26 files changed, 1632 insertions, 987 deletions

diff --git a/fs/nfs/Makefile b/fs/nfs/Makefile
index 8b3bb715d177..ec61fd56a1a9 100644
--- a/fs/nfs/Makefile
+++ b/fs/nfs/Makefile
@@ -13,4 +13,5 @@ nfs-$(CONFIG_NFS_V4)	+= nfs4proc.o nfs4xdr.o nfs4state.o nfs4renewd.o \
 			   delegation.o idmap.o \
 			   callback.o callback_xdr.o callback_proc.o
 nfs-$(CONFIG_NFS_DIRECTIO) += direct.o
+nfs-$(CONFIG_SYSCTL) += sysctl.o
 nfs-objs		:= $(nfs-y)
diff --git a/fs/nfs/callback.c b/fs/nfs/callback.c
index f2ca782aba33..fcd97406a778 100644
--- a/fs/nfs/callback.c
+++ b/fs/nfs/callback.c
@@ -14,6 +14,9 @@
 #include <linux/sunrpc/svc.h>
 #include <linux/sunrpc/svcsock.h>
 #include <linux/nfs_fs.h>
+
+#include <net/inet_sock.h>
+
 #include "nfs4_fs.h"
 #include "callback.h"
 
@@ -31,6 +34,7 @@ static struct nfs_callback_data nfs_callback_info;
 static DECLARE_MUTEX(nfs_callback_sema);
 static struct svc_program nfs4_callback_program;
 
+unsigned int nfs_callback_set_tcpport;
 unsigned short nfs_callback_tcpport;
 
 /*
@@ -95,7 +99,7 @@ int nfs_callback_up(void)
 	if (!serv)
 		goto out_err;
 	/* FIXME: We don't want to register this socket with the portmapper */
-	ret = svc_makesock(serv, IPPROTO_TCP, 0);
+	ret = svc_makesock(serv, IPPROTO_TCP, nfs_callback_set_tcpport);
 	if (ret < 0)
 		goto out_destroy;
 	if (!list_empty(&serv->sv_permsocks)) {
diff --git a/fs/nfs/callback.h b/fs/nfs/callback.h
index a0db2d4f9415..b252e7fe53a5 100644
--- a/fs/nfs/callback.h
+++ b/fs/nfs/callback.h
@@ -65,6 +65,7 @@ extern unsigned nfs4_callback_recall(struct cb_recallargs *args, void *dummy);
 extern int nfs_callback_up(void);
 extern int nfs_callback_down(void);
 
+extern unsigned int nfs_callback_set_tcpport;
 extern unsigned short nfs_callback_tcpport;
 
 #endif /* __LINUX_FS_NFS_CALLBACK_H */
diff --git a/fs/nfs/callback_proc.c b/fs/nfs/callback_proc.c
index 65f1e19e4d19..462cfceb50c5 100644
--- a/fs/nfs/callback_proc.c
+++ b/fs/nfs/callback_proc.c
@@ -35,7 +35,9 @@ unsigned nfs4_callback_getattr(struct cb_getattrargs *args, struct cb_getattrres
 	if (delegation == NULL || (delegation->type & FMODE_WRITE) == 0)
 		goto out_iput;
 	res->size = i_size_read(inode);
-	res->change_attr = NFS_CHANGE_ATTR(inode);
+	res->change_attr = delegation->change_attr;
+	if (nfsi->npages != 0)
+		res->change_attr++;
 	res->ctime = inode->i_ctime;
 	res->mtime = inode->i_mtime;
 	res->bitmap[0] = (FATTR4_WORD0_CHANGE|FATTR4_WORD0_SIZE) &
diff --git a/fs/nfs/delegation.c b/fs/nfs/delegation.c
index 618a327027b3..c6f07c1c71e6 100644
--- a/fs/nfs/delegation.c
+++ b/fs/nfs/delegation.c
@@ -8,6 +8,7 @@
  */
 #include <linux/config.h>
 #include <linux/completion.h>
+#include <linux/kthread.h>
 #include <linux/module.h>
 #include <linux/sched.h>
 #include <linux/spinlock.h>
@@ -130,6 +131,7 @@ int nfs_inode_set_delegation(struct inode *inode, struct rpc_cred *cred, struct
 			sizeof(delegation->stateid.data));
 	delegation->type = res->delegation_type;
 	delegation->maxsize = res->maxsize;
+	delegation->change_attr = nfsi->change_attr;
 	delegation->cred = get_rpccred(cred);
 	delegation->inode = inode;
 
@@ -157,8 +159,6 @@ static int nfs_do_return_delegation(struct inode *inode, struct nfs_delegation *
 {
 	int res = 0;
 
-	__nfs_revalidate_inode(NFS_SERVER(inode), inode);
-
 	res = nfs4_proc_delegreturn(inode, delegation->cred, &delegation->stateid);
 	nfs_free_delegation(delegation);
 	return res;
@@ -231,6 +231,49 @@ restart:
 	spin_unlock(&clp->cl_lock);
 }
 
+int nfs_do_expire_all_delegations(void *ptr)
+{
+	struct nfs4_client *clp = ptr;
+	struct nfs_delegation *delegation;
+	struct inode *inode;
+
+	allow_signal(SIGKILL);
+restart:
+	spin_lock(&clp->cl_lock);
+	if (test_bit(NFS4CLNT_STATE_RECOVER, &clp->cl_state) != 0)
+		goto out;
+	if (test_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state) == 0)
+		goto out;
+	list_for_each_entry(delegation, &clp->cl_delegations, super_list) {
+		inode = igrab(delegation->inode);
+		if (inode == NULL)
+			continue;
+		spin_unlock(&clp->cl_lock);
+		nfs_inode_return_delegation(inode);
+		iput(inode);
+		goto restart;
+	}
+out:
+	spin_unlock(&clp->cl_lock);
+	nfs4_put_client(clp);
+	module_put_and_exit(0);
+}
+
+void nfs_expire_all_delegations(struct nfs4_client *clp)
+{
+	struct task_struct *task;
+
+	__module_get(THIS_MODULE);
+	atomic_inc(&clp->cl_count);
+	task = kthread_run(nfs_do_expire_all_delegations, clp,
+			"%u.%u.%u.%u-delegreturn",
+			NIPQUAD(clp->cl_addr));
+	if (!IS_ERR(task))
+		return;
+	nfs4_put_client(clp);
+	module_put(THIS_MODULE);
+}
+
 /*
  * Return all delegations following an NFS4ERR_CB_PATH_DOWN error.
  */
diff --git a/fs/nfs/delegation.h b/fs/nfs/delegation.h
index 2fcc30de924b..7a0b2bfce771 100644
--- a/fs/nfs/delegation.h
+++ b/fs/nfs/delegation.h
@@ -21,6 +21,7 @@ struct nfs_delegation {
 #define NFS_DELEGATION_NEED_RECLAIM 1
 	long flags;
 	loff_t maxsize;
+	__u64 change_attr;
 };
 
 int nfs_inode_set_delegation(struct inode *inode, struct rpc_cred *cred, struct nfs_openres *res);
@@ -30,6 +31,7 @@ int nfs_async_inode_return_delegation(struct inode *inode, const nfs4_stateid *s
 
 struct inode *nfs_delegation_find_inode(struct nfs4_client *clp, const struct nfs_fh *fhandle);
 void nfs_return_all_delegations(struct super_block *sb);
+void nfs_expire_all_delegations(struct nfs4_client *clp);
 void nfs_handle_cb_pathdown(struct nfs4_client *clp);
 
 void nfs_delegation_mark_reclaim(struct nfs4_client *clp);
diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index c0d1a214572c..e9255198f767 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -1550,8 +1550,10 @@ go_ahead:
 	}
 	nfs_inode_return_delegation(old_inode);
 
-	if (new_inode)
+	if (new_inode != NULL) {
+		nfs_inode_return_delegation(new_inode);
 		d_delete(new_dentry);
+	}
 
 	nfs_begin_data_update(old_dir);
 	nfs_begin_data_update(new_dir);
diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c
index b497c71384e8..10ae377e68ff 100644
--- a/fs/nfs/direct.c
+++ b/fs/nfs/direct.c
@@ -122,9 +122,10 @@ nfs_free_user_pages(struct page **pages, int npages, int do_dirty)
 {
 	int i;
 	for (i = 0; i < npages; i++) {
-		if (do_dirty)
-			set_page_dirty_lock(pages[i]);
-		page_cache_release(pages[i]);
+		struct page *page = pages[i];
+		if (do_dirty && !PageCompound(page))
+			set_page_dirty_lock(page);
+		page_cache_release(page);
 	}
 	kfree(pages);
 }
@@ -154,6 +155,7 @@ static struct nfs_direct_req *nfs_direct_read_alloc(size_t nbytes, unsigned int
 	struct list_head *list;
 	struct nfs_direct_req *dreq;
 	unsigned int reads = 0;
+	unsigned int rpages = (rsize + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
 
 	dreq = kmem_cache_alloc(nfs_direct_cachep, SLAB_KERNEL);
 	if (!dreq)
@@ -167,7 +169,7 @@ static struct nfs_direct_req *nfs_direct_read_alloc(size_t nbytes, unsigned int
 
 	list = &dreq->list;
 	for(;;) {
-		struct nfs_read_data *data = nfs_readdata_alloc();
+		struct nfs_read_data *data = nfs_readdata_alloc(rpages);
 
 		if (unlikely(!data)) {
 			while (!list_empty(list)) {
@@ -268,8 +270,6 @@ static void nfs_direct_read_schedule(struct nfs_direct_req *dreq,
 		NFS_PROTO(inode)->read_setup(data);
 
 		data->task.tk_cookie = (unsigned long) inode;
-		data->task.tk_calldata = data;
-		data->task.tk_release = nfs_readdata_release;
 		data->complete = nfs_direct_read_result;
 
 		lock_kernel();
@@ -433,7 +433,7 @@ static ssize_t nfs_direct_write_seg(struct inode *inode,
 	struct nfs_writeverf first_verf;
 	struct nfs_write_data *wdata;
 
-	wdata = nfs_writedata_alloc();
+	wdata = nfs_writedata_alloc(NFS_SERVER(inode)->wpages);
 	if (!wdata)
 		return -ENOMEM;
 
@@ -662,10 +662,10 @@ nfs_file_direct_read(struct kiocb *iocb, char __user *buf, size_t count, loff_t
 		.iov_len = count,
 	};
 
-	dprintk("nfs: direct read(%s/%s, %lu@%lu)\n",
+	dprintk("nfs: direct read(%s/%s, %lu@%Ld)\n",
 		file->f_dentry->d_parent->d_name.name,
 		file->f_dentry->d_name.name,
-		(unsigned long) count, (unsigned long) pos);
+		(unsigned long) count, (long long) pos);
 
 	if (!is_sync_kiocb(iocb))
 		goto out;
@@ -678,15 +678,9 @@ nfs_file_direct_read(struct kiocb *iocb, char __user *buf, size_t count, loff_t
 	if (!count)
 		goto out;
 
-	if (mapping->nrpages) {
-		retval = filemap_fdatawrite(mapping);
-		if (retval == 0)
-			retval = nfs_wb_all(inode);
-		if (retval == 0)
-			retval = filemap_fdatawait(mapping);
-		if (retval)
-			goto out;
-	}
+	retval = nfs_sync_mapping(mapping);
+	if (retval)
+		goto out;
 
 	retval = nfs_direct_read(inode, ctx, &iov, pos, 1);
 	if (retval > 0)
@@ -724,9 +718,7 @@ out:
 ssize_t
 nfs_file_direct_write(struct kiocb *iocb, const char __user *buf, size_t count, loff_t pos)
 {
-	ssize_t retval = -EINVAL;
-	loff_t *ppos = &iocb->ki_pos;
-	unsigned long limit = current->signal->rlim[RLIMIT_FSIZE].rlim_cur;
+	ssize_t retval;
 	struct file *file = iocb->ki_filp;
 	struct nfs_open_context *ctx =
 			(struct nfs_open_context *) file->private_data;
@@ -734,51 +726,42 @@ nfs_file_direct_write(struct kiocb *iocb, const char __user *buf, size_t count,
 	struct inode *inode = mapping->host;
 	struct iovec iov = {
 		.iov_base = (char __user *)buf,
-		.iov_len = count,
 	};
 
-	dfprintk(VFS, "nfs: direct write(%s/%s(%ld), %lu@%lu)\n",
+	dfprintk(VFS, "nfs: direct write(%s/%s, %lu@%Ld)\n",
 		file->f_dentry->d_parent->d_name.name,
-		file->f_dentry->d_name.name, inode->i_ino,
-		(unsigned long) count, (unsigned long) pos);
+		file->f_dentry->d_name.name,
+		(unsigned long) count, (long long) pos);
 
+	retval = -EINVAL;
 	if (!is_sync_kiocb(iocb))
 		goto out;
-	if (count < 0)
-		goto out;
-        if (pos < 0)
+
+	retval = generic_write_checks(file, &pos, &count, 0);
+	if (retval)
 		goto out;
-	retval = -EFAULT;
-	if (!access_ok(VERIFY_READ, iov.iov_base, iov.iov_len))
+
+	retval = -EINVAL;
+	if ((ssize_t) count < 0)
 		goto out;
-	retval = -EFBIG;
-	if (limit != RLIM_INFINITY) {
-		if (pos >= limit) {
-			send_sig(SIGXFSZ, current, 0);
-			goto out;
-		}
-		if (count > limit - (unsigned long) pos)
-			count = limit - (unsigned long) pos;
-	}
 	retval = 0;
 	if (!count)
 		goto out;
+	iov.iov_len = count,
 
