation->inode = NULL;
 	nfsi->delegation_state = 0;
 	rcu_assign_pointer(nfsi->delegation, NULL);
+	spin_unlock(&delegation->lock);
 	return delegation;
+nomatch_unlock:
+	spin_unlock(&delegation->lock);
 nomatch:
 	return NULL;
 }
@@ -172,6 +209,8 @@ int nfs_inode_set_delegation(struct inode *inode, struct rpc_cred *cred, struct
 	delegation->change_attr = nfsi->change_attr;
 	delegation->cred = get_rpccred(cred);
 	delegation->inode = inode;
+	delegation->flags = 1<<NFS_DELEGATION_REFERENCED;
+	spin_lock_init(&delegation->lock);
 
 	spin_lock(&clp->cl_lock);
 	if (rcu_dereference(nfsi->delegation) != NULL) {
@@ -226,21 +265,46 @@ static void nfs_msync_inode(struct inode *inode)
  */
 static int __nfs_inode_return_delegation(struct inode *inode, struct nfs_delegation *delegation)
 {
-	struct nfs_client *clp = NFS_SERVER(inode)->nfs_client;
 	struct nfs_inode *nfsi = NFS_I(inode);
 
 	nfs_msync_inode(inode);
-	down_read(&clp->cl_sem);
 	/* Guard against new delegated open calls */
 	down_write(&nfsi->rwsem);
 	nfs_delegation_claim_opens(inode, &delegation->stateid);
 	up_write(&nfsi->rwsem);
-	up_read(&clp->cl_sem);
 	nfs_msync_inode(inode);
 
 	return nfs_do_return_delegation(inode, delegation, 1);
 }
 
+/*
+ * Return all delegations that have been marked for return
+ */
+void nfs_client_return_marked_delegations(struct nfs_client *clp)
+{
+	struct nfs_delegation *delegation;
+	struct inode *inode;
+
+restart:
+	rcu_read_lock();
+	list_for_each_entry_rcu(delegation, &clp->cl_delegations, super_list) {
+		if (!test_and_clear_bit(NFS_DELEGATION_RETURN, &delegation->flags))
+			continue;
+		inode = nfs_delegation_grab_inode(delegation);
+		if (inode == NULL)
+			continue;
+		spin_lock(&clp->cl_lock);
+		delegation = nfs_detach_delegation_locked(NFS_I(inode), NULL);
+		spin_unlock(&clp->cl_lock);
+		rcu_read_unlock();
+		if (delegation != NULL)
+			__nfs_inode_return_delegation(inode, delegation);
+		iput(inode);
+		goto restart;
+	}
+	rcu_read_unlock();
+}
+
 /*
  * This function returns the delegation without reclaiming opens
  * or protecting against delegation reclaims.
@@ -279,83 +343,55 @@ int nfs_inode_return_delegation(struct inode *inode)
 	return err;
 }
 
+static void nfs_mark_return_delegation(struct nfs_client *clp, struct nfs_delegation *delegation)
+{
+	set_bit(NFS_DELEGATION_RETURN, &delegation->flags);
+	set_bit(NFS4CLNT_DELEGRETURN, &clp->cl_state);
+}
+
 /*
  * Return all delegations associated to a super block
  */
-void nfs_return_all_delegations(struct super_block *sb)
+void nfs_super_return_all_delegations(struct super_block *sb)
 {
 	struct nfs_client *clp = NFS_SB(sb)->nfs_client;
 	struct nfs_delegation *delegation;
-	struct inode *inode;
 
 	if (clp == NULL)
 		return;
-restart:
 	rcu_read_lock();
 	list_for_each_entry_rcu(delegation, &clp->cl_delegations, super_list) {
-		if (delegation->inode->i_sb != sb)
-			continue;
-		inode = igrab(delegation->inode);
-		if (inode == NULL)
-			continue;
-		spin_lock(&clp->cl_lock);
-		delegation = nfs_detach_delegation_locked(NFS_I(inode), NULL);
-		spin_unlock(&clp->cl_lock);
-		rcu_read_unlock();
-		if (delegation != NULL)
-			__nfs_inode_return_delegation(inode, delegation);
-		iput(inode);
-		goto restart;
+		spin_lock(&delegation->lock);
+		if (delegation->inode != NULL && delegation->inode->i_sb == sb)
+			set_bit(NFS_DELEGATION_RETURN, &delegation->flags);
+		spin_unlock(&delegation->lock);
 	}
 	rcu_read_unlock();
+	nfs_client_return_marked_delegations(clp);
 }
 
-static int nfs_do_expire_all_delegations(void *ptr)
+static void nfs_client_mark_return_all_delegations(struct nfs_client *clp)
 {
-	struct nfs_client *clp = ptr;
 	struct nfs_delegation *delegation;
-	struct inode *inode;
 
-	allow_signal(SIGKILL);
-restart:
-	if (test_bit(NFS4CLNT_STATE_RECOVER, &clp->cl_state) != 0)
-		goto out;
-	if (test_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state) == 0)
-		goto out;
 	rcu_read_lock();
 	list_for_each_entry_rcu(delegation, &clp->cl_delegations, super_list) {
-		inode = igrab(delegation->inode);
-		if (inode == NULL)
-			continue;
-		spin_lock(&clp->cl_lock);
-		delegation = nfs_detach_delegation_locked(NFS_I(inode), NULL);
-		spin_unlock(&clp->cl_lock);
-		rcu_read_unlock();
-		if (delegation)
-			__nfs_inode_return_delegation(inode, delegation);
-		iput(inode);
-		goto restart;
+		set_bit(NFS_DELEGATION_RETURN, &delegation->flags);
+		set_bit(NFS4CLNT_DELEGRETURN, &clp->cl_state);
 	}
 	rcu_read_unlock();
-out:
-	nfs_put_client(clp);
-	module_put_and_exit(0);
+}
+
+static void nfs_delegation_run_state_manager(struct nfs_client *clp)
+{
+	if (test_bit(NFS4CLNT_DELEGRETURN, &clp->cl_state))
+		nfs4_schedule_state_manager(clp);
 }
 
 void nfs_expire_all_delegations(struct nfs_client *clp)
 {
-	struct task_struct *task;
-
-	__module_get(THIS_MODULE);
-	atomic_inc(&clp->cl_count);
-	task = kthread_run(nfs_do_expire_all_delegations, clp,
-				"%s-delegreturn",
-				rpc_peeraddr2str(clp->cl_rpcclient,
-							RPC_DISPLAY_ADDR));
-	if (!IS_ERR(task))
-		return;
-	nfs_put_client(clp);
-	module_put(THIS_MODULE);
+	nfs_client_mark_return_all_delegations(clp);
+	nfs_delegation_run_state_manager(clp);
 }
 
 /*
@@ -363,68 +399,29 @@ void nfs_expire_all_delegations(struct nfs_client *clp)
  */
 void nfs_handle_cb_pathdown(struct nfs_client *clp)
 {
-	struct nfs_delegation *delegation;
-	struct inode *inode;
-
 	if (clp == NULL)
 		return;
-restart:
+	nfs_client_mark_return_all_delegations(clp);
+}
+
+static void nfs_client_mark_return_unreferenced_delegations(struct nfs_client *clp)
+{
+	struct nfs_delegation *delegation;
+
 	rcu_read_lock();
 	list_for_each_entry_rcu(delegation, &clp->cl_delegations, super_list) {
-		inode = igrab(delegation->inode);
-		if (inode == NULL)
+		if (test_and_clear_bit(NFS_DELEGATION_REFERENCED, &delegation->flags))
 			continue;
-		spin_lock(&clp->cl_lock);
-		delegation = nfs_detach_delegation_locked(NFS_I(inode), NULL);
-		spin_unlock(&clp->cl_lock);
-		rcu_read_unlock();
-		if (delegation != NULL)
-			__nfs_inode_return_delegation(inode, delegation);
-		iput(inode);
-		goto restart;
+		set_bit(NFS_DELEGATION_RETURN, &delegation->flags);
+		set_bit(NFS4CLNT_DELEGRETURN, &clp->cl_state);
 	}
 	rcu_read_unlock();
 }
 