-	if (mapping->nrpages) {
-		retval = filemap_fdatawrite(mapping);
-		if (retval == 0)
-			retval = nfs_wb_all(inode);
-		if (retval == 0)
-			retval = filemap_fdatawait(mapping);
-		if (retval)
-			goto out;
-	}
+	retval = -EFAULT;
+	if (!access_ok(VERIFY_READ, iov.iov_base, iov.iov_len))
+		goto out;
+
+	retval = nfs_sync_mapping(mapping);
+	if (retval)
+		goto out;
 
 	retval = nfs_direct_write(inode, ctx, &iov, pos, 1);
 	if (mapping->nrpages)
 		invalidate_inode_pages2(mapping);
 	if (retval > 0)
-		*ppos = pos + retval;
+		iocb->ki_pos = pos + retval;
 
 out:
 	return retval;
diff --git a/fs/nfs/file.c b/fs/nfs/file.c
index 57d3e77d97ee..7a79fbe9f539 100644
--- a/fs/nfs/file.c
+++ b/fs/nfs/file.c
@@ -433,11 +433,7 @@ static int do_unlk(struct file *filp, int cmd, struct file_lock *fl)
 	 * Flush all pending writes before doing anything
 	 * with locks..
 	 */
-	filemap_fdatawrite(filp->f_mapping);
-	down(&inode->i_sem);
-	nfs_wb_all(inode);
-	up(&inode->i_sem);
-	filemap_fdatawait(filp->f_mapping);
+	nfs_sync_mapping(filp->f_mapping);
 
 	/* NOTE: special case
 	 * 	If we're signalled while cleaning up locks on process exit, we
@@ -465,15 +461,8 @@ static int do_setlk(struct file *filp, int cmd, struct file_lock *fl)
 	 * Flush all pending writes before doing anything
 	 * with locks..
 	 */
-	status = filemap_fdatawrite(filp->f_mapping);
-	if (status == 0) {
-		down(&inode->i_sem);
-		status = nfs_wb_all(inode);
-		up(&inode->i_sem);
-		if (status == 0)
-			status = filemap_fdatawait(filp->f_mapping);
-	}
-	if (status < 0)
+	status = nfs_sync_mapping(filp->f_mapping);
+	if (status != 0)
 		goto out;
 
 	lock_kernel();
@@ -497,11 +486,7 @@ static int do_setlk(struct file *filp, int cmd, struct file_lock *fl)
 	 * Make sure we clear the cache whenever we try to get the lock.
 	 * This makes locking act as a cache coherency point.
 	 */
-	filemap_fdatawrite(filp->f_mapping);
-	down(&inode->i_sem);
-	nfs_wb_all(inode);	/* we may have slept */
-	up(&inode->i_sem);
-	filemap_fdatawait(filp->f_mapping);
+	nfs_sync_mapping(filp->f_mapping);
 	nfs_zap_caches(inode);
 out:
 	rpc_clnt_sigunmask(NFS_CLIENT(inode), &oldset);
@@ -524,7 +509,8 @@ static int nfs_lock(struct file *filp, int cmd, struct file_lock *fl)
 		return -EINVAL;
 
 	/* No mandatory locks over NFS */
-	if ((inode->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID)
+	if ((inode->i_mode & (S_ISGID | S_IXGRP)) == S_ISGID &&
+	    fl->fl_type != F_UNLCK)
 		return -ENOLCK;
 
 	if (IS_GETLK(cmd))
diff --git a/fs/nfs/idmap.c b/fs/nfs/idmap.c
index ffb8df91dc34..821edd30333b 100644
--- a/fs/nfs/idmap.c
+++ b/fs/nfs/idmap.c
@@ -54,7 +54,11 @@
 
 #define IDMAP_HASH_SZ          128
 
+/* Default cache timeout is 10 minutes */
+unsigned int nfs_idmap_cache_timeout = 600 * HZ;
+
 struct idmap_hashent {
+	unsigned long ih_expires;
 	__u32 ih_id;
 	int ih_namelen;
 	char ih_name[IDMAP_NAMESZ];
@@ -149,6 +153,8 @@ idmap_lookup_name(struct idmap_hashtable *h, const char *name, size_t len)
 
 	if (he->ih_namelen != len || memcmp(he->ih_name, name, len) != 0)
 		return NULL;
+	if (time_after(jiffies, he->ih_expires))
+		return NULL;
 	return he;
 }
 
@@ -164,6 +170,8 @@ idmap_lookup_id(struct idmap_hashtable *h, __u32 id)
 	struct idmap_hashent *he = idmap_id_hash(h, id);
 	if (he->ih_id != id || he->ih_namelen == 0)
 		return NULL;
+	if (time_after(jiffies, he->ih_expires))
+		return NULL;
 	return he;
 }
 
@@ -192,6 +200,7 @@ idmap_update_entry(struct idmap_hashent *he, const char *name,
 	memcpy(he->ih_name, name, namelen);
 	he->ih_name[namelen] = '\0';
 	he->ih_namelen = namelen;
+	he->ih_expires = jiffies + nfs_idmap_cache_timeout;
 }
 
 /*
diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c
index afd75d0463fd..e7bd0d92600f 100644
--- a/fs/nfs/inode.c
+++ b/fs/nfs/inode.c
@@ -40,6 +40,7 @@
 #include <asm/uaccess.h>
 
 #include "nfs4_fs.h"
+#include "callback.h"
 #include "delegation.h"
 
 #define NFSDBG_FACILITY		NFSDBG_VFS
@@ -221,10 +222,10 @@ nfs_calc_block_size(u64 tsize)
 static inline unsigned long
 nfs_block_size(unsigned long bsize, unsigned char *nrbitsp)
 {
-	if (bsize < 1024)
-		bsize = NFS_DEF_FILE_IO_BUFFER_SIZE;
-	else if (bsize >= NFS_MAX_FILE_IO_BUFFER_SIZE)
-		bsize = NFS_MAX_FILE_IO_BUFFER_SIZE;
+	if (bsize < NFS_MIN_FILE_IO_SIZE)
+		bsize = NFS_DEF_FILE_IO_SIZE;
+	else if (bsize >= NFS_MAX_FILE_IO_SIZE)
+		bsize = NFS_MAX_FILE_IO_SIZE;
 
 	return nfs_block_bits(bsize, nrbitsp);
 }
@@ -307,20 +308,15 @@ nfs_sb_init(struct super_block *sb, rpc_authflavor_t authflavor)
 	max_rpc_payload = nfs_block_size(rpc_max_payload(server->client), NULL);
 	if (server->rsize > max_rpc_payload)
 		server->rsize = max_rpc_payload;
-	if (server->wsize > max_rpc_payload)
-		server->wsize = max_rpc_payload;
-
+	if (server->rsize > NFS_MAX_FILE_IO_SIZE)
+		server->rsize = NFS_MAX_FILE_IO_SIZE;
 	server->rpages = (server->rsize + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
-	if (server->rpages > NFS_READ_MAXIOV) {
-		server->rpages = NFS_READ_MAXIOV;
-		server->rsize = server->rpages << PAGE_CACHE_SHIFT;
-	}
 
+	if (server->wsize > max_rpc_payload)
+		server->wsize = max_rpc_payload;
+	if (server->wsize > NFS_MAX_FILE_IO_SIZE)
+		server->wsize = NFS_MAX_FILE_IO_SIZE;
 	server->wpages = (server->wsize + PAGE_CACHE_SIZE - 1) >> PAGE_CACHE_SHIFT;
-        if (server->wpages > NFS_WRITE_MAXIOV) {
-		server->wpages = NFS_WRITE_MAXIOV;
-                server->wsize = server->wpages << PAGE_CACHE_SHIFT;
-	}
 
 	if (sb->s_blocksize == 0)
 		sb->s_blocksize = nfs_block_bits(server->wsize,
@@ -417,7 +413,6 @@ nfs_create_client(struct nfs_server *server, const struct nfs_mount_data *data)
 
 	clnt->cl_intr     = 1;
 	clnt->cl_softrtry = 1;
-	clnt->cl_chatty   = 1;
 
 	return clnt;
 
@@ -575,11 +570,10 @@ nfs_statfs(struct super_block *sb, struct kstatfs *buf)
 	buf->f_namelen = server->namelen;
  out:
 	unlock_kernel();
-
 	return 0;
 
  out_err:
-	printk(KERN_WARNING "nfs_statfs: statfs error = %d\n", -error);
+	dprintk("%s: statfs error = %d\n", __FUNCTION__, -error);
 	buf->f_bsize = buf->f_blocks = buf->f_bfree = buf->f_bavail = -1;
 	goto out;
 
@@ -640,6 +634,27 @@ static int nfs_show_options(struct seq_file *m, struct vfsmount *mnt)
 	return 0;
 }
 
+/**
+ * nfs_sync_mapping - helper to flush all mmapped dirty data to disk
+ */
+int nfs_sync_mapping(struct address_space *mapping)
+{
+	int ret;
+
+	if (mapping->nrpages == 0)
+		return 0;
+	unmap_mapping_range(mapping, 0, 0, 0);
+	ret = filemap_fdatawrite(mapping);
+	if (ret != 0)
+		goto out;
+	ret = filemap_fdatawait(mapping);
+	if (ret != 0)
+		goto out;
+	ret = nfs_wb_all(mapping->host);
+out:
+	return ret;
+}
+
 /*
  * Invalidate the local caches
  */
@@ -937,6 +952,8 @@ int nfs_getattr(struct vfsmount *mnt, struct dentry *dentry, struct kstat *stat)
 	int need_atime = NFS_I(inode)->cache_validity & NFS_INO_INVALID_ATIME;
 	int err;
 
+	/* Flush out writes to the server in order to update c/mtime */
+	nfs_sync_inode(inode, 0, 0, FLUSH_WAIT|FLUSH_NOCOMMIT);
 	if (__IS_FLG(inode, MS_NOATIME))
 		need_atime = 0;
 	else if (__IS_FLG(inode, MS_NODIRATIME) && S_ISDIR(inode->i_mode))
@@ -1179,11 +1196,8 @@ void nfs_revalidate_mapping(struct inode *inode, struct address_space *mapping)
 	struct nfs_inode *nfsi = NFS_I(inode);
 
 	if (nfsi->cache_validity & NFS_INO_INVALID_DATA) {
-		if (S_ISREG(inode->i_mode)) {
-			if (filemap_fdatawrite(mapping) == 0)
-				filemap_fdatawait(mapping);
-			nfs_wb_all(inode);
-		}
+		if (S_ISREG(inode->i_mode))
+			nfs_sync_mapping(mapping);
 		invalidate_inode_pages2(mapping);
 
 		spin_lock(&inode->i_lock);
@@ -1234,6 +1248,33 @@ void nfs_end_data_update(struct inode *inode)
 	atomic_dec(&nfsi->data_updates);
 }
 
+static void nfs_wcc_update_inode(struct inode *inode, struct nfs_fattr *fattr)
+{
+	struct nfs_inode *nfsi = NFS_I(inode);
+
+	if ((fattr->valid & NFS_ATTR_PRE_CHANGE) != 0
+			&& nfsi->change_attr == fattr->pre_change_attr) {
+		nfsi->change_attr = fattr->change_attr;
+		nfsi->cache_change_attribute = jiffies;
+	}
+
+	/* If we have atomic WCC data, we may update some attributes */
+	if ((fattr->valid & NFS_ATTR_WCC) != 0) {
+		if (timespec_equal(&inode->i_ctime, &fattr->pre_ctime)) {
+			memcpy(&inode->i_ctime, &fattr->ctime, sizeof(inode->i_ctime));
+			nfsi->cache_change_attribute = jiffies;
+		}
+		if (timespec_equal(&inode->i_mtime, &fattr->pre_mtime)) {
+			memcpy(&inode->i_mtime, &fattr->mtime, sizeof(inode->i_mtime));
+			nfsi->cache_change_attribute = jiffies;
+		}
+		if (inode->i_size == fattr->pre_size && nfsi->npages == 0) {
+			inode->i_size = fattr->size;
+			nfsi->cache_change_attribute = jiffies;
+		}
+	}
+}
+
 /**
  * nfs_check_inode_attributes - verify consistency of the inode attribute cache
  * @inode - pointer to inode
@@ -1250,22 +1291,20 @@ static int nfs_check_inode_attributes(struct inode *inode, struct nfs_fattr *fat
 	int data_unstable;
 
 
+	if ((fattr->valid & NFS_ATTR_FATTR) == 0)
+		return 0;
+
 	/* Are we in the process of updating data on the server? */
 	data_unstable = nfs_caches_unstable(inode);
 
-	if (fattr->valid & NFS_ATTR_FATTR_V4) {
-		if ((fattr->valid & NFS_ATTR_PRE_CHANGE) != 0
-				&& nfsi->change_attr == fattr->pre_change_attr)
-			nfsi->change_attr = fattr->change_attr;
-		if (nfsi->change_attr != fattr->change_attr) {
-			nfsi->cache_validity |= NFS_INO_INVALID_ATTR;
-			if (!data_unstable)
-				nfsi->cache_validity |= NFS_INO_REVAL_PAGECACHE;
-		}
-	}
+	/* Do atomic weak cache consistency updates */
+	nfs_wcc_update_inode(inode, fattr);
 
-	if ((fattr->valid & NFS_ATTR_FATTR) == 0) {
-		return 0;
+	if ((fattr->valid & NFS_ATTR_FATTR_V4) != 0 &&
+			nfsi->change_attr != fattr->change_attr) {
+		nfsi->cache_validity |= NFS_INO_INVALID_ATTR;
+		if (!data_unstable)
+			nfsi->cache_validity |= NFS_INO_REVAL_PAGECACHE;
 	}
 
 	/* Has the inode gone and changed behind our back? */
@@ -1277,14 +1316,6 @@ static int nfs_check_inode_attributes(struct inode *inode, struct nfs_fattr *fat
 	cur_size = i_size_read(inode);
  	new_isize = nfs_size_to_loff_t(fattr->size);
 
-	/* If we have atomic WCC data, we may update some attributes */
-	if ((fattr->valid & NFS_ATTR_WCC) != 0) {
-		if (timespec_equal(&inode->i_ctime, &fattr->pre_ctime))
-			memcpy(&inode->i_ctime, &fattr->ctime, sizeof(inode->i_ctime));
-		if (timespec_equal(&inode->i_mtime, &fattr->pre_mtime))
-			memcpy(&inode->i_mtime, &fattr->mtime, sizeof(inode->i_mtime));
-	}
-
 	/* Verify a few of the more important attributes */
 	if (!timespec_equal(&inode->i_mtime, &fattr->mtime)) {
 		nfsi->cache_validity |= NFS_INO_INVALID_ATTR;
@@ -1392,14 +1423,8 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 	if ((fattr->valid & NFS_ATTR_FATTR) == 0)
 		return 0;
 
-	if (nfsi->fileid != fattr->fileid) {
-		printk(KERN_ERR "%s: inode number mismatch\n"
-		       "expected (%s/0x%Lx), got (%s/0x%Lx)\n",
-		       __FUNCTION__,
-		       inode->i_sb->s_id, (long long)nfsi->fileid,
-		       inode->i_sb->s_id, (long long)fattr->fileid);
-		goto out_err;
-	}
+	if (nfsi->fileid != fattr->fileid)
+		goto out_fileid;
 
 	/*
 	 * Make sure the inode's type hasn't changed.
@@ -1418,6 +1443,9 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 	if (data_stable)
 		nfsi->cache_validity &= ~(NFS_INO_INVALID_ATTR|NFS_INO_INVALID_ATIME);
 
+	/* Do atomic weak cache consistency updates */
+	nfs_wcc_update_inode(inode, fattr);
+
 	/* Check if our cached file size is stale */
  	new_isize = nfs_size_to_loff_t(fattr->size);
 	cur_isize = i_size_read(inode);
@@ -1521,6 +1549,13 @@ static int nfs_update_inode(struct inode *inode, struct nfs_fattr *fattr)
 	 */
 	nfs_invalidate_inode(inode);
 	return -ESTALE;
+
+ out_fileid:
+	printk(KERN_ERR "NFS: server %s error: fileid changed\n"
+		"fsid %s: expected fileid 0x%Lx, got 0x%Lx\n",
+		NFS_SERVER(inode)->hostname, inode->i_sb->s_id,
+		(long long)nfsi->fileid, (long long)fattr->fileid);
+	goto out_err;
 }
 
 /*
@@ -1802,25 +1837,10 @@ static int nfs4_fill_super(struct super_block *sb, struct nfs4_mount_data *data,
 		}
 		clnt->cl_intr     = 1;
 		clnt->cl_softrtry = 1;
-		clnt->cl_chatty   = 1;
 		clp->cl_rpcclient = clnt;
-		clp->cl_cred = rpcauth_lookupcred(clnt->cl_auth, 0);
-		if (IS_ERR(clp->cl_cred)) {
-			up_write(&clp->cl_sem);
-			err = PTR_ERR(clp->cl_cred);
-			clp->cl_cred = NULL;
-			goto out_fail;
-		}
 		memcpy(clp->cl_ipaddr, server->ip_addr, sizeof(clp->cl_ipaddr));
 		nfs_idmap_new(clp);
 	}
-	if (list_empty(&clp->cl_superblocks)) {
-		err = nfs4_init_client(clp);
-		if (err != 0) {
-			up_write(&clp->cl_sem);
-			goto out_fail;
-		}
-	}
 	list_add_tail(&server->nfs4_siblings, &clp->cl_superblocks);
 	clnt = rpc_clone_client(clp->cl_rpcclient);
 	if (!IS_ERR(clnt))
@@ -2015,6 +2035,35 @@ static struct file_system_type nfs4_fs_type = {
 	.fs_flags	= FS_ODD_RENAME|FS_REVAL_DOT|FS_BINARY_MOUNTDATA,
 };
 
+static const int nfs_set_port_min = 0;
+static const int nfs_set_port_max = 65535;
+static int param_set_port(const char *val, struct kernel_param *kp)
+{
+	char *endp;
+	int num = simple_strtol(val, &endp, 0);
+	if (endp == val || *endp || num < nfs_set_port_min || num > nfs_set_port_max)
+		return -EINVAL;
+	*((int *)kp->arg) = num;
+	return 0;
+}
+
+module_param_call(callback_tcpport, param_set_port, param_get_int,
+		 &nfs_callback_set_tcpport, 0644);
+
+static int param_set_idmap_timeout(const char *val, struct kernel_param *kp)
+{
+	char *endp;
+	int num = simple_strtol(val, &endp, 0);
+	int jif = num * HZ;
+	if (endp == val || *endp || num < 0 || jif < num)
+		return -EINVAL;
+	*((int *)kp->arg) = jif;
+	return 0;
+}
+
+module_param_call(idmap_cache_timeout, param_set_idmap_timeout, param_get_int,
+		 &nfs_idmap_cache_timeout, 0644);
+
 #define nfs4_init_once(nfsi) \
 	do { \
 		INIT_LIST_HEAD(&(nfsi)->open_states); \
@@ -2022,8 +2071,25 @@ static struct file_system_type nfs4_fs_type = {
 		nfsi->delegation_state = 0; \
 		init_rwsem(&nfsi->rwsem); \
 	} while(0)
-#define register_nfs4fs() register_filesystem(&nfs4_fs_type)
-#define unregister_nfs4fs() unregister_filesystem(&nfs4_fs_type)
+
+static inline int register_nfs4fs(void)
+{
+	int ret;
+
+	ret = nfs_register_sysctl();
+	if (ret != 0)
+		return ret;
+	ret = register_filesystem(&nfs4_fs_type);
+	if (ret != 0)
+		nfs_unregister_sysctl();
+	return ret;
+}
+
+static inline void unregister_nfs4fs(void)
+{
+	unregister_filesystem(&nfs4_fs_type);
+	nfs_unregister_sysctl();
+}
 #else
 #define nfs4_init_once(nfsi) \
 	do { } while (0)
@@ -2148,11 +2214,11 @@ out:
 #ifdef CONFIG_PROC_FS
 	rpc_proc_unregister("nfs");
 #endif
-	nfs_destroy_writepagecache();
 #ifdef CONFIG_NFS_DIRECTIO
-out0:
 	nfs_destroy_directcache();
+out0:
 #endif
+	nfs_destroy_writepagecache();
 out1:
 	nfs_destroy_readpagecache();
 out2:
diff --git a/fs/nfs/mount_clnt.c b/fs/nfs/mount_clnt.c
index 0e82617f2de0..db99b8f678f8 100644
--- a/fs/nfs/mount_clnt.c
+++ b/fs/nfs/mount_clnt.c
@@ -82,7 +82,6 @@ mnt_create(char *hostname, struct sockaddr_in *srvaddr, int version,
 				RPC_AUTH_UNIX);
 	if (!IS_ERR(clnt)) {
 		clnt->cl_softrtry = 1;
-		clnt->cl_chatty   = 1;
 		clnt->cl_oneshot  = 1;
 		clnt->cl_intr = 1;
 	}
diff --git a/fs/nfs/nfs2xdr.c b/fs/nfs/nfs2xdr.c
index 59049e864ca7..7fc0560c89c9 100644
--- a/fs/nfs/nfs2xdr.c
+++ b/fs/nfs/nfs2xdr.c
@@ -146,23 +146,23 @@ xdr_decode_fattr(u32 *p, struct nfs_fattr *fattr)
 	return p;
 }
 
-#define SATTR(p, attr, flag, field) \
-        *p++ = (attr->ia_valid & flag) ? htonl(attr->field) : ~(u32) 0
 static inline u32 *
 xdr_encode_sattr(u32 *p, struct iattr *attr)
 {
-	SATTR(p, attr, ATTR_MODE, ia_mode);
-	SATTR(p, attr, ATTR_UID, ia_uid);
-	SATTR(p, attr, ATTR_GID, ia_gid);
-	SATTR(p, attr, ATTR_SIZE, ia_size);
+	const u32 not_set = __constant_htonl(0xFFFFFFFF);
+
+	*p++ = (attr->ia_valid & ATTR_MODE) ? htonl(attr->ia_mode) : not_set;
+	*p++ = (attr->ia_valid & ATTR_UID) ? htonl(attr->ia_uid) : not_set;
+	*p++ = (attr->ia_valid & ATTR_GID) ? htonl(attr->ia_gid) : not_set;
+	*p++ = (attr->ia_valid & ATTR_SIZE) ? htonl(attr->ia_size) : not_set;
 
 	if (attr->ia_valid & ATTR_ATIME_SET) {
 		p = xdr_encode_time(p, &attr->ia_atime);
 	} else if (attr->ia_valid & ATTR_ATIME) {
 		p = xdr_encode_current_server_time(p, &attr->ia_atime);
 	} else {
-		*p++ = ~(u32) 0;
-		*p++ = ~(u32) 0;
+		*p++ = not_set;
+		*p++ = not_set;
 	}
 
 	if (attr->ia_valid & ATTR_MTIME_SET) {
@@ -170,12 +170,11 @@ xdr_encode_sattr(u32 *p, struct iattr *attr)
 	} else if (attr->ia_valid & ATTR_MTIME) {
 		p = xdr_encode_current_server_time(p, &attr->ia_mtime);
 	} else {
-		*p++ = ~(u32) 0;	
-		*p++ = ~(u32) 0;
+		*p++ = not_set;	
+		*p++ = not_set;
 	}
   	return p;
 }
-#undef SATTR
 
 /*
  * NFS encode functions
diff --git a/fs/nfs/nfs3proc.c b/fs/nfs/nfs3proc.c
index 92c870d19ccd..ed67567f0556 100644
--- a/fs/nfs/nfs3proc.c
+++ b/fs/nfs/nfs3proc.c
@@ -68,27 +68,39 @@ nfs3_async_handle_jukebox(struct rpc_task *task)
 	return 1;
 }
 
-/*
- * Bare-bones access to getattr: this is for nfs_read_super.
- */
 static int
-nfs3_proc_get_root(struct nfs_server *server, struct nfs_fh *fhandle,
-		   struct nfs_fsinfo *info)
+do_proc_get_root(struct rpc_clnt *client, struct nfs_fh *fhandle,
+		 struct nfs_fsinfo *info)
 {
 	int	status;
 
 	dprintk("%s: call  fsinfo\n", __FUNCTION__);
 	nfs_fattr_init(info->fattr);
-	status = rpc_call(server->client_sys, NFS3PROC_FSINFO, fhandle, info, 0);
+	status = rpc_call(client, NFS3PROC_FSINFO, fhandle, info, 0);
 	dprintk("%s: reply fsinfo: %d\n", __FUNCTION__, status);
 	if (!(info->fattr->valid & NFS_ATTR_FATTR)) {
-		status = rpc_call(server->client_sys, NFS3PROC_GETATTR, fhandle, info->fattr, 0);
+		status = rpc_call(client, NFS3PROC_GETATTR, fhandle, info->fattr, 0);
 		dprintk("%s: reply getattr: %d\n", __FUNCTION__, status);
 	}
 	return status;
 }
 
 /*
+ * Bare-bones access to getattr: this is for nfs_read_super.
+ */
+static int
+nfs3_proc_get_root(struct nfs_server *server, struct nfs_fh *fhandle,
+		   struct nfs_fsinfo *info)
+{
+	int	status;
+
+	status = do_proc_get_root(server->client, fhandle, info);
+	if (status && server->client_sys != server->client)
+		status = do_proc_get_root(server->client_sys, fhandle, info);
+	return status;
+}
+
+/*
  * One function for each procedure in the NFS protocol.
  */
 static int
@@ -732,19 +744,23 @@ nfs3_proc_pathconf(struct nfs_server *server, struct nfs_fh *fhandle,
 
 extern u32 *nfs3_decode_dirent(u32 *, struct nfs_entry *, int);
 
-static void
-nfs3_read_done(struct rpc_task *task)
+static void nfs3_read_done(struct rpc_task *task, void *calldata)
 {
-	struct nfs_read_data *data = (struct nfs_read_data *) task->tk_calldata;
+	struct nfs_read_data *data = calldata;
 
 	if (nfs3_async_handle_jukebox(task))
 		return;
 	/* Call back common NFS readpage processing */
 	if (task->tk_status >= 0)
 		nfs_refresh_inode(data->inode, &data->fattr);
-	nfs_readpage_result(task);
+	nfs_readpage_result(task, calldata);
 }
 
+static const struct rpc_call_ops nfs3_read_ops = {
+	.rpc_call_done = nfs3_read_done,
+	.rpc_release = nfs_readdata_release,
+};
+
 static void
 nfs3_proc_read_setup(struct nfs_read_data *data)
 {
@@ -762,23 +778,26 @@ nfs3_proc_read_setup(struct nfs_read_data *data)
 	flags = RPC_TASK_ASYNC | (IS_SWAPFILE(inode)? NFS_RPC_SWAPFLAGS : 0);
 
 	/* Finalize the task. */
-	rpc_init_task(task, NFS_CLIENT(inode), nfs3_read_done, flags);
+	rpc_init_task(task, NFS_CLIENT(inode), flags, &nfs3_read_ops, data);
 	rpc_call_setup(task, &msg, 0);
 }
 
-static void
-nfs3_write_done(struct rpc_task *task)
+static void nfs3_write_done(struct rpc_task *task, void *calldata)
 {
-	struct nfs_write_data *data;
+	struct nfs_write_data *data = calldata;
 
 	if (nfs3_async_handle_jukebox(task))
 		return;
-	data = (struct nfs_write_data *)task->tk_calldata;
 	if (task->tk_status >= 0)
 		nfs_post_op_update_inode(data->inode, data->res.fattr);
-	nfs_writeback_done(task);
+	nfs_writeback_done(task, calldata);
 }
 
+static const struct rpc_call_ops nfs3_write_ops = {
+	.rpc_call_done = nfs3_write_done,
+	.rpc_release = nfs_writedata_release,
+};
+
 static void
 nfs3_proc_write_setup(struct nfs_write_data *data, int how)
 {
@@ -806,23 +825,26 @@ nfs3_proc_write_setup(struct nfs_write_data *data, int how)
 	flags = (how & FLUSH_SYNC) ? 0 : RPC_TASK_ASYNC;
 