-struct recall_threadargs {
-	struct inode *inode;
-	struct nfs_client *clp;
-	const nfs4_stateid *stateid;
-
-	struct completion started;
-	int result;
-};
-
-static int recall_thread(void *data)
+void nfs_expire_unreferenced_delegations(struct nfs_client *clp)
 {
-	struct recall_threadargs *args = (struct recall_threadargs *)data;
-	struct inode *inode = igrab(args->inode);
-	struct nfs_client *clp = NFS_SERVER(inode)->nfs_client;
-	struct nfs_inode *nfsi = NFS_I(inode);
-	struct nfs_delegation *delegation;
-
-	daemonize("nfsv4-delegreturn");
-
-	nfs_msync_inode(inode);
-	down_read(&clp->cl_sem);
-	down_write(&nfsi->rwsem);
-	spin_lock(&clp->cl_lock);
-	delegation = nfs_detach_delegation_locked(nfsi, args->stateid);
-	if (delegation != NULL)
-		args->result = 0;
-	else
-		args->result = -ENOENT;
-	spin_unlock(&clp->cl_lock);
-	complete(&args->started);
-	nfs_delegation_claim_opens(inode, args->stateid);
-	up_write(&nfsi->rwsem);
-	up_read(&clp->cl_sem);
-	nfs_msync_inode(inode);
-
-	if (delegation != NULL)
-		nfs_do_return_delegation(inode, delegation, 1);
-	iput(inode);
-	module_put_and_exit(0);
+	nfs_client_mark_return_unreferenced_delegations(clp);
+	nfs_delegation_run_state_manager(clp);
 }
 
 /*
@@ -432,22 +429,20 @@ static int recall_thread(void *data)
  */
 int nfs_async_inode_return_delegation(struct inode *inode, const nfs4_stateid *stateid)
 {
-	struct recall_threadargs data = {
-		.inode = inode,
-		.stateid = stateid,
-	};
-	int status;
+	struct nfs_client *clp = NFS_SERVER(inode)->nfs_client;
+	struct nfs_delegation *delegation;
 
-	init_completion(&data.started);
-	__module_get(THIS_MODULE);
-	status = kernel_thread(recall_thread, &data, CLONE_KERNEL);
-	if (status < 0)
-		goto out_module_put;
-	wait_for_completion(&data.started);
-	return data.result;
-out_module_put:
-	module_put(THIS_MODULE);
-	return status;
+	rcu_read_lock();
+	delegation = rcu_dereference(NFS_I(inode)->delegation);
+	if (delegation == NULL || memcmp(delegation->stateid.data, stateid->data,
+				sizeof(delegation->stateid.data)) != 0) {
+		rcu_read_unlock();
+		return -ENOENT;
+	}
+	nfs_mark_return_delegation(clp, delegation);
+	rcu_read_unlock();
+	nfs_delegation_run_state_manager(clp);
+	return 0;
 }
 
 /*
@@ -459,10 +454,14 @@ struct inode *nfs_delegation_find_inode(struct nfs_client *clp, const struct nfs
 	struct inode *res = NULL;
 	rcu_read_lock();
 	list_for_each_entry_rcu(delegation, &clp->cl_delegations, super_list) {
-		if (nfs_compare_fh(fhandle, &NFS_I(delegation->inode)->fh) == 0) {
+		spin_lock(&delegation->lock);
+		if (delegation->inode != NULL &&
+		    nfs_compare_fh(fhandle, &NFS_I(delegation->inode)->fh) == 0) {
 			res = igrab(delegation->inode);
-			break;
 		}
+		spin_unlock(&delegation->lock);
+		if (res != NULL)
+			break;
 	}
 	rcu_read_unlock();
 	return res;
@@ -476,7 +475,7 @@ void nfs_delegation_mark_reclaim(struct nfs_client *clp)
 	struct nfs_delegation *delegation;
 	rcu_read_lock();
 	list_for_each_entry_rcu(delegation, &clp->cl_delegations, super_list)
-		delegation->flags |= NFS_DELEGATION_NEED_RECLAIM;
+		set_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags);
 	rcu_read_unlock();
 }
 
@@ -486,17 +485,22 @@ void nfs_delegation_mark_reclaim(struct nfs_client *clp)
 void nfs_delegation_reap_unclaimed(struct nfs_client *clp)
 {
 	struct nfs_delegation *delegation;
+	struct inode *inode;
 restart:
 	rcu_read_lock();
 	list_for_each_entry_rcu(delegation, &clp->cl_delegations, super_list) {
-		if ((delegation->flags & NFS_DELEGATION_NEED_RECLAIM) == 0)
+		if (test_bit(NFS_DELEGATION_NEED_RECLAIM, &delegation->flags) == 0)
+			continue;
+		inode = nfs_delegation_grab_inode(delegation);
+		if (inode == NULL)
 			continue;
 		spin_lock(&clp->cl_lock);
-		delegation = nfs_detach_delegation_locked(NFS_I(delegation->inode), NULL);
+		delegation = nfs_detach_delegation_locked(NFS_I(inode), NULL);
 		spin_unlock(&clp->cl_lock);
 		rcu_read_unlock();
 		if (delegation != NULL)
 			nfs_free_delegation(delegation);
+		iput(inode);
 		goto restart;
 	}
 	rcu_read_unlock();

fs/nfs/delegation.h

@@ -17,14 +17,20 @@ struct nfs_delegation {
 	struct rpc_cred *cred;
 	struct inode *inode;
 	nfs4_stateid stateid;
-	int type;
-#define NFS_DELEGATION_NEED_RECLAIM 1
-	long flags;
+	fmode_t type;
 	loff_t maxsize;
 	__u64 change_attr;
+	unsigned long flags;
+	spinlock_t lock;
 	struct rcu_head rcu;
 };
 
+enum {
+	NFS_DELEGATION_NEED_RECLAIM = 0,
+	NFS_DELEGATION_RETURN,
+	NFS_DELEGATION_REFERENCED,
+};
+
 int nfs_inode_set_delegation(struct inode *inode, struct rpc_cred *cred, struct nfs_openres *res);
 void nfs_inode_reclaim_delegation(struct inode *inode, struct rpc_cred *cred, struct nfs_openres *res);
 int nfs_inode_return_delegation(struct inode *inode);
@@ -32,9 +38,11 @@ int nfs_async_inode_return_delegation(struct inode *inode, const nfs4_stateid *s
 void nfs_inode_return_delegation_noreclaim(struct inode *inode);
 
 struct inode *nfs_delegation_find_inode(struct nfs_client *clp, const struct nfs_fh *fhandle);
-void nfs_return_all_delegations(struct super_block *sb);
+void nfs_super_return_all_delegations(struct super_block *sb);
 void nfs_expire_all_delegations(struct nfs_client *clp);
+void nfs_expire_unreferenced_delegations(struct nfs_client *clp);
 void nfs_handle_cb_pathdown(struct nfs_client *clp);
+void nfs_client_return_marked_delegations(struct nfs_client *clp);
 
 void nfs_delegation_mark_reclaim(struct nfs_client *clp);
 void nfs_delegation_reap_unclaimed(struct nfs_client *clp);
@@ -45,22 +53,11 @@ int nfs4_open_delegation_recall(struct nfs_open_context *ctx, struct nfs4_state
 int nfs4_lock_delegation_recall(struct nfs4_state *state, struct file_lock *fl);
 int nfs4_copy_delegation_stateid(nfs4_stateid *dst, struct inode *inode);
 
-static inline int nfs_have_delegation(struct inode *inode, int flags)
-{
-	struct nfs_delegation *delegation;
-	int ret = 0;
-
-	flags &= FMODE_READ|FMODE_WRITE;
-	rcu_read_lock();
-	delegation = rcu_dereference(NFS_I(inode)->delegation);
-	if (delegation != NULL && (delegation->type & flags) == flags)
-		ret = 1;
-	rcu_read_unlock();
-	return ret;
-}
+void nfs_mark_delegation_referenced(struct nfs_delegation *delegation);
+int nfs_have_delegation(struct inode *inode, fmode_t flags);
 
 #else
-static inline int nfs_have_delegation(struct inode *inode, int flags)
+static inline int nfs_have_delegation(struct inode *inode, fmode_t flags)
 {
 	return 0;
 }

fs/nfs/dir.c

@@ -799,6 +799,9 @@ static int nfs_lookup_revalidate(struct dentry * dentry, struct nameidata *nd)
 		goto out_bad;
 	}
 
+	if (nfs_have_delegation(inode, FMODE_READ))
+		goto out_set_verifier;
+
 	/* Force a full look up iff the parent directory has changed */
 	if (!nfs_is_exclusive_create(dir, nd) && nfs_check_verifier(dir, dentry)) {
 		if (nfs_lookup_verify_inode(inode, nd))
@@ -817,6 +820,7 @@ static int nfs_lookup_revalidate(struct dentry * dentry, struct nameidata *nd)
 	if ((error = nfs_refresh_inode(inode, &fattr)) != 0)
 		goto out_bad;
 
+out_set_verifier:
 	nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
  out_valid:
 	dput(parent);
@@ -973,7 +977,7 @@ struct dentry_operations nfs4_dentry_operations = {
  * Use intent information to determine whether we need to substitute
  * the NFSv4-style stateful OPEN for the LOOKUP call
  */
-static int is_atomic_open(struct inode *dir, struct nameidata *nd)
+static int is_atomic_open(struct nameidata *nd)
 {
 	if (nd == NULL || nfs_lookup_check_intent(nd, LOOKUP_OPEN) == 0)
 		return 0;
@@ -996,7 +1000,7 @@ static struct dentry *nfs_atomic_lookup(struct inode *dir, struct dentry *dentry
 			dir->i_sb->s_id, dir->i_ino, dentry->d_name.name);
 
 	/* Check that we are indeed trying to open this file */
-	if (!is_atomic_open(dir, nd))
+	if (!is_atomic_open(nd))
 		goto no_open;
 
 	if (dentry->d_name.len > NFS_SERVER(dir)->namelen) {
@@ -1047,10 +1051,10 @@ static int nfs_open_revalidate(struct dentry *dentry, struct nameidata *nd)
 	struct inode *dir;
 	int openflags, ret = 0;
 
+	if (!is_atomic_open(nd))
+		goto no_open;
 	parent = dget_parent(dentry);
 	dir = parent->d_inode;
-	if (!is_atomic_open(dir, nd))
-		goto no_open;
 	/* We can't create new files in nfs_open_revalidate(), so we
 	 * optimize away revalidation of negative dentries.
 	 */
@@ -1062,11 +1066,11 @@ static int nfs_open_revalidate(struct dentry *dentry, struct nameidata *nd)
 
 	/* NFS only supports OPEN on regular files */
 	if (!S_ISREG(inode->i_mode))
-		goto no_open;
+		goto no_open_dput;
 	openflags = nd->intent.open.flags;
 	/* We cannot do exclusive creation on a positive dentry */
 	if ((openflags & (O_CREAT|O_EXCL)) == (O_CREAT|O_EXCL))
-		goto no_open;
+		goto no_open_dput;
 	/* We can't create new files, or truncate existing ones here */
 	openflags &= ~(O_CREAT|O_TRUNC);
 
@@ -1081,10 +1085,9 @@ static int nfs_open_revalidate(struct dentry *dentry, struct nameidata *nd)
 	if (!ret)
 		d_drop(dentry);
 	return ret;
-no_open:
+no_open_dput:
 	dput(parent);
-	if (inode != NULL && nfs_have_delegation(inode, FMODE_READ))
-		return 1;
+no_open:
 	return nfs_lookup_revalidate(dentry, nd);
 }
 #endif /* CONFIG_NFSV4 */
@@ -1794,7 +1797,8 @@ static int nfs_access_get_cached(struct inode *inode, struct rpc_cred *cred, str
 	cache = nfs_access_search_rbtree(inode, cred);
 	if (cache == NULL)
 		goto out;
-	if (!time_in_range(jiffies, cache->jiffies, cache->jiffies + nfsi->attrtimeo))
+	if (!nfs_have_delegation(inode, FMODE_READ) &&
+	    !time_in_range_open(jiffies, cache->jiffies, cache->jiffies + nfsi->attrtimeo))
 		goto out_stale;
 	res->jiffies = cache->jiffies;
 	res->cred = cache->cred;

fs/nfs/inode.c

@@ -592,7 +592,7 @@ static void nfs_file_set_open_context(struct file *filp, struct nfs_open_context
 /*
  * Given an inode, search for an open context with the desired characteristics
  */
-struct nfs_open_context *nfs_find_open_context(struct inode *inode, struct rpc_cred *cred, int mode)
+struct nfs_open_context *nfs_find_open_context(struct inode *inode, struct rpc_cred *cred, fmode_t mode)
 {
 	struct nfs_inode *nfsi = NFS_I(inode);
 	struct nfs_open_context *pos, *ctx = NULL;
@@ -712,14 +712,7 @@ int nfs_attribute_timeout(struct inode *inode)
 
 	if (nfs_have_delegation(inode, FMODE_READ))
 		return 0;
-	/*
-	 * Special case: if the attribute timeout is set to 0, then always
-	 * 		 treat the cache as having expired (unless holding
-	 * 		 a delegation).
-	 */
-	if (nfsi->attrtimeo == 0)
-		return 1;
-	return !time_in_range(jiffies, nfsi->read_cache_jiffies, nfsi->read_cache_jiffies + nfsi->attrtimeo);
+	return !time_in_range_open(jiffies, nfsi->read_cache_jiffies, nfsi->read_cache_jiffies + nfsi->attrtimeo);
 }
 
 /**
@@ -1182,7 +1175,7 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 		nfsi->attrtimeo_timestamp = now;
 		nfsi->attr_gencount = nfs_inc_attr_generation_counter();
 	} else {
-		if (!time_in_range(now, nfsi->attrtimeo_timestamp, nfsi->attrtimeo_timestamp + nfsi->attrtimeo)) {
+		if (!time_in_range_open(now, nfsi->attrtimeo_timestamp, nfsi->attrtimeo_timestamp + nfsi->attrtimeo)) {
 			if ((nfsi->attrtimeo <<= 1) > NFS_MAXATTRTIMEO(inode))
 				nfsi->attrtimeo = NFS_MAXATTRTIMEO(inode);
 			nfsi->attrtimeo_timestamp = now;

fs/nfs/internal.h

@@ -63,6 +63,20 @@ struct nfs_parsed_mount_data {
 	struct security_mnt_opts lsm_opts;
 };
 
+/* mount_clnt.c */
+struct nfs_mount_request {
+	struct sockaddr		*sap;
+	size_t			salen;
+	char			*hostname;
+	char			*dirpath;
+	u32			version;
+	unsigned short		protocol;
+	struct nfs_fh		*fh;
+	int			noresvport;
+};
+
+extern int nfs_mount(struct nfs_mount_request *info);
+
 /* client.c */
 extern struct rpc_program nfs_program;
 

fs/nfs/mount_clnt.c

@@ -29,47 +29,43 @@ struct mnt_fhstatus {
 
 /**
  * nfs_mount - Obtain an NFS file handle for the given host and path
- * @addr: pointer to server's address
- * @len: size of server's address
- * @hostname: name of server host, or NULL
- * @path: pointer to string containing export path to mount
- * @version: mount version to use for this request
- * @protocol: transport protocol to use for thie request
- * @fh: pointer to location to place returned file handle
+ * @info: pointer to mount request arguments
  *
  * Uses default timeout parameters specified by underlying transport.
  */
-int nfs_mount(struct sockaddr *addr, size_t len, char *hostname, char *path,
-	      int version, int protocol, struct nfs_fh *fh)
+int nfs_mount(struct nfs_mount_request *info)
 {
 	struct mnt_fhstatus	result = {
-		.fh		= fh
+		.fh		= info->fh
 	};
 	struct rpc_message msg	= {
-		.rpc_argp	= path,
+		.rpc_argp	= info->dirpath,
 		.rpc_resp	= &result,
 	};
 	struct rpc_create_args args = {
-		.protocol	= protocol,
-		.address	= addr,
-		.addrsize	= len,
-		.servername	= hostname,
+		.protocol	= info->protocol,
+		.address	= info->sap,
+		.addrsize	= info->salen,
+		.servername	= info->hostname,
 		.program	= &mnt_program,
-		.version	= version,
+		.version	= info->version,
 		.authflavor	= RPC_AUTH_UNIX,
-		.flags		= 0,
 	};
 	struct rpc_clnt		*mnt_clnt;
 	int			status;
 
 	dprintk("NFS: sending MNT request for %s:%s\n",
-		(hostname ? hostname : "server"), path);
+		(info->hostname ? info->hostname : "server"),
+			info->dirpath);
+
+	if (info->noresvport)
+		args.flags |= RPC_CLNT_CREATE_NONPRIVPORT;
 
 	mnt_clnt = rpc_create(&args);
 	if (IS_ERR(mnt_clnt))
 		goto out_clnt_err;
 