 	/* Finalize the task. */
-	rpc_init_task(task, NFS_CLIENT(inode), nfs3_write_done, flags);
+	rpc_init_task(task, NFS_CLIENT(inode), flags, &nfs3_write_ops, data);
 	rpc_call_setup(task, &msg, 0);
 }
 
-static void
-nfs3_commit_done(struct rpc_task *task)
+static void nfs3_commit_done(struct rpc_task *task, void *calldata)
 {
-	struct nfs_write_data *data;
+	struct nfs_write_data *data = calldata;
 
 	if (nfs3_async_handle_jukebox(task))
 		return;
-	data = (struct nfs_write_data *)task->tk_calldata;
 	if (task->tk_status >= 0)
 		nfs_post_op_update_inode(data->inode, data->res.fattr);
-	nfs_commit_done(task);
+	nfs_commit_done(task, calldata);
 }
 
+static const struct rpc_call_ops nfs3_commit_ops = {
+	.rpc_call_done = nfs3_commit_done,
+	.rpc_release = nfs_commit_release,
+};
+
 static void
 nfs3_proc_commit_setup(struct nfs_write_data *data, int how)
 {
@@ -840,7 +862,7 @@ nfs3_proc_commit_setup(struct nfs_write_data *data, int how)
 	flags = (how & FLUSH_SYNC) ? 0 : RPC_TASK_ASYNC;
 
 	/* Finalize the task. */
-	rpc_init_task(task, NFS_CLIENT(inode), nfs3_commit_done, flags);
+	rpc_init_task(task, NFS_CLIENT(inode), flags, &nfs3_commit_ops, data);
 	rpc_call_setup(task, &msg, 0);
 }
 
diff --git a/fs/nfs/nfs3xdr.c b/fs/nfs/nfs3xdr.c
index 0498bd36602c..b6c0b5012bce 100644
--- a/fs/nfs/nfs3xdr.c
+++ b/fs/nfs/nfs3xdr.c
@@ -182,7 +182,7 @@ xdr_encode_sattr(u32 *p, struct iattr *attr)
 {
 	if (attr->ia_valid & ATTR_MODE) {
 		*p++ = xdr_one;
-		*p++ = htonl(attr->ia_mode);
+		*p++ = htonl(attr->ia_mode & S_IALLUGO);
 	} else {
 		*p++ = xdr_zero;
 	}
diff --git a/fs/nfs/nfs4_fs.h b/fs/nfs/nfs4_fs.h
index b7f262dcb6e3..0f5e4e7cddec 100644
--- a/fs/nfs/nfs4_fs.h
+++ b/fs/nfs/nfs4_fs.h
@@ -38,7 +38,8 @@ struct idmap;
  ((err) != NFSERR_NOFILEHANDLE))
 
 enum nfs4_client_state {
-	NFS4CLNT_OK  = 0,
+	NFS4CLNT_STATE_RECOVER  = 0,
+	NFS4CLNT_LEASE_EXPIRED,
 };
 
 /*
@@ -67,7 +68,6 @@ struct nfs4_client {
 	atomic_t		cl_count;
 
 	struct rpc_clnt *	cl_rpcclient;
-	struct rpc_cred *	cl_cred;
 
 	struct list_head	cl_superblocks;	/* List of nfs_server structs */
 
@@ -76,7 +76,6 @@ struct nfs4_client {
 	struct work_struct	cl_renewd;
 	struct work_struct	cl_recoverd;
 
-	wait_queue_head_t	cl_waitq;
 	struct rpc_wait_queue	cl_rpcwaitq;
 
 	/* used for the setclientid verifier */
@@ -182,8 +181,9 @@ struct nfs4_state {
 
 	nfs4_stateid stateid;
 
-	unsigned int nreaders;
-	unsigned int nwriters;
+	unsigned int n_rdonly;
+	unsigned int n_wronly;
+	unsigned int n_rdwr;
 	int state;			/* State on the server (R,W, or RW) */
 	atomic_t count;
 };
@@ -210,10 +210,10 @@ 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 nfs4_client *, u32, unsigned short);
-extern int nfs4_proc_setclientid_confirm(struct nfs4_client *);
-extern int nfs4_proc_async_renew(struct nfs4_client *);
-extern int nfs4_proc_renew(struct nfs4_client *);
+extern int nfs4_proc_setclientid(struct nfs4_client *, u32, unsigned short, struct rpc_cred *);
+extern int nfs4_proc_setclientid_confirm(struct nfs4_client *, struct rpc_cred *);
+extern int nfs4_proc_async_renew(struct nfs4_client *, struct rpc_cred *);
+extern int nfs4_proc_renew(struct nfs4_client *, struct rpc_cred *);
 extern int nfs4_do_close(struct inode *inode, struct nfs4_state *state);
 extern struct dentry *nfs4_atomic_open(struct inode *, struct dentry *, struct nameidata *);
 extern int nfs4_open_revalidate(struct inode *, struct dentry *, int, struct nameidata *);
@@ -237,8 +237,8 @@ extern void init_nfsv4_state(struct nfs_server *);
 extern void destroy_nfsv4_state(struct nfs_server *);
 extern struct nfs4_client *nfs4_get_client(struct in_addr *);
 extern void nfs4_put_client(struct nfs4_client *clp);
-extern int nfs4_init_client(struct nfs4_client *clp);
 extern struct nfs4_client *nfs4_find_client(struct in_addr *);
+struct rpc_cred *nfs4_get_renew_cred(struct nfs4_client *clp);
 extern u32 nfs4_alloc_lockowner_id(struct nfs4_client *);
 
 extern struct nfs4_state_owner * nfs4_get_state_owner(struct nfs_server *, struct rpc_cred *);
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index f988a9417b13..984ca3454d04 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -57,11 +57,13 @@
 #define NFS4_POLL_RETRY_MIN	(1*HZ)
 #define NFS4_POLL_RETRY_MAX	(15*HZ)
 
-static int _nfs4_proc_open_confirm(struct rpc_clnt *clnt, const struct nfs_fh *fh, struct nfs4_state_owner *sp, nfs4_stateid *stateid, struct nfs_seqid *seqid);
+struct nfs4_opendata;
+static int _nfs4_proc_open(struct nfs4_opendata *data);
 static int nfs4_do_fsinfo(struct nfs_server *, struct nfs_fh *, struct nfs_fsinfo *);
 static int nfs4_async_handle_error(struct rpc_task *, const struct nfs_server *);
 static int _nfs4_proc_access(struct inode *inode, struct nfs_access_entry *entry);
 static int nfs4_handle_exception(const struct nfs_server *server, int errorcode, struct nfs4_exception *exception);
+static int nfs4_wait_clnt_recover(struct rpc_clnt *clnt, struct nfs4_client *clp);
 extern u32 *nfs4_decode_dirent(u32 *p, struct nfs_entry *entry, int plus);
 extern struct rpc_procinfo nfs4_procedures[];
 
@@ -173,8 +175,7 @@ static void nfs4_setup_readdir(u64 cookie, u32 *verifier, struct dentry *dentry,
 	kunmap_atomic(start, KM_USER0);
 }
 
-static void
-renew_lease(struct nfs_server *server, unsigned long timestamp)
+static void renew_lease(const struct nfs_server *server, unsigned long timestamp)
 {
 	struct nfs4_client *clp = server->nfs4_state;
 	spin_lock(&clp->cl_lock);
@@ -194,21 +195,123 @@ static void update_changeattr(struct inode *inode, struct nfs4_change_info *cinf
 	spin_unlock(&inode->i_lock);
 }
 
+struct nfs4_opendata {
+	atomic_t count;
+	struct nfs_openargs o_arg;
+	struct nfs_openres o_res;
+	struct nfs_open_confirmargs c_arg;
+	struct nfs_open_confirmres c_res;
+	struct nfs_fattr f_attr;
+	struct nfs_fattr dir_attr;
+	struct dentry *dentry;
+	struct dentry *dir;
+	struct nfs4_state_owner *owner;
+	struct iattr attrs;
+	unsigned long timestamp;
+	int rpc_status;
+	int cancelled;
+};
+
+static struct nfs4_opendata *nfs4_opendata_alloc(struct dentry *dentry,
+		struct nfs4_state_owner *sp, int flags,
+		const struct iattr *attrs)
+{
+	struct dentry *parent = dget_parent(dentry);
+	struct inode *dir = parent->d_inode;
+	struct nfs_server *server = NFS_SERVER(dir);
+	struct nfs4_opendata *p;
+
+	p = kzalloc(sizeof(*p), GFP_KERNEL);
+	if (p == NULL)
+		goto err;
+	p->o_arg.seqid = nfs_alloc_seqid(&sp->so_seqid);
+	if (p->o_arg.seqid == NULL)
+		goto err_free;
+	atomic_set(&p->count, 1);
+	p->dentry = dget(dentry);
+	p->dir = parent;
+	p->owner = sp;
+	atomic_inc(&sp->so_count);
+	p->o_arg.fh = NFS_FH(dir);
+	p->o_arg.open_flags = flags,
+	p->o_arg.clientid = server->nfs4_state->cl_clientid;
+	p->o_arg.id = sp->so_id;
+	p->o_arg.name = &dentry->d_name;
+	p->o_arg.server = server;
+	p->o_arg.bitmask = server->attr_bitmask;
+	p->o_arg.claim = NFS4_OPEN_CLAIM_NULL;
+	p->o_res.f_attr = &p->f_attr;
+	p->o_res.dir_attr = &p->dir_attr;
+	p->o_res.server = server;
+	nfs_fattr_init(&p->f_attr);
+	nfs_fattr_init(&p->dir_attr);
+	if (flags & O_EXCL) {
+		u32 *s = (u32 *) p->o_arg.u.verifier.data;
+		s[0] = jiffies;
+		s[1] = current->pid;
+	} else if (flags & O_CREAT) {
+		p->o_arg.u.attrs = &p->attrs;
+		memcpy(&p->attrs, attrs, sizeof(p->attrs));
+	}
+	p->c_arg.fh = &p->o_res.fh;
+	p->c_arg.stateid = &p->o_res.stateid;
+	p->c_arg.seqid = p->o_arg.seqid;
+	return p;
+err_free:
+	kfree(p);
+err:
+	dput(parent);
+	return NULL;
+}
+
+static void nfs4_opendata_free(struct nfs4_opendata *p)
+{
+	if (p != NULL && atomic_dec_and_test(&p->count)) {
+		nfs_free_seqid(p->o_arg.seqid);
+		nfs4_put_state_owner(p->owner);
+		dput(p->dir);
+		dput(p->dentry);
+		kfree(p);
+	}
+}
+
 /* Helper for asynchronous RPC calls */
-static int nfs4_call_async(struct rpc_clnt *clnt, rpc_action tk_begin,
-		rpc_action tk_exit, void *calldata)
+static int nfs4_call_async(struct rpc_clnt *clnt,
+		const struct rpc_call_ops *tk_ops, void *calldata)
 {
 	struct rpc_task *task;
 
-	if (!(task = rpc_new_task(clnt, tk_exit, RPC_TASK_ASYNC)))
+	if (!(task = rpc_new_task(clnt, RPC_TASK_ASYNC, tk_ops, calldata)))
 		return -ENOMEM;
-
-	task->tk_calldata = calldata;
-	task->tk_action = tk_begin;
 	rpc_execute(task);
 	return 0;
 }
 
+static int nfs4_wait_for_completion_rpc_task(struct rpc_task *task)
+{
+	sigset_t oldset;
+	int ret;
+
+	rpc_clnt_sigmask(task->tk_client, &oldset);
+	ret = rpc_wait_for_completion_task(task);
+	rpc_clnt_sigunmask(task->tk_client, &oldset);
+	return ret;
+}
+
+static inline void update_open_stateflags(struct nfs4_state *state, mode_t open_flags)
+{
+	switch (open_flags) {
+		case FMODE_WRITE:
+			state->n_wronly++;
+			break;
+		case FMODE_READ:
+			state->n_rdonly++;
+			break;
+		case FMODE_READ|FMODE_WRITE:
+			state->n_rdwr++;
+	}
+}
+
 static void update_open_stateid(struct nfs4_state *state, nfs4_stateid *stateid, int open_flags)
 {
 	struct inode *inode = state->inode;
@@ -218,41 +321,134 @@ static void update_open_stateid(struct nfs4_state *state, nfs4_stateid *stateid,
 	spin_lock(&state->owner->so_lock);
 	spin_lock(&inode->i_lock);
 	memcpy(&state->stateid, stateid, sizeof(state->stateid));
-	if ((open_flags & FMODE_WRITE))
-		state->nwriters++;
-	if (open_flags & FMODE_READ)
-		state->nreaders++;
+	update_open_stateflags(state, open_flags);
 	nfs4_state_set_mode_locked(state, state->state | open_flags);
 	spin_unlock(&inode->i_lock);
 	spin_unlock(&state->owner->so_lock);
 }
 