-	if (version == NFS_MNT3_VERSION)
+	if (info->version == NFS_MNT3_VERSION)
 		msg.rpc_proc = &mnt_clnt->cl_procinfo[MOUNTPROC3_MNT];
 	else
 		msg.rpc_proc = &mnt_clnt->cl_procinfo[MNTPROC_MNT];

fs/nfs/nfs4_fs.h

@@ -38,8 +38,12 @@ struct idmap;
  ((err) != NFSERR_NOFILEHANDLE))
 
 enum nfs4_client_state {
-	NFS4CLNT_STATE_RECOVER  = 0,
+	NFS4CLNT_MANAGER_RUNNING  = 0,
+	NFS4CLNT_CHECK_LEASE,
 	NFS4CLNT_LEASE_EXPIRED,
+	NFS4CLNT_RECLAIM_REBOOT,
+	NFS4CLNT_RECLAIM_NOGRACE,
+	NFS4CLNT_DELEGRETURN,
 };
 
 /*
@@ -90,12 +94,18 @@ struct nfs4_state_owner {
 
 	spinlock_t	     so_lock;
 	atomic_t	     so_count;
+	unsigned long	     so_flags;
 	struct list_head     so_states;
 	struct list_head     so_delegations;
 	struct nfs_seqid_counter so_seqid;
 	struct rpc_sequence  so_sequence;
 };
 
+enum {
+	NFS_OWNER_RECLAIM_REBOOT,
+	NFS_OWNER_RECLAIM_NOGRACE
+};
+
 /*
  * struct nfs4_state maintains the client-side state for a given
  * (state_owner,inode) tuple (OPEN) or state_owner (LOCK).
@@ -128,6 +138,8 @@ enum {
 	NFS_O_RDONLY_STATE,		/* OPEN stateid has read-only state */
 	NFS_O_WRONLY_STATE,		/* OPEN stateid has write-only state */
 	NFS_O_RDWR_STATE,		/* OPEN stateid has read/write state */
+	NFS_STATE_RECLAIM_REBOOT,	/* OPEN stateid server rebooted */
+	NFS_STATE_RECLAIM_NOGRACE,	/* OPEN stateid needs to recover state */
 };
 
 struct nfs4_state {
@@ -149,7 +161,7 @@ struct nfs4_state {
 	unsigned int n_rdonly;		/* Number of read-only references */
 	unsigned int n_wronly;		/* Number of write-only references */
 	unsigned int n_rdwr;		/* Number of read/write references */
-	int state;			/* State on the server (R,W, or RW) */
+	fmode_t state;			/* State on the server (R,W, or RW) */
 	atomic_t count;
 };
 
@@ -157,9 +169,12 @@ struct nfs4_state {
 struct nfs4_exception {
 	long timeout;
 	int retry;
+	struct nfs4_state *state;
 };
 
 struct nfs4_state_recovery_ops {
+	int owner_flag_bit;
+	int state_flag_bit;
 	int (*recover_open)(struct nfs4_state_owner *, struct nfs4_state *);
 	int (*recover_lock)(struct nfs4_state *, struct file_lock *);
 };
@@ -174,7 +189,6 @@ extern ssize_t nfs4_listxattr(struct dentry *, char *, size_t);
 
 
 /* nfs4proc.c */
-extern int nfs4_map_errors(int err);
 extern int nfs4_proc_setclientid(struct nfs_client *, u32, unsigned short, struct rpc_cred *);
 extern int nfs4_proc_setclientid_confirm(struct nfs_client *, struct rpc_cred *);
 extern int nfs4_proc_async_renew(struct nfs_client *, struct rpc_cred *);
@@ -187,7 +201,7 @@ extern int nfs4_proc_fs_locations(struct inode *dir, const struct qstr *name,
 		struct nfs4_fs_locations *fs_locations, struct page *page);
 
 extern struct nfs4_state_recovery_ops nfs4_reboot_recovery_ops;
-extern struct nfs4_state_recovery_ops nfs4_network_partition_recovery_ops;
+extern struct nfs4_state_recovery_ops nfs4_nograce_recovery_ops;
 
 extern const u32 nfs4_fattr_bitmap[2];
 extern const u32 nfs4_statfs_bitmap[2];
@@ -202,16 +216,18 @@ extern void nfs4_kill_renewd(struct nfs_client *);
 extern void nfs4_renew_state(struct work_struct *);
 
 /* nfs4state.c */
-struct rpc_cred *nfs4_get_renew_cred(struct nfs_client *clp);
+struct rpc_cred *nfs4_get_renew_cred_locked(struct nfs_client *clp);
 
 extern struct nfs4_state_owner * nfs4_get_state_owner(struct nfs_server *, struct rpc_cred *);
 extern void nfs4_put_state_owner(struct nfs4_state_owner *);
 extern struct nfs4_state * nfs4_get_open_state(struct inode *, struct nfs4_state_owner *);
 extern void nfs4_put_open_state(struct nfs4_state *);
-extern void nfs4_close_state(struct path *, struct nfs4_state *, mode_t);
-extern void nfs4_close_sync(struct path *, struct nfs4_state *, mode_t);
-extern void nfs4_state_set_mode_locked(struct nfs4_state *, mode_t);
+extern void nfs4_close_state(struct path *, struct nfs4_state *, fmode_t);
+extern void nfs4_close_sync(struct path *, struct nfs4_state *, fmode_t);
+extern void nfs4_state_set_mode_locked(struct nfs4_state *, fmode_t);
 extern void nfs4_schedule_state_recovery(struct nfs_client *);
+extern void nfs4_schedule_state_manager(struct nfs_client *);
+extern int nfs4_state_mark_reclaim_nograce(struct nfs_client *clp, struct nfs4_state *state);
 extern void nfs4_put_lock_state(struct nfs4_lock_state *lsp);
 extern int nfs4_set_lock_state(struct nfs4_state *state, struct file_lock *fl);
 extern void nfs4_copy_stateid(nfs4_stateid *, struct nfs4_state *, fl_owner_t);

fs/nfs/nfs4renewd.c

@@ -65,7 +65,6 @@ nfs4_renew_state(struct work_struct *work)
 	long lease, timeout;
 	unsigned long last, now;
 
-	down_read(&clp->cl_sem);
 	dprintk("%s: start\n", __func__);
 	/* Are there any active superblocks? */
 	if (list_empty(&clp->cl_superblocks))
@@ -77,17 +76,19 @@ nfs4_renew_state(struct work_struct *work)
 	timeout = (2 * lease) / 3 + (long)last - (long)now;
 	/* Are we close to a lease timeout? */
 	if (time_after(now, last + lease/3)) {
-		cred = nfs4_get_renew_cred(clp);
+		cred = nfs4_get_renew_cred_locked(clp);
+		spin_unlock(&clp->cl_lock);
 		if (cred == NULL) {
-			set_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state);
-			spin_unlock(&clp->cl_lock);
+			if (list_empty(&clp->cl_delegations)) {
+				set_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state);
+				goto out;
+			}
 			nfs_expire_all_delegations(clp);
-			goto out;
+		} else {
+			/* Queue an asynchronous RENEW. */
+			nfs4_proc_async_renew(clp, cred);
+			put_rpccred(cred);
 		}
-		spin_unlock(&clp->cl_lock);
-		/* Queue an asynchronous RENEW. */
-		nfs4_proc_async_renew(clp, cred);
-		put_rpccred(cred);
 		timeout = (2 * lease) / 3;
 		spin_lock(&clp->cl_lock);
 	} else
@@ -100,12 +101,11 @@ nfs4_renew_state(struct work_struct *work)
 	cancel_delayed_work(&clp->cl_renewd);
 	schedule_delayed_work(&clp->cl_renewd, timeout);
 	spin_unlock(&clp->cl_lock);
+	nfs_expire_unreferenced_delegations(clp);
 out:
-	up_read(&clp->cl_sem);
 	dprintk("%s: done\n", __func__);
 }
 