+static struct nfs4_state *nfs4_opendata_to_nfs4_state(struct nfs4_opendata *data)
+{
+	struct inode *inode;
+	struct nfs4_state *state = NULL;
+
+	if (!(data->f_attr.valid & NFS_ATTR_FATTR))
+		goto out;
+	inode = nfs_fhget(data->dir->d_sb, &data->o_res.fh, &data->f_attr);
+	if (inode == NULL)
+		goto out;
+	state = nfs4_get_open_state(inode, data->owner);
+	if (state == NULL)
+		goto put_inode;
+	update_open_stateid(state, &data->o_res.stateid, data->o_arg.open_flags);
+put_inode:
+	iput(inode);
+out:
+	return state;
+}
+
+static struct nfs_open_context *nfs4_state_find_open_context(struct nfs4_state *state)
+{
+	struct nfs_inode *nfsi = NFS_I(state->inode);
+	struct nfs_open_context *ctx;
+
+	spin_lock(&state->inode->i_lock);
+	list_for_each_entry(ctx, &nfsi->open_files, list) {
+		if (ctx->state != state)
+			continue;
+		get_nfs_open_context(ctx);
+		spin_unlock(&state->inode->i_lock);
+		return ctx;
+	}
+	spin_unlock(&state->inode->i_lock);
+	return ERR_PTR(-ENOENT);
+}
+
+static int nfs4_open_recover_helper(struct nfs4_opendata *opendata, mode_t openflags, nfs4_stateid *stateid)
+{
+	int ret;
+
+	opendata->o_arg.open_flags = openflags;
+	ret = _nfs4_proc_open(opendata);
+	if (ret != 0)
+		return ret; 
+	memcpy(stateid->data, opendata->o_res.stateid.data,
+			sizeof(stateid->data));
+	return 0;
+}
+
+static int nfs4_open_recover(struct nfs4_opendata *opendata, struct nfs4_state *state)
+{
+	nfs4_stateid stateid;
+	struct nfs4_state *newstate;
+	int mode = 0;
+	int delegation = 0;
+	int ret;
+
+	/* memory barrier prior to reading state->n_* */
+	smp_rmb();
+	if (state->n_rdwr != 0) {
+		ret = nfs4_open_recover_helper(opendata, FMODE_READ|FMODE_WRITE, &stateid);
+		if (ret != 0)
+			return ret;
+		mode |= FMODE_READ|FMODE_WRITE;
+		if (opendata->o_res.delegation_type != 0)
+			delegation = opendata->o_res.delegation_type;
+		smp_rmb();
+	}
+	if (state->n_wronly != 0) {
+		ret = nfs4_open_recover_helper(opendata, FMODE_WRITE, &stateid);
+		if (ret != 0)
+			return ret;
+		mode |= FMODE_WRITE;
+		if (opendata->o_res.delegation_type != 0)
+			delegation = opendata->o_res.delegation_type;
+		smp_rmb();
+	}
+	if (state->n_rdonly != 0) {
+		ret = nfs4_open_recover_helper(opendata, FMODE_READ, &stateid);
+		if (ret != 0)
+			return ret;
+		mode |= FMODE_READ;
+	}
+	clear_bit(NFS_DELEGATED_STATE, &state->flags);
+	if (mode == 0)
+		return 0;
+	if (opendata->o_res.delegation_type == 0)
+		opendata->o_res.delegation_type = delegation;
+	opendata->o_arg.open_flags |= mode;
+	newstate = nfs4_opendata_to_nfs4_state(opendata);
+	if (newstate != NULL) {
+		if (opendata->o_res.delegation_type != 0) {
+			struct nfs_inode *nfsi = NFS_I(newstate->inode);
+			int delegation_flags = 0;
+			if (nfsi->delegation)
+				delegation_flags = nfsi->delegation->flags;
+			if (!(delegation_flags & NFS_DELEGATION_NEED_RECLAIM))
+				nfs_inode_set_delegation(newstate->inode,
+						opendata->owner->so_cred,
+						&opendata->o_res);
+			else
+				nfs_inode_reclaim_delegation(newstate->inode,
+						opendata->owner->so_cred,
+						&opendata->o_res);
+		}
+		nfs4_close_state(newstate, opendata->o_arg.open_flags);
+	}
+	if (newstate != state)
+		return -ESTALE;
+	return 0;
+}
+
 /*
  * OPEN_RECLAIM:
  * 	reclaim state on the server after a reboot.
  */
-static int _nfs4_open_reclaim(struct nfs4_state_owner *sp, struct nfs4_state *state)
+static int _nfs4_do_open_reclaim(struct nfs4_state_owner *sp, struct nfs4_state *state, struct dentry *dentry)
 {
-	struct inode *inode = state->inode;
-	struct nfs_server *server = NFS_SERVER(inode);
-	struct nfs_delegation *delegation = NFS_I(inode)->delegation;
-	struct nfs_openargs o_arg = {
-		.fh = NFS_FH(inode),
-		.id = sp->so_id,
-		.open_flags = state->state,
-		.clientid = server->nfs4_state->cl_clientid,
-		.claim = NFS4_OPEN_CLAIM_PREVIOUS,
-		.bitmask = server->attr_bitmask,
-	};
-	struct nfs_openres o_res = {
-		.server = server,	/* Grrr */
-	};
-	struct rpc_message msg = {
-		.rpc_proc       = &nfs4_procedures[NFSPROC4_CLNT_OPEN_NOATTR],
-		.rpc_argp       = &o_arg,
-		.rpc_resp	= &o_res,
-		.rpc_cred	= sp->so_cred,
-	};
+	struct nfs_delegation *delegation = NFS_I(state->inode)->delegation;
+	struct nfs4_opendata *opendata;
+	int delegation_type = 0;
 	int status;
 
 	if (delegation != NULL) {
@@ -262,38 +458,27 @@ static int _nfs4_open_reclaim(struct nfs4_state_owner *sp, struct nfs4_state *st
 			set_bit(NFS_DELEGATED_STATE, &state->flags);
 			return 0;
 		}
-		o_arg.u.delegation_type = delegation->type;
+		delegation_type = delegation->type;
 	}
-	o_arg.seqid = nfs_alloc_seqid(&sp->so_seqid);
-	if (o_arg.seqid == NULL)
+	opendata = nfs4_opendata_alloc(dentry, sp, 0, NULL);
+	if (opendata == NULL)
 		return -ENOMEM;
-	status = rpc_call_sync(server->client, &msg, RPC_TASK_NOINTR);
-	/* Confirm the sequence as being established */
-	nfs_confirm_seqid(&sp->so_seqid, status);
-	nfs_increment_open_seqid(status, o_arg.seqid);
-	if (status == 0) {
-		memcpy(&state->stateid, &o_res.stateid, sizeof(state->stateid));
-		if (o_res.delegation_type != 0) {
-			nfs_inode_reclaim_delegation(inode, sp->so_cred, &o_res);
-			/* Did the server issue an immediate delegation recall? */
-			if (o_res.do_recall)
-				nfs_async_inode_return_delegation(inode, &o_res.stateid);
-		}
-	}
-	nfs_free_seqid(o_arg.seqid);
-	clear_bit(NFS_DELEGATED_STATE, &state->flags);
-	/* Ensure we update the inode attributes */
-	NFS_CACHEINV(inode);
+	opendata->o_arg.claim = NFS4_OPEN_CLAIM_PREVIOUS;
+	opendata->o_arg.fh = NFS_FH(state->inode);
+	nfs_copy_fh(&opendata->o_res.fh, opendata->o_arg.fh);
+	opendata->o_arg.u.delegation_type = delegation_type;
+	status = nfs4_open_recover(opendata, state);
+	nfs4_opendata_free(opendata);
 	return status;
 }
 
-static int nfs4_open_reclaim(struct nfs4_state_owner *sp, struct nfs4_state *state)
+static int nfs4_do_open_reclaim(struct nfs4_state_owner *sp, struct nfs4_state *state, struct dentry *dentry)
 {
 	struct nfs_server *server = NFS_SERVER(state->inode);
 	struct nfs4_exception exception = { };
 	int err;
 	do {
-		err = _nfs4_open_reclaim(sp, state);
+		err = _nfs4_do_open_reclaim(sp, state, dentry);
 		if (err != -NFS4ERR_DELAY)
 			break;
 		nfs4_handle_exception(server, err, &exception);
@@ -301,63 +486,36 @@ static int nfs4_open_reclaim(struct nfs4_state_owner *sp, struct nfs4_state *sta
 	return err;
 }
 
+static int nfs4_open_reclaim(struct nfs4_state_owner *sp, struct nfs4_state *state)
+{
+	struct nfs_open_context *ctx;
+	int ret;
+
+	ctx = nfs4_state_find_open_context(state);
+	if (IS_ERR(ctx))
+		return PTR_ERR(ctx);
+	ret = nfs4_do_open_reclaim(sp, state, ctx->dentry);
+	put_nfs_open_context(ctx);
+	return ret;
+}
+
 static int _nfs4_open_delegation_recall(struct dentry *dentry, struct nfs4_state *state)
 {
 	struct nfs4_state_owner  *sp  = state->owner;
-	struct inode *inode = dentry->d_inode;
-	struct nfs_server *server = NFS_SERVER(inode);
-	struct dentry *parent = dget_parent(dentry);
-	struct nfs_openargs arg = {
-		.fh = NFS_FH(parent->d_inode),
-		.clientid = server->nfs4_state->cl_clientid,
-		.name = &dentry->d_name,
-		.id = sp->so_id,
-		.server = server,
-		.bitmask = server->attr_bitmask,
-		.claim = NFS4_OPEN_CLAIM_DELEGATE_CUR,
-	};
-	struct nfs_openres res = {
-		.server = server,
-	};
-	struct 	rpc_message msg = {
-		.rpc_proc       = &nfs4_procedures[NFSPROC4_CLNT_OPEN_NOATTR],
-		.rpc_argp       = &arg,
-		.rpc_resp       = &res,
-		.rpc_cred	= sp->so_cred,
-	};
-	int status = 0;
+	struct nfs4_opendata *opendata;
+	int ret;
 
 	if (!test_bit(NFS_DELEGATED_STATE, &state->flags))
-		goto out;
-	if (state->state == 0)
-		goto out;
-	arg.seqid = nfs_alloc_seqid(&sp->so_seqid);
-	status = -ENOMEM;
-	if (arg.seqid == NULL)
-		goto out;
-	arg.open_flags = state->state;
-	memcpy(arg.u.delegation.data, state->stateid.data, sizeof(arg.u.delegation.data));
-	status = rpc_call_sync(server->client, &msg, RPC_TASK_NOINTR);
-	nfs_increment_open_seqid(status, arg.seqid);
-	if (status != 0)
-		goto out_free;
-	if(res.rflags & NFS4_OPEN_RESULT_CONFIRM) {
-		status = _nfs4_proc_open_confirm(server->client, NFS_FH(inode),
-				sp, &res.stateid, arg.seqid);
-		if (status != 0)
-			goto out_free;
-	}
-	nfs_confirm_seqid(&sp->so_seqid, 0);
-	if (status >= 0) {
-		memcpy(state->stateid.data, res.stateid.data,
-				sizeof(state->stateid.data));
-		clear_bit(NFS_DELEGATED_STATE, &state->flags);
-	}
-out_free:
-	nfs_free_seqid(arg.seqid);
-out:
-	dput(parent);
-	return status;
+		return 0;
+	opendata = nfs4_opendata_alloc(dentry, sp, 0, NULL);
+	if (opendata == NULL)
+		return -ENOMEM;
+	opendata->o_arg.claim = NFS4_OPEN_CLAIM_DELEGATE_CUR;
+	memcpy(opendata->o_arg.u.delegation.data, state->stateid.data,
+			sizeof(opendata->o_arg.u.delegation.data));
+	ret = nfs4_open_recover(opendata, state);
+	nfs4_opendata_free(opendata);
+	return ret;
 }
 
 int nfs4_open_delegation_recall(struct dentry *dentry, struct nfs4_state *state)
@@ -382,82 +540,202 @@ int nfs4_open_delegation_recall(struct dentry *dentry, struct nfs4_state *state)
 	return err;
 }
 
-static int _nfs4_proc_open_confirm(struct rpc_clnt *clnt, const struct nfs_fh *fh, struct nfs4_state_owner *sp, nfs4_stateid *stateid, struct nfs_seqid *seqid)
+static void nfs4_open_confirm_prepare(struct rpc_task *task, void *calldata)
 {
-	struct nfs_open_confirmargs arg = {
-		.fh             = fh,
-		.seqid          = seqid,
-		.stateid	= *stateid,
-	};
-	struct nfs_open_confirmres res;
-	struct 	rpc_message msg = {
-		.rpc_proc       = &nfs4_procedures[NFSPROC4_CLNT_OPEN_CONFIRM],
-		.rpc_argp       = &arg,
-		.rpc_resp       = &res,
-		.rpc_cred	= sp->so_cred,
+	struct nfs4_opendata *data = calldata;
+	struct  rpc_message msg = {
+		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_CONFIRM],
+		.rpc_argp = &data->c_arg,
+		.rpc_resp = &data->c_res,
+		.rpc_cred = data->owner->so_cred,
 	};
+	data->timestamp = jiffies;
+	rpc_call_setup(task, &msg, 0);
+}
+
+static void nfs4_open_confirm_done(struct rpc_task *task, void *calldata)
+{
+	struct nfs4_opendata *data = calldata;
+
+	data->rpc_status = task->tk_status;
+	if (RPC_ASSASSINATED(task))
+		return;
+	if (data->rpc_status == 0) {
+		memcpy(data->o_res.stateid.data, data->c_res.stateid.data,
+				sizeof(data->o_res.stateid.data));
+		renew_lease(data->o_res.server, data->timestamp);
+	}
+	nfs_increment_open_seqid(data->rpc_status, data->c_arg.seqid);
+	nfs_confirm_seqid(&data->owner->so_seqid, data->rpc_status);
+}
+
+static void nfs4_open_confirm_release(void *calldata)
+{
+	struct nfs4_opendata *data = calldata;
+	struct nfs4_state *state = NULL;
+
+	/* If this request hasn't been cancelled, do nothing */
+	if (data->cancelled == 0)
+		goto out_free;
+	/* In case of error, no cleanup! */
+	if (data->rpc_status != 0)
+		goto out_free;
+	nfs_confirm_seqid(&data->owner->so_seqid, 0);
+	state = nfs4_opendata_to_nfs4_state(data);
+	if (state != NULL)
+		nfs4_close_state(state, data->o_arg.open_flags);
+out_free:
+	nfs4_opendata_free(data);
+}
+
+static const struct rpc_call_ops nfs4_open_confirm_ops = {
+	.rpc_call_prepare = nfs4_open_confirm_prepare,
+	.rpc_call_done = nfs4_open_confirm_done,
+	.rpc_release = nfs4_open_confirm_release,
+};
+
+/*
+ * Note: On error, nfs4_proc_open_confirm will free the struct nfs4_opendata
+ */
+static int _nfs4_proc_open_confirm(struct nfs4_opendata *data)
+{
+	struct nfs_server *server = NFS_SERVER(data->dir->d_inode);
+	struct rpc_task *task;
 	int status;
 