-/* Must be called with clp->cl_sem locked for writes */
 void
 nfs4_schedule_state_renewal(struct nfs_client *clp)
 {

fs/nfs/nfs4state.c

@@ -71,14 +71,12 @@ static int nfs4_init_client(struct nfs_client *clp, struct rpc_cred *cred)
 	return status;
 }
 
-static struct rpc_cred *nfs4_get_machine_cred(struct nfs_client *clp)
+static struct rpc_cred *nfs4_get_machine_cred_locked(struct nfs_client *clp)
 {
 	struct rpc_cred *cred = NULL;
 
-	spin_lock(&clp->cl_lock);
 	if (clp->cl_machine_cred != NULL)
 		cred = get_rpccred(clp->cl_machine_cred);
-	spin_unlock(&clp->cl_lock);
 	return cred;
 }
 
@@ -94,7 +92,7 @@ static void nfs4_clear_machine_cred(struct nfs_client *clp)
 		put_rpccred(cred);
 }
 
-struct rpc_cred *nfs4_get_renew_cred(struct nfs_client *clp)
+struct rpc_cred *nfs4_get_renew_cred_locked(struct nfs_client *clp)
 {
 	struct nfs4_state_owner *sp;
 	struct rb_node *pos;
@@ -110,13 +108,24 @@ struct rpc_cred *nfs4_get_renew_cred(struct nfs_client *clp)
 	return cred;
 }
 
+static struct rpc_cred *nfs4_get_renew_cred(struct nfs_client *clp)
+{
+	struct rpc_cred *cred;
+
+	spin_lock(&clp->cl_lock);
+	cred = nfs4_get_renew_cred_locked(clp);
+	spin_unlock(&clp->cl_lock);
+	return cred;
+}
+
 static struct rpc_cred *nfs4_get_setclientid_cred(struct nfs_client *clp)
 {
 	struct nfs4_state_owner *sp;
 	struct rb_node *pos;
 	struct rpc_cred *cred;
 
-	cred = nfs4_get_machine_cred(clp);
+	spin_lock(&clp->cl_lock);
+	cred = nfs4_get_machine_cred_locked(clp);
 	if (cred != NULL)
 		goto out;
 	pos = rb_first(&clp->cl_state_owners);
@@ -125,6 +134,7 @@ static struct rpc_cred *nfs4_get_setclientid_cred(struct nfs_client *clp)
 		cred = get_rpccred(sp->so_cred);
 	}
 out:
+	spin_unlock(&clp->cl_lock);
 	return cred;
 }
 
@@ -295,10 +305,6 @@ nfs4_drop_state_owner(struct nfs4_state_owner *sp)
 	}
 }
 
-/*
- * Note: must be called with clp->cl_sem held in order to prevent races
- *       with reboot recovery!
- */
 struct nfs4_state_owner *nfs4_get_state_owner(struct nfs_server *server, struct rpc_cred *cred)
 {
 	struct nfs_client *clp = server->nfs_client;
@@ -327,10 +333,6 @@ struct nfs4_state_owner *nfs4_get_state_owner(struct nfs_server *server, struct
 	return sp;
 }
 
-/*
- * Must be called with clp->cl_sem held in order to avoid races
- * with state recovery...
- */
 void nfs4_put_state_owner(struct nfs4_state_owner *sp)
 {
 	struct nfs_client *clp = sp->so_client;
@@ -361,18 +363,18 @@ nfs4_alloc_open_state(void)
 }
 
 void
-nfs4_state_set_mode_locked(struct nfs4_state *state, mode_t mode)
+nfs4_state_set_mode_locked(struct nfs4_state *state, fmode_t fmode)
 {
-	if (state->state == mode)
+	if (state->state == fmode)
 		return;
 	/* NB! List reordering - see the reclaim code for why.  */
-	if ((mode & FMODE_WRITE) != (state->state & FMODE_WRITE)) {
-		if (mode & FMODE_WRITE)
+	if ((fmode & FMODE_WRITE) != (state->state & FMODE_WRITE)) {
+		if (fmode & FMODE_WRITE)
 			list_move(&state->open_states, &state->owner->so_states);
 		else
 			list_move_tail(&state->open_states, &state->owner->so_states);
 	}
-	state->state = mode;
+	state->state = fmode;
 }
 
 static struct nfs4_state *
@@ -432,10 +434,6 @@ nfs4_get_open_state(struct inode *inode, struct nfs4_state_owner *owner)
 	return state;
 }
 
-/*
- * Beware! Caller must be holding exactly one
- * reference to clp->cl_sem!
- */
 void nfs4_put_open_state(struct nfs4_state *state)
 {
 	struct inode *inode = state->inode;
@@ -456,16 +454,16 @@ void nfs4_put_open_state(struct nfs4_state *state)
 /*
  * Close the current file.
  */
-static void __nfs4_close(struct path *path, struct nfs4_state *state, mode_t mode, int wait)
+static void __nfs4_close(struct path *path, struct nfs4_state *state, fmode_t fmode, int wait)
 {
 	struct nfs4_state_owner *owner = state->owner;
 	int call_close = 0;
-	int newstate;
+	fmode_t newstate;
 
 	atomic_inc(&owner->so_count);
 	/* Protect against nfs4_find_state() */
 	spin_lock(&owner->so_lock);
-	switch (mode & (FMODE_READ | FMODE_WRITE)) {
+	switch (fmode & (FMODE_READ | FMODE_WRITE)) {
 		case FMODE_READ:
 			state->n_rdonly--;
 			break;
@@ -500,14 +498,14 @@ static void __nfs4_close(struct path *path, struct nfs4_state *state, mode_t mod
 		nfs4_do_close(path, state, wait);
 }
 
-void nfs4_close_state(struct path *path, struct nfs4_state *state, mode_t mode)
+void nfs4_close_state(struct path *path, struct nfs4_state *state, fmode_t fmode)
 {
-	__nfs4_close(path, state, mode, 0);
+	__nfs4_close(path, state, fmode, 0);
 }
 
-void nfs4_close_sync(struct path *path, struct nfs4_state *state, mode_t mode)
+void nfs4_close_sync(struct path *path, struct nfs4_state *state, fmode_t fmode)
 {
-	__nfs4_close(path, state, mode, 1);
+	__nfs4_close(path, state, fmode, 1);
 }
 
 /*
@@ -568,7 +566,6 @@ static void nfs4_free_lock_state(struct nfs4_lock_state *lsp)
  * Return a compatible lock_state. If no initialized lock_state structure
  * exists, return an uninitialized one.
  *
- * The caller must be holding clp->cl_sem
  */
 static struct nfs4_lock_state *nfs4_get_lock_state(struct nfs4_state *state, fl_owner_t owner)
 {
@@ -770,32 +767,34 @@ int nfs_wait_on_sequence(struct nfs_seqid *seqid, struct rpc_task *task)
 	return status;
 }
 
-static int reclaimer(void *);
+static int nfs4_run_state_manager(void *);
 
-static inline void nfs4_clear_recover_bit(struct nfs_client *clp)
+static void nfs4_clear_state_manager_bit(struct nfs_client *clp)
 {
 	smp_mb__before_clear_bit();
-	clear_bit(NFS4CLNT_STATE_RECOVER, &clp->cl_state);
+	clear_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state);
 	smp_mb__after_clear_bit();
-	wake_up_bit(&clp->cl_state, NFS4CLNT_STATE_RECOVER);
+	wake_up_bit(&clp->cl_state, NFS4CLNT_MANAGER_RUNNING);
 	rpc_wake_up(&clp->cl_rpcwaitq);
 }
 
 /*
- * State recovery routine
+ * Schedule the nfs_client asynchronous state management routine
  */
-static void nfs4_recover_state(struct nfs_client *clp)
+void nfs4_schedule_state_manager(struct nfs_client *clp)
 {
 	struct task_struct *task;
 
+	if (test_and_set_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state) != 0)
+		return;
 	__module_get(THIS_MODULE);
 	atomic_inc(&clp->cl_count);
-	task = kthread_run(reclaimer, clp, "%s-reclaim",
+	task = kthread_run(nfs4_run_state_manager, clp, "%s-manager",
 				rpc_peeraddr2str(clp->cl_rpcclient,
 							RPC_DISPLAY_ADDR));
 	if (!IS_ERR(task))
 		return;
-	nfs4_clear_recover_bit(clp);
+	nfs4_clear_state_manager_bit(clp);
 	nfs_put_client(clp);
 	module_put(THIS_MODULE);
 }
@@ -807,16 +806,42 @@ void nfs4_schedule_state_recovery(struct nfs_client *clp)
 {
 	if (!clp)
 		return;
-	if (test_and_set_bit(NFS4CLNT_STATE_RECOVER, &clp->cl_state) == 0)
-		nfs4_recover_state(clp);
+	if (!test_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state))
+		set_bit(NFS4CLNT_CHECK_LEASE, &clp->cl_state);
+	nfs4_schedule_state_manager(clp);
 }
 
-static int nfs4_reclaim_locks(struct nfs4_state_recovery_ops *ops, struct nfs4_state *state)
+static int nfs4_state_mark_reclaim_reboot(struct nfs_client *clp, struct nfs4_state *state)
+{
+
+	set_bit(NFS_STATE_RECLAIM_REBOOT, &state->flags);
+	/* Don't recover state that expired before the reboot */
+	if (test_bit(NFS_STATE_RECLAIM_NOGRACE, &state->flags)) {
+		clear_bit(NFS_STATE_RECLAIM_REBOOT, &state->flags);
+		return 0;
+	}
+	set_bit(NFS_OWNER_RECLAIM_REBOOT, &state->owner->so_flags);
+	set_bit(NFS4CLNT_RECLAIM_REBOOT, &clp->cl_state);
+	return 1;
+}
+
+int nfs4_state_mark_reclaim_nograce(struct nfs_client *clp, struct nfs4_state *state)
+{
+	set_bit(NFS_STATE_RECLAIM_NOGRACE, &state->flags);
+	clear_bit(NFS_STATE_RECLAIM_REBOOT, &state->flags);
+	set_bit(NFS_OWNER_RECLAIM_NOGRACE, &state->owner->so_flags);
+	set_bit(NFS4CLNT_RECLAIM_NOGRACE, &clp->cl_state);
+	return 1;
+}
+
+static int nfs4_reclaim_locks(struct nfs4_state *state, const struct nfs4_state_recovery_ops *ops)
 {
 	struct inode *inode = state->inode;
+	struct nfs_inode *nfsi = NFS_I(inode);
 	struct file_lock *fl;
 	int status = 0;
 
+	down_write(&nfsi->rwsem);
 	for (fl = inode->i_flock; fl != NULL; fl = fl->fl_next) {
 		if (!(fl->fl_flags & (FL_POSIX|FL_FLOCK)))
 			continue;
@@ -839,12 +864,14 @@ static int nfs4_reclaim_locks(struct nfs4_state_recovery_ops *ops, struct nfs4_s
 				goto out_err;
 		}
 	}
+	up_write(&nfsi->rwsem);
 	return 0;
 out_err:
+	up_write(&nfsi->rwsem);
 	return status;
 }
 
-static int nfs4_reclaim_open_state(struct nfs4_state_recovery_ops *ops, struct nfs4_state_owner *sp)
+static int nfs4_reclaim_open_state(struct nfs4_state_owner *sp, const struct nfs4_state_recovery_ops *ops)
 {
 	struct nfs4_state *state;
 	struct nfs4_lock_state *lock;
@@ -858,28 +885,34 @@ static int nfs4_reclaim_open_state(struct nfs4_state_recovery_ops *ops, struct n
 	 * recovering after a network partition or a reboot from a
 	 * server that doesn't support a grace period.
 	 */
+restart:
+	spin_lock(&sp->so_lock);
 	list_for_each_entry(state, &sp->so_states, open_states) {
+		if (!test_and_clear_bit(ops->state_flag_bit, &state->flags))
+			continue;
 		if (state->state == 0)
 			continue;
+		atomic_inc(&state->count);
+		spin_unlock(&sp->so_lock);
 		status = ops->recover_open(sp, state);
 		if (status >= 0) {
-			status = nfs4_reclaim_locks(ops, state);
-			if (status < 0)
-				goto out_err;
-			list_for_each_entry(lock, &state->lock_states, ls_locks) {
-				if (!(lock->ls_flags & NFS_LOCK_INITIALIZED))
-					printk("%s: Lock reclaim failed!\n",
+			status = nfs4_reclaim_locks(state, ops);
+			if (status >= 0) {
+				list_for_each_entry(lock, &state->lock_states, ls_locks) {
+					if (!(lock->ls_flags & NFS_LOCK_INITIALIZED))
+						printk("%s: Lock reclaim failed!\n",
 							__func__);
+				}
+				nfs4_put_open_state(state);
+				goto restart;
 			}
-			continue;
 		}
 		switch (status) {
 			default:
 				printk(KERN_ERR "%s: unhandled error %d. Zeroing state\n",
 						__func__, status);
 			case -ENOENT:
-			case -NFS4ERR_RECLAIM_BAD:
-			case -NFS4ERR_RECLAIM_CONFLICT:
+			case -ESTALE:
 				/*
 				 * Open state on this file cannot be recovered
 				 * All we can do is revert to using the zero stateid.
@@ -889,84 +922,176 @@ static int nfs4_reclaim_open_state(struct nfs4_state_recovery_ops *ops, struct n
 				/* Mark the file as being 'closed' */
 				state->state = 0;
 				break;
+			case -NFS4ERR_RECLAIM_BAD:
+			case -NFS4ERR_RECLAIM_CONFLICT:
+				nfs4_state_mark_reclaim_nograce(sp->so_client, state);
+				break;
 			case -NFS4ERR_EXPIRED:
 			case -NFS4ERR_NO_GRACE:
+				nfs4_state_mark_reclaim_nograce(sp->so_client, state);
 			case -NFS4ERR_STALE_CLIENTID:
 				goto out_err;
 		}
+		nfs4_put_open_state(state);
+		goto restart;
 	}
+	spin_unlock(&sp->so_lock);
 	return 0;
 out_err:
+	nfs4_put_open_state(state);
 	return status;
 }
 
-static void nfs4_state_mark_reclaim(struct nfs_client *clp)
+static void nfs4_clear_open_state(struct nfs4_state *state)
+{
+	struct nfs4_lock_state *lock;
+
+	clear_bit(NFS_DELEGATED_STATE, &state->flags);
+	clear_bit(NFS_O_RDONLY_STATE, &state->flags);
+	clear_bit(NFS_O_WRONLY_STATE, &state->flags);
+	clear_bit(NFS_O_RDWR_STATE, &state->flags);
+	list_for_each_entry(lock, &state->lock_states, ls_locks) {
+		lock->ls_seqid.flags = 0;
+		lock->ls_flags &= ~NFS_LOCK_INITIALIZED;
+	}
+}
+
+static void nfs4_state_mark_reclaim_helper(struct nfs_client *clp, int (*mark_reclaim)(struct nfs_client *clp, struct nfs4_state *state))
 {
 	struct nfs4_state_owner *sp;
 	struct rb_node *pos;
 	struct nfs4_state *state;
-	struct nfs4_lock_state *lock;
 