-	status = rpc_call_sync(clnt, &msg, RPC_TASK_NOINTR);
-	/* Confirm the sequence as being established */
-	nfs_confirm_seqid(&sp->so_seqid, status);
-	nfs_increment_open_seqid(status, seqid);
-	if (status >= 0)
-		memcpy(stateid, &res.stateid, sizeof(*stateid));
+	atomic_inc(&data->count);
+	task = rpc_run_task(server->client, RPC_TASK_ASYNC, &nfs4_open_confirm_ops, data);
+	if (IS_ERR(task)) {
+		nfs4_opendata_free(data);
+		return PTR_ERR(task);
+	}
+	status = nfs4_wait_for_completion_rpc_task(task);
+	if (status != 0) {
+		data->cancelled = 1;
+		smp_wmb();
+	} else
+		status = data->rpc_status;
+	rpc_release_task(task);
 	return status;
 }
 
-static int _nfs4_proc_open(struct inode *dir, struct nfs4_state_owner  *sp, struct nfs_openargs *o_arg, struct nfs_openres *o_res)
+static void nfs4_open_prepare(struct rpc_task *task, void *calldata)
 {
-	struct nfs_server *server = NFS_SERVER(dir);
+	struct nfs4_opendata *data = calldata;
+	struct nfs4_state_owner *sp = data->owner;
 	struct rpc_message msg = {
 		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN],
-		.rpc_argp = o_arg,
-		.rpc_resp = o_res,
+		.rpc_argp = &data->o_arg,
+		.rpc_resp = &data->o_res,
 		.rpc_cred = sp->so_cred,
 	};
-	int status;
+	
+	if (nfs_wait_on_sequence(data->o_arg.seqid, task) != 0)
+		return;
+	/* Update sequence id. */
+	data->o_arg.id = sp->so_id;
+	data->o_arg.clientid = sp->so_client->cl_clientid;
+	if (data->o_arg.claim == NFS4_OPEN_CLAIM_PREVIOUS)
+		msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_NOATTR];
+	data->timestamp = jiffies;
+	rpc_call_setup(task, &msg, 0);
+}
 
-	/* Update sequence id. The caller must serialize! */
-	o_arg->id = sp->so_id;
-	o_arg->clientid = sp->so_client->cl_clientid;
+static void nfs4_open_done(struct rpc_task *task, void *calldata)
+{
+	struct nfs4_opendata *data = calldata;
 
-	status = rpc_call_sync(server->client, &msg, RPC_TASK_NOINTR);
-	if (status == 0) {
-		/* OPEN on anything except a regular file is disallowed in NFSv4 */
-		switch (o_res->f_attr->mode & S_IFMT) {
+	data->rpc_status = task->tk_status;
+	if (RPC_ASSASSINATED(task))
+		return;
+	if (task->tk_status == 0) {
+		switch (data->o_res.f_attr->mode & S_IFMT) {
 			case S_IFREG:
 				break;
 			case S_IFLNK:
-				status = -ELOOP;
+				data->rpc_status = -ELOOP;
 				break;
 			case S_IFDIR:
-				status = -EISDIR;
+				data->rpc_status = -EISDIR;
 				break;
 			default:
-				status = -ENOTDIR;
+				data->rpc_status = -ENOTDIR;
 		}
+		renew_lease(data->o_res.server, data->timestamp);
 	}
+	nfs_increment_open_seqid(data->rpc_status, data->o_arg.seqid);
+}
+
+static void nfs4_open_release(void *calldata)
+{
+	struct nfs4_opendata *data = calldata;
+	struct nfs4_state *state = NULL;
 
-	nfs_increment_open_seqid(status, o_arg->seqid);
+	/* If this request hasn't been cancelled, do nothing */
+	if (data->cancelled == 0)
+		goto out_free;
+	/* In case of error, no cleanup! */
+	if (data->rpc_status != 0)
+		goto out_free;
+	/* In case we need an open_confirm, no cleanup! */
+	if (data->o_res.rflags & NFS4_OPEN_RESULT_CONFIRM)
+		goto out_free;
+	nfs_confirm_seqid(&data->owner->so_seqid, 0);
+	state = nfs4_opendata_to_nfs4_state(data);
+	if (state != NULL)
+		nfs4_close_state(state, data->o_arg.open_flags);
+out_free:
+	nfs4_opendata_free(data);
+}
+
+static const struct rpc_call_ops nfs4_open_ops = {
+	.rpc_call_prepare = nfs4_open_prepare,
+	.rpc_call_done = nfs4_open_done,
+	.rpc_release = nfs4_open_release,
+};
+
+/*
+ * Note: On error, nfs4_proc_open will free the struct nfs4_opendata
+ */
+static int _nfs4_proc_open(struct nfs4_opendata *data)
+{
+	struct inode *dir = data->dir->d_inode;
+	struct nfs_server *server = NFS_SERVER(dir);
+	struct nfs_openargs *o_arg = &data->o_arg;
+	struct nfs_openres *o_res = &data->o_res;
+	struct rpc_task *task;
+	int status;
+
+	atomic_inc(&data->count);
+	task = rpc_run_task(server->client, RPC_TASK_ASYNC, &nfs4_open_ops, data);
+	if (IS_ERR(task)) {
+		nfs4_opendata_free(data);
+		return PTR_ERR(task);
+	}
+	status = nfs4_wait_for_completion_rpc_task(task);
+	if (status != 0) {
+		data->cancelled = 1;
+		smp_wmb();
+	} else
+		status = data->rpc_status;
+	rpc_release_task(task);
 	if (status != 0)
-		goto out;
+		return status;
+
 	if (o_arg->open_flags & O_CREAT) {
 		update_changeattr(dir, &o_res->cinfo);
 		nfs_post_op_update_inode(dir, o_res->dir_attr);
 	} else
 		nfs_refresh_inode(dir, o_res->dir_attr);
 	if(o_res->rflags & NFS4_OPEN_RESULT_CONFIRM) {
-		status = _nfs4_proc_open_confirm(server->client, &o_res->fh,
-				sp, &o_res->stateid, o_arg->seqid);
+		status = _nfs4_proc_open_confirm(data);
 		if (status != 0)
-			goto out;
+			return status;
 	}
-	nfs_confirm_seqid(&sp->so_seqid, 0);
+	nfs_confirm_seqid(&data->owner->so_seqid, 0);
 	if (!(o_res->f_attr->valid & NFS_ATTR_FATTR))
-		status = server->rpc_ops->getattr(server, &o_res->fh, o_res->f_attr);
-out:
-	return status;
+		return server->rpc_ops->getattr(server, &o_res->fh, o_res->f_attr);
+	return 0;
 }
 
 static int _nfs4_do_access(struct inode *inode, struct rpc_cred *cred, int openflags)
@@ -488,6 +766,15 @@ out:
 	return -EACCES;
 }
 
+int nfs4_recover_expired_lease(struct nfs_server *server)
+{
+	struct nfs4_client *clp = server->nfs4_state;
+
+	if (test_and_clear_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state))
+		nfs4_schedule_state_recovery(clp);
+	return nfs4_wait_clnt_recover(server->client, clp);
+}
+
 /*
  * OPEN_EXPIRED:
  * 	reclaim state on the server after a network partition.
@@ -495,77 +782,31 @@ out:
  */
 static int _nfs4_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *state, struct dentry *dentry)
 {
-	struct dentry *parent = dget_parent(dentry);
-	struct inode *dir = parent->d_inode;
 	struct inode *inode = state->inode;
-	struct nfs_server *server = NFS_SERVER(dir);
 	struct nfs_delegation *delegation = NFS_I(inode)->delegation;
-	struct nfs_fattr f_attr, dir_attr;
-	struct nfs_openargs o_arg = {
-		.fh = NFS_FH(dir),
-		.open_flags = state->state,
-		.name = &dentry->d_name,
-		.bitmask = server->attr_bitmask,
-		.claim = NFS4_OPEN_CLAIM_NULL,
-	};
-	struct nfs_openres o_res = {
-		.f_attr = &f_attr,
-		.dir_attr = &dir_attr,
-		.server = server,
-	};
-	int status = 0;
+	struct nfs4_opendata *opendata;
+	int openflags = state->state & (FMODE_READ|FMODE_WRITE);
+	int ret;
 
 	if (delegation != NULL && !(delegation->flags & NFS_DELEGATION_NEED_RECLAIM)) {
-		status = _nfs4_do_access(inode, sp->so_cred, state->state);
-		if (status < 0)
-			goto out;
+		ret = _nfs4_do_access(inode, sp->so_cred, openflags);
+		if (ret < 0)
+			return ret;
 		memcpy(&state->stateid, &delegation->stateid, sizeof(state->stateid));
 		set_bit(NFS_DELEGATED_STATE, &state->flags);
-		goto out;
+		return 0;
 	}
-	o_arg.seqid = nfs_alloc_seqid(&sp->so_seqid);
-	status = -ENOMEM;
-	if (o_arg.seqid == NULL)
-		goto out;
-	nfs_fattr_init(&f_attr);
-	nfs_fattr_init(&dir_attr);
-	status = _nfs4_proc_open(dir, sp, &o_arg, &o_res);
-	if (status != 0)
-		goto out_nodeleg;
-	/* Check if files differ */
-	if ((f_attr.mode & S_IFMT) != (inode->i_mode & S_IFMT))
-		goto out_stale;
-	/* Has the file handle changed? */
-	if (nfs_compare_fh(&o_res.fh, NFS_FH(inode)) != 0) {
-		/* Verify if the change attributes are the same */
-		if (f_attr.change_attr != NFS_I(inode)->change_attr)
-			goto out_stale;
-		if (nfs_size_to_loff_t(f_attr.size) != inode->i_size)
-			goto out_stale;
-		/* Lets just pretend that this is the same file */
-		nfs_copy_fh(NFS_FH(inode), &o_res.fh);
-		NFS_I(inode)->fileid = f_attr.fileid;
-	}
-	memcpy(&state->stateid, &o_res.stateid, sizeof(state->stateid));
-	if (o_res.delegation_type != 0) {
-		if (!(delegation->flags & NFS_DELEGATION_NEED_RECLAIM))
-			nfs_inode_set_delegation(inode, sp->so_cred, &o_res);
-		else
-			nfs_inode_reclaim_delegation(inode, sp->so_cred, &o_res);
+	opendata = nfs4_opendata_alloc(dentry, sp, openflags, NULL);
+	if (opendata == NULL)
+		return -ENOMEM;
+	ret = nfs4_open_recover(opendata, state);
+	if (ret == -ESTALE) {
+		/* Invalidate the state owner so we don't ever use it again */
+		nfs4_drop_state_owner(sp);
+		d_drop(dentry);
 	}
-out_nodeleg:
-	nfs_free_seqid(o_arg.seqid);
-	clear_bit(NFS_DELEGATED_STATE, &state->flags);
-out:
-	dput(parent);
-	return status;
-out_stale:
-	status = -ESTALE;
-	/* Invalidate the state owner so we don't ever use it again */
-	nfs4_drop_state_owner(sp);
-	d_drop(dentry);
-	/* Should we be trying to close that stateid? */
-	goto out_nodeleg;
+	nfs4_opendata_free(opendata);
+	return ret;
 }
 
 static inline int nfs4_do_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *state, struct dentry *dentry)
@@ -584,26 +825,19 @@ static inline int nfs4_do_open_expired(struct nfs4_state_owner *sp, struct nfs4_
 
 static int nfs4_open_expired(struct nfs4_state_owner *sp, struct nfs4_state *state)
 {
-	struct nfs_inode *nfsi = NFS_I(state->inode);
 	struct nfs_open_context *ctx;
-	int status;
+	int ret;
 
-	spin_lock(&state->inode->i_lock);
-	list_for_each_entry(ctx, &nfsi->open_files, list) {
-		if (ctx->state != state)
-			continue;
-		get_nfs_open_context(ctx);
-		spin_unlock(&state->inode->i_lock);
-		status = nfs4_do_open_expired(sp, state, ctx->dentry);
-		put_nfs_open_context(ctx);
-		return status;
-	}
-	spin_unlock(&state->inode->i_lock);
-	return -ENOENT;
+	ctx = nfs4_state_find_open_context(state);
+	if (IS_ERR(ctx))
+		return PTR_ERR(ctx);
+	ret = nfs4_do_open_expired(sp, state, ctx->dentry);
+	put_nfs_open_context(ctx);
+	return ret;
 }
 
 /*
- * Returns an nfs4_state + an extra reference to the inode
+ * Returns a referenced nfs4_state if there is an open delegation on the file
  */
 static int _nfs4_open_delegated(struct inode *inode, int flags, struct rpc_cred *cred, struct nfs4_state **res)
 {
@@ -616,6 +850,14 @@ static int _nfs4_open_delegated(struct inode *inode, int flags, struct rpc_cred
 	int open_flags = flags & (FMODE_READ|FMODE_WRITE);
 	int err;
 