 	/* Reset all sequence ids to zero */
 	for (pos = rb_first(&clp->cl_state_owners); pos != NULL; pos = rb_next(pos)) {
 		sp = rb_entry(pos, struct nfs4_state_owner, so_client_node);
-		sp->so_seqid.counter = 0;
 		sp->so_seqid.flags = 0;
 		spin_lock(&sp->so_lock);
 		list_for_each_entry(state, &sp->so_states, open_states) {
-			clear_bit(NFS_DELEGATED_STATE, &state->flags);
-			clear_bit(NFS_O_RDONLY_STATE, &state->flags);
-			clear_bit(NFS_O_WRONLY_STATE, &state->flags);
-			clear_bit(NFS_O_RDWR_STATE, &state->flags);
-			list_for_each_entry(lock, &state->lock_states, ls_locks) {
-				lock->ls_seqid.counter = 0;
-				lock->ls_seqid.flags = 0;
-				lock->ls_flags &= ~NFS_LOCK_INITIALIZED;
-			}
+			if (mark_reclaim(clp, state))
+				nfs4_clear_open_state(state);
 		}
 		spin_unlock(&sp->so_lock);
 	}
 }
 
-static int reclaimer(void *ptr)
+static void nfs4_state_start_reclaim_reboot(struct nfs_client *clp)
+{
+	/* Mark all delegations for reclaim */
+	nfs_delegation_mark_reclaim(clp);
+	nfs4_state_mark_reclaim_helper(clp, nfs4_state_mark_reclaim_reboot);
+}
+
+static void nfs4_state_end_reclaim_reboot(struct nfs_client *clp)
 {
-	struct nfs_client *clp = ptr;
 	struct nfs4_state_owner *sp;
 	struct rb_node *pos;
-	struct nfs4_state_recovery_ops *ops;
-	struct rpc_cred *cred;
+	struct nfs4_state *state;
+
+	if (!test_and_clear_bit(NFS4CLNT_RECLAIM_REBOOT, &clp->cl_state))
+		return;
+
+	for (pos = rb_first(&clp->cl_state_owners); pos != NULL; pos = rb_next(pos)) {
+		sp = rb_entry(pos, struct nfs4_state_owner, so_client_node);
+		spin_lock(&sp->so_lock);
+		list_for_each_entry(state, &sp->so_states, open_states) {
+			if (!test_and_clear_bit(NFS_STATE_RECLAIM_REBOOT, &state->flags))
+				continue;
+			nfs4_state_mark_reclaim_nograce(clp, state);
+		}
+		spin_unlock(&sp->so_lock);
+	}
+
+	nfs_delegation_reap_unclaimed(clp);
+}
+
+static void nfs_delegation_clear_all(struct nfs_client *clp)
+{
+	nfs_delegation_mark_reclaim(clp);
+	nfs_delegation_reap_unclaimed(clp);
+}
+
+static void nfs4_state_start_reclaim_nograce(struct nfs_client *clp)
+{
+	nfs_delegation_clear_all(clp);
+	nfs4_state_mark_reclaim_helper(clp, nfs4_state_mark_reclaim_nograce);
+}
+
+static void nfs4_state_end_reclaim_nograce(struct nfs_client *clp)
+{
+	clear_bit(NFS4CLNT_RECLAIM_NOGRACE, &clp->cl_state);
+}
+
+static void nfs4_recovery_handle_error(struct nfs_client *clp, int error)
+{
+	switch (error) {
+		case -NFS4ERR_CB_PATH_DOWN:
+			nfs_handle_cb_pathdown(clp);
+			break;
+		case -NFS4ERR_STALE_CLIENTID:
+		case -NFS4ERR_LEASE_MOVED:
+			set_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state);
+			nfs4_state_start_reclaim_reboot(clp);
+			break;
+		case -NFS4ERR_EXPIRED:
+			set_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state);
+			nfs4_state_start_reclaim_nograce(clp);
+	}
+}
+
+static int nfs4_do_reclaim(struct nfs_client *clp, const struct nfs4_state_recovery_ops *ops)
+{
+	struct rb_node *pos;
 	int status = 0;
 
-	allow_signal(SIGKILL);
+restart:
+	spin_lock(&clp->cl_lock);
+	for (pos = rb_first(&clp->cl_state_owners); pos != NULL; pos = rb_next(pos)) {
+		struct nfs4_state_owner *sp = rb_entry(pos, struct nfs4_state_owner, so_client_node);
+		if (!test_and_clear_bit(ops->owner_flag_bit, &sp->so_flags))
+			continue;
+		atomic_inc(&sp->so_count);
+		spin_unlock(&clp->cl_lock);
+		status = nfs4_reclaim_open_state(sp, ops);
+		if (status < 0) {
+			set_bit(ops->owner_flag_bit, &sp->so_flags);
+			nfs4_put_state_owner(sp);
+			nfs4_recovery_handle_error(clp, status);
+			return status;
+		}
+		nfs4_put_state_owner(sp);
+		goto restart;
+	}
+	spin_unlock(&clp->cl_lock);
+	return status;
+}
 
-	/* Ensure exclusive access to NFSv4 state */
-	down_write(&clp->cl_sem);
-	/* Are there any NFS mounts out there? */
-	if (list_empty(&clp->cl_superblocks))
-		goto out;
-restart_loop:
-	ops = &nfs4_network_partition_recovery_ops;
-	/* Are there any open files on this volume? */
+static int nfs4_check_lease(struct nfs_client *clp)
+{
+	struct rpc_cred *cred;
+	int status = -NFS4ERR_EXPIRED;
+
+	/* Is the client already known to have an expired lease? */
+	if (test_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state))
+		return 0;
 	cred = nfs4_get_renew_cred(clp);
-	if (cred != NULL) {
-		/* Yes there are: try to renew the old lease */
-		status = nfs4_proc_renew(clp, cred);
-		put_rpccred(cred);
-		switch (status) {
-			case 0:
-			case -NFS4ERR_CB_PATH_DOWN:
-				goto out;
-			case -NFS4ERR_STALE_CLIENTID:
-			case -NFS4ERR_LEASE_MOVED:
-				ops = &nfs4_reboot_recovery_ops;
-		}
-	} else {
-		/* "reboot" to ensure we clear all state on the server */
-		clp->cl_boot_time = CURRENT_TIME;
+	if (cred == NULL) {
+		cred = nfs4_get_setclientid_cred(clp);
+		if (cred == NULL)
+			goto out;
 	}
-	/* We're going to have to re-establish a clientid */
-	nfs4_state_mark_reclaim(clp);
-	status = -ENOENT;
+	status = nfs4_proc_renew(clp, cred);
+	put_rpccred(cred);
+out:
+	nfs4_recovery_handle_error(clp, status);
+	return status;
+}
+
+static int nfs4_reclaim_lease(struct nfs_client *clp)
+{
+	struct rpc_cred *cred;
+	int status = -ENOENT;
+
 	cred = nfs4_get_setclientid_cred(clp);
 	if (cred != NULL) {
 		status = nfs4_init_client(clp, cred);
@@ -974,42 +1099,90 @@ static int reclaimer(void *ptr)
 		/* Handle case where the user hasn't set up machine creds */
 		if (status == -EACCES && cred == clp->cl_machine_cred) {
 			nfs4_clear_machine_cred(clp);
-			goto restart_loop;
+			status = -EAGAIN;
 		}
 	}
-	if (status)
-		goto out_error;
-	/* Mark all delegations for reclaim */
-	nfs_delegation_mark_reclaim(clp);
-	/* Note: list is protected by exclusive lock on cl->cl_sem */
-	for (pos = rb_first(&clp->cl_state_owners); pos != NULL; pos = rb_next(pos)) {
-		sp = rb_entry(pos, struct nfs4_state_owner, so_client_node);
-		status = nfs4_reclaim_open_state(ops, sp);
-		if (status < 0) {
-			if (status == -NFS4ERR_NO_GRACE) {
-				ops = &nfs4_network_partition_recovery_ops;
-				status = nfs4_reclaim_open_state(ops, sp);
+	return status;
+}
+
+static void nfs4_state_manager(struct nfs_client *clp)
+{
+	int status = 0;
+
+	/* Ensure exclusive access to NFSv4 state */
+	for(;;) {
+		if (test_and_clear_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state)) {
+			/* We're going to have to re-establish a clientid */
+			status = nfs4_reclaim_lease(clp);
+			if (status) {
+				set_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state);
+				if (status == -EAGAIN)
+					continue;
+				goto out_error;
 			}
+			clear_bit(NFS4CLNT_CHECK_LEASE, &clp->cl_state);
+		}
+
+		if (test_and_clear_bit(NFS4CLNT_CHECK_LEASE, &clp->cl_state)) {
+			status = nfs4_check_lease(clp);
+			if (status != 0)
+				continue;
+		}
+
+		/* First recover reboot state... */
+		if (test_and_clear_bit(NFS4CLNT_RECLAIM_REBOOT, &clp->cl_state)) {
+			status = nfs4_do_reclaim(clp, &nfs4_reboot_recovery_ops);
 			if (status == -NFS4ERR_STALE_CLIENTID)
-				goto restart_loop;
-			if (status == -NFS4ERR_EXPIRED)
-				goto restart_loop;
+				continue;
+			nfs4_state_end_reclaim_reboot(clp);
+			continue;
+		}
+
+		/* Now recover expired state... */
+		if (test_and_clear_bit(NFS4CLNT_RECLAIM_NOGRACE, &clp->cl_state)) {
+			status = nfs4_do_reclaim(clp, &nfs4_nograce_recovery_ops);
+			if (status < 0) {
+				set_bit(NFS4CLNT_RECLAIM_NOGRACE, &clp->cl_state);
+				if (status == -NFS4ERR_STALE_CLIENTID)
+					continue;
+				if (status == -NFS4ERR_EXPIRED)
+					continue;
+				goto out_error;
+			} else
+				nfs4_state_end_reclaim_nograce(clp);
+			continue;
 		}
+
+		if (test_and_clear_bit(NFS4CLNT_DELEGRETURN, &clp->cl_state)) {
+			nfs_client_return_marked_delegations(clp);
+			continue;
+		}
+
+		nfs4_clear_state_manager_bit(clp);
+		/* Did we race with an attempt to give us more work? */
+		if (clp->cl_state == 0)
+			break;
+		if (test_and_set_bit(NFS4CLNT_MANAGER_RUNNING, &clp->cl_state) != 0)
+			break;
 	}
-	nfs_delegation_reap_unclaimed(clp);
-out:
-	up_write(&clp->cl_sem);
-	if (status == -NFS4ERR_CB_PATH_DOWN)
-		nfs_handle_cb_pathdown(clp);
-	nfs4_clear_recover_bit(clp);
+	return;
+out_error:
+	printk(KERN_WARNING "Error: state manager failed on NFSv4 server %s"
+			" with error %d\n", clp->cl_hostname, -status);
+	if (test_bit(NFS4CLNT_RECLAIM_REBOOT, &clp->cl_state))
+		nfs4_state_end_reclaim_reboot(clp);
+	nfs4_clear_state_manager_bit(clp);
+}
+
+static int nfs4_run_state_manager(void *ptr)
+{
+	struct nfs_client *clp = ptr;
+
+	allow_signal(SIGKILL);
+	nfs4_state_manager(clp);
 	nfs_put_client(clp);
 	module_put_and_exit(0);
 	return 0;
-out_error:
-	printk(KERN_WARNING "Error: state recovery failed on NFSv4 server %s"
-			" with error %d\n", clp->cl_hostname, -status);
-	set_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state);
-	goto out;
 }
 