+	err = -ENOMEM;
+	if (!(sp = nfs4_get_state_owner(server, cred))) {
+		dprintk("%s: nfs4_get_state_owner failed!\n", __FUNCTION__);
+		return err;
+	}
+	err = nfs4_recover_expired_lease(server);
+	if (err != 0)
+		goto out_put_state_owner;
 	/* Protect against reboot recovery - NOTE ORDER! */
 	down_read(&clp->cl_sem);
 	/* Protect against delegation recall */
@@ -625,10 +867,6 @@ static int _nfs4_open_delegated(struct inode *inode, int flags, struct rpc_cred
 	if (delegation == NULL || (delegation->type & open_flags) != open_flags)
 		goto out_err;
 	err = -ENOMEM;
-	if (!(sp = nfs4_get_state_owner(server, cred))) {
-		dprintk("%s: nfs4_get_state_owner failed!\n", __FUNCTION__);
-		goto out_err;
-	}
 	state = nfs4_get_open_state(inode, sp);
 	if (state == NULL)
 		goto out_err;
@@ -636,39 +874,34 @@ static int _nfs4_open_delegated(struct inode *inode, int flags, struct rpc_cred
 	err = -ENOENT;
 	if ((state->state & open_flags) == open_flags) {
 		spin_lock(&inode->i_lock);
-		if (open_flags & FMODE_READ)
-			state->nreaders++;
-		if (open_flags & FMODE_WRITE)
-			state->nwriters++;
+		update_open_stateflags(state, open_flags);
 		spin_unlock(&inode->i_lock);
 		goto out_ok;
 	} else if (state->state != 0)
-		goto out_err;
+		goto out_put_open_state;
 
 	lock_kernel();
 	err = _nfs4_do_access(inode, cred, open_flags);
 	unlock_kernel();
 	if (err != 0)
-		goto out_err;
+		goto out_put_open_state;
 	set_bit(NFS_DELEGATED_STATE, &state->flags);
 	update_open_stateid(state, &delegation->stateid, open_flags);
 out_ok:
 	nfs4_put_state_owner(sp);
 	up_read(&nfsi->rwsem);
 	up_read(&clp->cl_sem);
-	igrab(inode);
 	*res = state;
-	return 0; 
+	return 0;
+out_put_open_state:
+	nfs4_put_open_state(state);
 out_err:
-	if (sp != NULL) {
-		if (state != NULL)
-			nfs4_put_open_state(state);
-		nfs4_put_state_owner(sp);
-	}
 	up_read(&nfsi->rwsem);
 	up_read(&clp->cl_sem);
 	if (err != -EACCES)
 		nfs_inode_return_delegation(inode);
+out_put_state_owner:
+	nfs4_put_state_owner(sp);
 	return err;
 }
 
@@ -689,7 +922,7 @@ static struct nfs4_state *nfs4_open_delegated(struct inode *inode, int flags, st
 }
 
 /*
- * Returns an nfs4_state + an referenced inode
+ * Returns a referenced nfs4_state
  */
 static int _nfs4_do_open(struct inode *dir, struct dentry *dentry, int flags, struct iattr *sattr, struct rpc_cred *cred, struct nfs4_state **res)
 {
@@ -697,73 +930,46 @@ static int _nfs4_do_open(struct inode *dir, struct dentry *dentry, int flags, st
 	struct nfs4_state     *state = NULL;
 	struct nfs_server       *server = NFS_SERVER(dir);
 	struct nfs4_client *clp = server->nfs4_state;
-	struct inode *inode = NULL;
+	struct nfs4_opendata *opendata;
 	int                     status;
-	struct nfs_fattr f_attr, dir_attr;
-	struct nfs_openargs o_arg = {
-		.fh             = NFS_FH(dir),
-		.open_flags	= flags,
-		.name           = &dentry->d_name,
-		.server         = server,
-		.bitmask = server->attr_bitmask,
-		.claim = NFS4_OPEN_CLAIM_NULL,
-	};
-	struct nfs_openres o_res = {
-		.f_attr         = &f_attr,
-		.dir_attr	= &dir_attr,
-		.server         = server,
-	};
 
 	/* Protect against reboot recovery conflicts */
-	down_read(&clp->cl_sem);
 	status = -ENOMEM;
 	if (!(sp = nfs4_get_state_owner(server, cred))) {
 		dprintk("nfs4_do_open: nfs4_get_state_owner failed!\n");
 		goto out_err;
 	}
-	if (flags & O_EXCL) {
-		u32 *p = (u32 *) o_arg.u.verifier.data;
-		p[0] = jiffies;
-		p[1] = current->pid;
-	} else
-		o_arg.u.attrs = sattr;
-	/* Serialization for the sequence id */
+	status = nfs4_recover_expired_lease(server);
+	if (status != 0)
+		goto err_put_state_owner;
+	down_read(&clp->cl_sem);
+	status = -ENOMEM;
+	opendata = nfs4_opendata_alloc(dentry, sp, flags, sattr);
+	if (opendata == NULL)
+		goto err_put_state_owner;
 
-	o_arg.seqid = nfs_alloc_seqid(&sp->so_seqid);
-	if (o_arg.seqid == NULL)
-		return -ENOMEM;
-	nfs_fattr_init(&f_attr);
-	nfs_fattr_init(&dir_attr);
-	status = _nfs4_proc_open(dir, sp, &o_arg, &o_res);
+	status = _nfs4_proc_open(opendata);
 	if (status != 0)
-		goto out_err;
+		goto err_opendata_free;
 
 	status = -ENOMEM;
-	inode = nfs_fhget(dir->i_sb, &o_res.fh, &f_attr);
-	if (!inode)
-		goto out_err;
-	state = nfs4_get_open_state(inode, sp);
-	if (!state)
-		goto out_err;
-	update_open_stateid(state, &o_res.stateid, flags);
-	if (o_res.delegation_type != 0)
-		nfs_inode_set_delegation(inode, cred, &o_res);
-	nfs_free_seqid(o_arg.seqid);
+	state = nfs4_opendata_to_nfs4_state(opendata);
+	if (state == NULL)
+		goto err_opendata_free;
+	if (opendata->o_res.delegation_type != 0)
+		nfs_inode_set_delegation(state->inode, cred, &opendata->o_res);
+	nfs4_opendata_free(opendata);
 	nfs4_put_state_owner(sp);
 	up_read(&clp->cl_sem);
 	*res = state;
 	return 0;
+err_opendata_free:
+	nfs4_opendata_free(opendata);
+err_put_state_owner:
+	nfs4_put_state_owner(sp);
 out_err:
-	if (sp != NULL) {
-		if (state != NULL)
-			nfs4_put_open_state(state);
-		nfs_free_seqid(o_arg.seqid);
-		nfs4_put_state_owner(sp);
-	}
 	/* Note: clp->cl_sem must be released before nfs4_put_open_state()! */
 	up_read(&clp->cl_sem);
-	if (inode != NULL)
-		iput(inode);
 	*res = NULL;
 	return status;
 }
@@ -830,6 +1036,7 @@ static int _nfs4_do_setattr(struct nfs_server *server, struct nfs_fattr *fattr,
                 .rpc_argp       = &arg,
                 .rpc_resp       = &res,
         };
+	unsigned long timestamp = jiffies;
 	int status;
 
 	nfs_fattr_init(fattr);
@@ -841,6 +1048,8 @@ static int _nfs4_do_setattr(struct nfs_server *server, struct nfs_fattr *fattr,
 		memcpy(&arg.stateid, &zero_stateid, sizeof(arg.stateid));
 
 	status = rpc_call_sync(server->client, &msg, 0);
+	if (status == 0 && state != NULL)
+		renew_lease(server, timestamp);
 	return status;
 }
 
@@ -865,12 +1074,13 @@ struct nfs4_closedata {
 	struct nfs_closeargs arg;
 	struct nfs_closeres res;
 	struct nfs_fattr fattr;
+	unsigned long timestamp;
 };
 
-static void nfs4_free_closedata(struct nfs4_closedata *calldata)
+static void nfs4_free_closedata(void *data)
 {
-	struct nfs4_state *state = calldata->state;
-	struct nfs4_state_owner *sp = state->owner;
+	struct nfs4_closedata *calldata = data;
+	struct nfs4_state_owner *sp = calldata->state->owner;
 
 	nfs4_put_open_state(calldata->state);
 	nfs_free_seqid(calldata->arg.seqid);
@@ -878,12 +1088,14 @@ static void nfs4_free_closedata(struct nfs4_closedata *calldata)
 	kfree(calldata);
 }
 
-static void nfs4_close_done(struct rpc_task *task)
+static void nfs4_close_done(struct rpc_task *task, void *data)
 {
-	struct nfs4_closedata *calldata = (struct nfs4_closedata *)task->tk_calldata;
+	struct nfs4_closedata *calldata = data;
 	struct nfs4_state *state = calldata->state;
 	struct nfs_server *server = NFS_SERVER(calldata->inode);
 
+	if (RPC_ASSASSINATED(task))
+		return;
         /* hmm. we are done with the inode, and in the process of freeing
 	 * the state_owner. we keep this around to process errors
 	 */
@@ -892,6 +1104,7 @@ static void nfs4_close_done(struct rpc_task *task)
 		case 0:
 			memcpy(&state->stateid, &calldata->res.stateid,
 					sizeof(state->stateid));
+			renew_lease(server, calldata->timestamp);
 			break;
 		case -NFS4ERR_STALE_STATEID:
 		case -NFS4ERR_EXPIRED:
@@ -904,12 +1117,11 @@ static void nfs4_close_done(struct rpc_task *task)
 			}
 	}
 	nfs_refresh_inode(calldata->inode, calldata->res.fattr);
-	nfs4_free_closedata(calldata);
 }
 
-static void nfs4_close_begin(struct rpc_task *task)
+static void nfs4_close_prepare(struct rpc_task *task, void *data)
 {
-	struct nfs4_closedata *calldata = (struct nfs4_closedata *)task->tk_calldata;
+	struct nfs4_closedata *calldata = data;
 	struct nfs4_state *state = calldata->state;
 	struct rpc_message msg = {
 		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_CLOSE],
@@ -918,10 +1130,8 @@ static void nfs4_close_begin(struct rpc_task *task)
 		.rpc_cred = state->owner->so_cred,
 	};
 	int mode = 0, old_mode;
-	int status;
 
-	status = nfs_wait_on_sequence(calldata->arg.seqid, task);
-	if (status != 0)
+	if (nfs_wait_on_sequence(calldata->arg.seqid, task) != 0)
 		return;
 	/* Recalculate the new open mode in case someone reopened the file
 	 * while we were waiting in line to be scheduled.
@@ -929,26 +1139,34 @@ static void nfs4_close_begin(struct rpc_task *task)
 	spin_lock(&state->owner->so_lock);
 	spin_lock(&calldata->inode->i_lock);
 	mode = old_mode = state->state;
-	if (state->nreaders == 0)
-		mode &= ~FMODE_READ;
-	if (state->nwriters == 0)
-		mode &= ~FMODE_WRITE;
+	if (state->n_rdwr == 0) {
+		if (state->n_rdonly == 0)
+			mode &= ~FMODE_READ;
+		if (state->n_wronly == 0)
+			mode &= ~FMODE_WRITE;
+	}
 	nfs4_state_set_mode_locked(state, mode);
 	spin_unlock(&calldata->inode->i_lock);
 	spin_unlock(&state->owner->so_lock);
 	if (mode == old_mode || test_bit(NFS_DELEGATED_STATE, &state->flags)) {
-		nfs4_free_closedata(calldata);
-		task->tk_exit = NULL;
-		rpc_exit(task, 0);
+		/* Note: exit _without_ calling nfs4_close_done */
+		task->tk_action = NULL;
 		return;
 	}
 	nfs_fattr_init(calldata->res.fattr);
 	if (mode != 0)
 		msg.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_OPEN_DOWNGRADE];
 	calldata->arg.open_flags = mode;
+	calldata->timestamp = jiffies;
 	rpc_call_setup(task, &msg, 0);
 }
 
+static const struct rpc_call_ops nfs4_close_ops = {
+	.rpc_call_prepare = nfs4_close_prepare,
+	.rpc_call_done = nfs4_close_done,
+	.rpc_release = nfs4_free_closedata,
+};
+
 /* 
  * It is possible for data to be read/written from a mem-mapped file 
  * after the sys_close call (which hits the vfs layer as a flush).
@@ -981,8 +1199,7 @@ int nfs4_do_close(struct inode *inode, struct nfs4_state *state)
 	calldata->res.fattr = &calldata->fattr;
 	calldata->res.server = server;
 
-	status = nfs4_call_async(server->client, nfs4_close_begin,
-			nfs4_close_done, calldata);
+	status = nfs4_call_async(server->client, &nfs4_close_ops, calldata);
 	if (status == 0)
 		goto out;
 
@@ -1034,7 +1251,7 @@ nfs4_atomic_open(struct inode *dir, struct dentry *dentry, struct nameidata *nd)
 			d_add(dentry, NULL);
 		return (struct dentry *)state;
 	}
-	res = d_add_unique(dentry, state->inode);
+	res = d_add_unique(dentry, igrab(state->inode));
 	if (res != NULL)
 		dentry = res;
 	nfs4_intent_set_file(nd, dentry, state);
@@ -1046,7 +1263,6 @@ nfs4_open_revalidate(struct inode *dir, struct dentry *dentry, int openflags, st
 {
 	struct rpc_cred *cred;
 	struct nfs4_state *state;
-	struct inode *inode;
 
 	cred = rpcauth_lookupcred(NFS_SERVER(dir)->client->cl_auth, 0);
 	if (IS_ERR(cred))
@@ -1070,9 +1286,7 @@ nfs4_open_revalidate(struct inode *dir, struct dentry *dentry, int openflags, st
 		}
 		goto out_drop;
 	}
-	inode = state->inode;
-	iput(inode);
-	if (inode == dentry->d_inode) {
+	if (state->inode == dentry->d_inode) {
 		nfs4_intent_set_file(nd, dentry, state);
 		return 1;
 	}
@@ -1508,11 +1722,13 @@ static int _nfs4_proc_write(struct nfs_write_data *wdata)
 
 	wdata->args.bitmask = server->attr_bitmask;
 	wdata->res.server = server;
+	wdata->timestamp = jiffies;
 	nfs_fattr_init(fattr);
 	status = rpc_call_sync(server->client, &msg, rpcflags);
 	dprintk("NFS reply write: %d\n", status);
 	if (status < 0)
 		return status;
+	renew_lease(server, wdata->timestamp);
 	nfs_post_op_update_inode(inode, fattr);
 	return wdata->res.count;
 }
@@ -1547,8 +1763,11 @@ static int _nfs4_proc_commit(struct nfs_write_data *cdata)
 
 	cdata->args.bitmask = server->attr_bitmask;
 	cdata->res.server = server;
+	cdata->timestamp = jiffies;
 	nfs_fattr_init(fattr);
 	status = rpc_call_sync(server->client, &msg, 0);
+	if (status >= 0)
+		renew_lease(server, cdata->timestamp);
 	dprintk("NFS reply commit: %d\n", status);
 	if (status >= 0)
 		nfs_post_op_update_inode(inode, fattr);
@@ -1601,7 +1820,7 @@ nfs4_proc_create(struct inode *dir, struct dentry *dentry, struct iattr *sattr,
 		status = PTR_ERR(state);
 		goto out;
 	}
-	d_instantiate(dentry, state->inode);
+	d_instantiate(dentry, igrab(state->inode));
 	if (flags & O_EXCL) {
 		struct nfs_fattr fattr;
 		status = nfs4_do_setattr(NFS_SERVER(dir), &fattr,
@@ -2125,10 +2344,9 @@ static int nfs4_proc_pathconf(struct nfs_server *server, struct nfs_fh *fhandle,
 	return err;
 }
 