 /*

fs/nfs/nfsroot.c

@@ -86,6 +86,8 @@
 #include <net/ipconfig.h>
 #include <linux/parser.h>
 
+#include "internal.h"
+
 /* Define this to allow debugging output */
 #undef NFSROOT_DEBUG
 #define NFSDBG_FACILITY NFSDBG_ROOT
@@ -100,7 +102,7 @@ static char nfs_root_name[256] __initdata = "";
 static __be32 servaddr __initdata = 0;
 
 /* Name of directory to mount */
-static char nfs_path[NFS_MAXPATHLEN] __initdata = { 0, };
+static char nfs_export_path[NFS_MAXPATHLEN] __initdata = { 0, };
 
 /* NFS-related data */
 static struct nfs_mount_data nfs_data __initdata = { 0, };/* NFS mount info */
@@ -312,7 +314,7 @@ static int __init root_nfs_name(char *name)
 		printk(KERN_ERR "Root-NFS: Pathname for remote directory too long.\n");
 		return -1;
 	}
-	sprintf(nfs_path, buf, cp);
+	sprintf(nfs_export_path, buf, cp);
 
 	return 1;
 }
@@ -340,7 +342,7 @@ static int __init root_nfs_addr(void)
 static void __init root_nfs_print(void)
 {
 	printk(KERN_NOTICE "Root-NFS: Mounting %s on server %s as root\n",
-		nfs_path, nfs_data.hostname);
+		nfs_export_path, nfs_data.hostname);
 	printk(KERN_NOTICE "Root-NFS:     rsize = %d, wsize = %d, timeo = %d, retrans = %d\n",
 		nfs_data.rsize, nfs_data.wsize, nfs_data.timeo, nfs_data.retrans);
 	printk(KERN_NOTICE "Root-NFS:     acreg (min,max) = (%d,%d), acdir (min,max) = (%d,%d)\n",
@@ -485,18 +487,23 @@ static int __init root_nfs_get_handle(void)
 {
 	struct nfs_fh fh;
 	struct sockaddr_in sin;
+	struct nfs_mount_request request = {
+		.sap		= (struct sockaddr *)&sin,
+		.salen		= sizeof(sin),
+		.dirpath	= nfs_export_path,
+		.version	= (nfs_data.flags & NFS_MOUNT_VER3) ?
+					NFS_MNT3_VERSION : NFS_MNT_VERSION,
+		.protocol	= (nfs_data.flags & NFS_MOUNT_TCP) ?
+					XPRT_TRANSPORT_TCP : XPRT_TRANSPORT_UDP,
+		.fh		= &fh,
+	};
 	int status;
-	int protocol = (nfs_data.flags & NFS_MOUNT_TCP) ?
-					XPRT_TRANSPORT_TCP : XPRT_TRANSPORT_UDP;
-	int version = (nfs_data.flags & NFS_MOUNT_VER3) ?
-					NFS_MNT3_VERSION : NFS_MNT_VERSION;
 
 	set_sockaddr(&sin, servaddr, htons(mount_port));
-	status = nfs_mount((struct sockaddr *) &sin, sizeof(sin), NULL,
-			   nfs_path, version, protocol, &fh);
+	status = nfs_mount(&request);
 	if (status < 0)
 		printk(KERN_ERR "Root-NFS: Server returned error %d "
-				"while mounting %s\n", status, nfs_path);
+				"while mounting %s\n", status, nfs_export_path);
 	else {
 		nfs_data.root.size = fh.size;
 		memcpy(nfs_data.root.data, fh.data, fh.size);

fs/nfs/read.c

@@ -533,12 +533,6 @@ readpage_async_filler(void *data, struct page *page)
 	unsigned int len;
 	int error;
 
-	error = nfs_wb_page(inode, page);
-	if (error)
-		goto out_unlock;
-	if (PageUptodate(page))
-		goto out_unlock;
-
 	len = nfs_page_length(page);
 	if (len == 0)
 		return nfs_return_empty_page(page);

fs/nfs_common/nfsacl.c

@@ -29,8 +29,8 @@
 
 MODULE_LICENSE("GPL");
 
-EXPORT_SYMBOL(nfsacl_encode);
-EXPORT_SYMBOL(nfsacl_decode);
+EXPORT_SYMBOL_GPL(nfsacl_encode);
+EXPORT_SYMBOL_GPL(nfsacl_decode);
 
 struct nfsacl_encode_desc {
 	struct xdr_array2_desc desc;

include/linux/lockd/bind.h

@@ -41,6 +41,7 @@ struct nlmclnt_initdata {
 	size_t			addrlen;
 	unsigned short		protocol;
 	u32			nfs_version;
+	int			noresvport;
 };
 
 /*

include/linux/nfsd/state.h

@@ -124,6 +124,8 @@ struct nfs4_client {
 	nfs4_verifier		cl_verifier; 	/* generated by client */
 	time_t                  cl_time;        /* time of last lease renewal */
 	__be32			cl_addr; 	/* client ipaddress */
+	u32			cl_flavor;	/* setclientid pseudoflavor */
+	char			*cl_principal;	/* setclientid principal name */
 	struct svc_cred		cl_cred; 	/* setclientid principal */
 	clientid_t		cl_clientid;	/* generated by server */
 	nfs4_verifier		cl_confirm;	/* generated by server */