-static void
-nfs4_read_done(struct rpc_task *task)
+static void nfs4_read_done(struct rpc_task *task, void *calldata)
 {
-	struct nfs_read_data *data = (struct nfs_read_data *) task->tk_calldata;
+	struct nfs_read_data *data = calldata;
 	struct inode *inode = data->inode;
 
 	if (nfs4_async_handle_error(task, NFS_SERVER(inode)) == -EAGAIN) {
@@ -2138,9 +2356,14 @@ nfs4_read_done(struct rpc_task *task)
 	if (task->tk_status > 0)
 		renew_lease(NFS_SERVER(inode), data->timestamp);
 	/* Call back common NFS readpage processing */
-	nfs_readpage_result(task);
+	nfs_readpage_result(task, calldata);
 }
 
+static const struct rpc_call_ops nfs4_read_ops = {
+	.rpc_call_done = nfs4_read_done,
+	.rpc_release = nfs_readdata_release,
+};
+
 static void
 nfs4_proc_read_setup(struct nfs_read_data *data)
 {
@@ -2160,14 +2383,13 @@ nfs4_proc_read_setup(struct nfs_read_data *data)
 	flags = RPC_TASK_ASYNC | (IS_SWAPFILE(inode)? NFS_RPC_SWAPFLAGS : 0);
 
 	/* Finalize the task. */
-	rpc_init_task(task, NFS_CLIENT(inode), nfs4_read_done, flags);
+	rpc_init_task(task, NFS_CLIENT(inode), flags, &nfs4_read_ops, data);
 	rpc_call_setup(task, &msg, 0);
 }
 
-static void
-nfs4_write_done(struct rpc_task *task)
+static void nfs4_write_done(struct rpc_task *task, void *calldata)
 {
-	struct nfs_write_data *data = (struct nfs_write_data *) task->tk_calldata;
+	struct nfs_write_data *data = calldata;
 	struct inode *inode = data->inode;
 	
 	if (nfs4_async_handle_error(task, NFS_SERVER(inode)) == -EAGAIN) {
@@ -2179,9 +2401,14 @@ nfs4_write_done(struct rpc_task *task)
 		nfs_post_op_update_inode(inode, data->res.fattr);
 	}
 	/* Call back common NFS writeback processing */
-	nfs_writeback_done(task);
+	nfs_writeback_done(task, calldata);
 }
 
+static const struct rpc_call_ops nfs4_write_ops = {
+	.rpc_call_done = nfs4_write_done,
+	.rpc_release = nfs_writedata_release,
+};
+
 static void
 nfs4_proc_write_setup(struct nfs_write_data *data, int how)
 {
@@ -2214,14 +2441,13 @@ nfs4_proc_write_setup(struct nfs_write_data *data, int how)
 	flags = (how & FLUSH_SYNC) ? 0 : RPC_TASK_ASYNC;
 
 	/* Finalize the task. */
-	rpc_init_task(task, NFS_CLIENT(inode), nfs4_write_done, flags);
+	rpc_init_task(task, NFS_CLIENT(inode), flags, &nfs4_write_ops, data);
 	rpc_call_setup(task, &msg, 0);
 }
 
-static void
-nfs4_commit_done(struct rpc_task *task)
+static void nfs4_commit_done(struct rpc_task *task, void *calldata)
 {
-	struct nfs_write_data *data = (struct nfs_write_data *) task->tk_calldata;
+	struct nfs_write_data *data = calldata;
 	struct inode *inode = data->inode;
 	
 	if (nfs4_async_handle_error(task, NFS_SERVER(inode)) == -EAGAIN) {
@@ -2231,9 +2457,14 @@ nfs4_commit_done(struct rpc_task *task)
 	if (task->tk_status >= 0)
 		nfs_post_op_update_inode(inode, data->res.fattr);
 	/* Call back common NFS writeback processing */
-	nfs_commit_done(task);
+	nfs_commit_done(task, calldata);
 }
 
+static const struct rpc_call_ops nfs4_commit_ops = {
+	.rpc_call_done = nfs4_commit_done,
+	.rpc_release = nfs_commit_release,
+};
+
 static void
 nfs4_proc_commit_setup(struct nfs_write_data *data, int how)
 {
@@ -2255,7 +2486,7 @@ nfs4_proc_commit_setup(struct nfs_write_data *data, int how)
 	flags = (how & FLUSH_SYNC) ? 0 : RPC_TASK_ASYNC;
 
 	/* Finalize the task. */
-	rpc_init_task(task, NFS_CLIENT(inode), nfs4_commit_done, flags);
+	rpc_init_task(task, NFS_CLIENT(inode), flags, &nfs4_commit_ops, data);
 	rpc_call_setup(task, &msg, 0);	
 }
 
@@ -2263,11 +2494,10 @@ nfs4_proc_commit_setup(struct nfs_write_data *data, int how)
  * nfs4_proc_async_renew(): This is not one of the nfs_rpc_ops; it is a special
  * standalone procedure for queueing an asynchronous RENEW.
  */
-static void
-renew_done(struct rpc_task *task)
+static void nfs4_renew_done(struct rpc_task *task, void *data)
 {
 	struct nfs4_client *clp = (struct nfs4_client *)task->tk_msg.rpc_argp;
-	unsigned long timestamp = (unsigned long)task->tk_calldata;
+	unsigned long timestamp = (unsigned long)data;
 
 	if (task->tk_status < 0) {
 		switch (task->tk_status) {
@@ -2284,26 +2514,28 @@ renew_done(struct rpc_task *task)
 	spin_unlock(&clp->cl_lock);
 }
 
-int
-nfs4_proc_async_renew(struct nfs4_client *clp)
+static const struct rpc_call_ops nfs4_renew_ops = {
+	.rpc_call_done = nfs4_renew_done,
+};
+
+int nfs4_proc_async_renew(struct nfs4_client *clp, struct rpc_cred *cred)
 {
 	struct rpc_message msg = {
 		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_RENEW],
 		.rpc_argp	= clp,
-		.rpc_cred	= clp->cl_cred,
+		.rpc_cred	= cred,
 	};
 
 	return rpc_call_async(clp->cl_rpcclient, &msg, RPC_TASK_SOFT,
-			renew_done, (void *)jiffies);
+			&nfs4_renew_ops, (void *)jiffies);
 }
 
-int
-nfs4_proc_renew(struct nfs4_client *clp)
+int nfs4_proc_renew(struct nfs4_client *clp, struct rpc_cred *cred)
 {
 	struct rpc_message msg = {
 		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_RENEW],
 		.rpc_argp	= clp,
-		.rpc_cred	= clp->cl_cred,
+		.rpc_cred	= cred,
 	};
 	unsigned long now = jiffies;
 	int status;
@@ -2519,7 +2751,7 @@ nfs4_async_handle_error(struct rpc_task *task, const struct nfs_server *server)
 		case -NFS4ERR_EXPIRED:
 			rpc_sleep_on(&clp->cl_rpcwaitq, task, NULL, NULL);
 			nfs4_schedule_state_recovery(clp);
-			if (test_bit(NFS4CLNT_OK, &clp->cl_state))
+			if (test_bit(NFS4CLNT_STATE_RECOVER, &clp->cl_state) == 0)
 				rpc_wake_up_task(task);
 			task->tk_status = 0;
 			return -EAGAIN;
@@ -2536,25 +2768,25 @@ nfs4_async_handle_error(struct rpc_task *task, const struct nfs_server *server)
 	return 0;
 }
 
+static int nfs4_wait_bit_interruptible(void *word)
+{
+	if (signal_pending(current))
+		return -ERESTARTSYS;
+	schedule();
+	return 0;
+}
+
 static int nfs4_wait_clnt_recover(struct rpc_clnt *clnt, struct nfs4_client *clp)
 {
-	DEFINE_WAIT(wait);
 	sigset_t oldset;
-	int interruptible, res = 0;
+	int res;
 
 	might_sleep();
 
 	rpc_clnt_sigmask(clnt, &oldset);
-	interruptible = TASK_UNINTERRUPTIBLE;
-	if (clnt->cl_intr)
-		interruptible = TASK_INTERRUPTIBLE;
-	prepare_to_wait(&clp->cl_waitq, &wait, interruptible);
-	nfs4_schedule_state_recovery(clp);
-	if (clnt->cl_intr && signalled())
-		res = -ERESTARTSYS;
-	else if (!test_bit(NFS4CLNT_OK, &clp->cl_state))
-		schedule();
-	finish_wait(&clp->cl_waitq, &wait);
+	res = wait_on_bit(&clp->cl_state, NFS4CLNT_STATE_RECOVER,
+			nfs4_wait_bit_interruptible,
+			TASK_INTERRUPTIBLE);
 	rpc_clnt_sigunmask(clnt, &oldset);
 	return res;
 }
@@ -2597,6 +2829,7 @@ int nfs4_handle_exception(const struct nfs_server *server, int errorcode, struct
 		case -NFS4ERR_STALE_CLIENTID:
 		case -NFS4ERR_STALE_STATEID:
 		case -NFS4ERR_EXPIRED:
+			nfs4_schedule_state_recovery(clp);
 			ret = nfs4_wait_clnt_recover(server->client, clp);
 			if (ret == 0)
 				exception->retry = 1;
@@ -2613,7 +2846,7 @@ int nfs4_handle_exception(const struct nfs_server *server, int errorcode, struct
 	return nfs4_map_errors(ret);
 }
 
-int nfs4_proc_setclientid(struct nfs4_client *clp, u32 program, unsigned short port)
+int nfs4_proc_setclientid(struct nfs4_client *clp, u32 program, unsigned short port, struct rpc_cred *cred)
 {
 	nfs4_verifier sc_verifier;
 	struct nfs4_setclientid setclientid = {
@@ -2624,7 +2857,7 @@ int nfs4_proc_setclientid(struct nfs4_client *clp, u32 program, unsigned short p
 		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETCLIENTID],
 		.rpc_argp = &setclientid,
 		.rpc_resp = clp,
-		.rpc_cred = clp->cl_cred,
+		.rpc_cred = cred,
 	};
 	u32 *p;
 	int loop = 0;
@@ -2638,7 +2871,7 @@ int nfs4_proc_setclientid(struct nfs4_client *clp, u32 program, unsigned short p
 		setclientid.sc_name_len = scnprintf(setclientid.sc_name,
 				sizeof(setclientid.sc_name), "%s/%u.%u.%u.%u %s %u",
 				clp->cl_ipaddr, NIPQUAD(clp->cl_addr.s_addr),
-				clp->cl_cred->cr_ops->cr_name,
+				cred->cr_ops->cr_name,
 				clp->cl_id_uniquifier);
 		setclientid.sc_netid_len = scnprintf(setclientid.sc_netid,
 				sizeof(setclientid.sc_netid), "tcp");
@@ -2661,14 +2894,14 @@ int nfs4_proc_setclientid(struct nfs4_client *clp, u32 program, unsigned short p
 }
 
 int
-nfs4_proc_setclientid_confirm(struct nfs4_client *clp)
+nfs4_proc_setclientid_confirm(struct nfs4_client *clp, struct rpc_cred *cred)
 {
 	struct nfs_fsinfo fsinfo;
 	struct rpc_message msg = {
 		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_SETCLIENTID_CONFIRM],
 		.rpc_argp = clp,
 		.rpc_resp = &fsinfo,
-		.rpc_cred = clp->cl_cred,
+		.rpc_cred = cred,
 	};
 	unsigned long now;
 	int status;
@@ -2679,24 +2912,92 @@ nfs4_proc_setclientid_confirm(struct nfs4_client *clp)
 		spin_lock(&clp->cl_lock);
 		clp->cl_lease_time = fsinfo.lease_time * HZ;
 		clp->cl_last_renewal = now;
+		clear_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state);
 		spin_unlock(&clp->cl_lock);
 	}
 	return status;
 }
 
-static int _nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4_stateid *stateid)
+struct nfs4_delegreturndata {
+	struct nfs4_delegreturnargs args;
+	struct nfs4_delegreturnres res;
+	struct nfs_fh fh;
+	nfs4_stateid stateid;
+	struct rpc_cred *cred;
+	unsigned long timestamp;
+	struct nfs_fattr fattr;
+	int rpc_status;
+};
+
+static void nfs4_delegreturn_prepare(struct rpc_task *task, void *calldata)
 {
-	struct nfs4_delegreturnargs args = {
-		.fhandle = NFS_FH(inode),
-		.stateid = stateid,
-	};
+	struct nfs4_delegreturndata *data = calldata;
 	struct rpc_message msg = {
 		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_DELEGRETURN],
-		.rpc_argp = &args,
-		.rpc_cred = cred,
+		.rpc_argp = &data->args,
+		.rpc_resp = &data->res,
+		.rpc_cred = data->cred,
 	};
+	nfs_fattr_init(data->res.fattr);
+	rpc_call_setup(task, &msg, 0);
+}
 
-	return rpc_call_sync(NFS_CLIENT(inode), &msg, 0);
+static void nfs4_delegreturn_done(struct rpc_task *task, void *calldata)
+{
+	struct nfs4_delegreturndata *data = calldata;
+	data->rpc_status = task->tk_status;
+	if (data->rpc_status == 0)
+		renew_lease(data->res.server, data->timestamp);
+}
+
+static void nfs4_delegreturn_release(void *calldata)
+{
+	struct nfs4_delegreturndata *data = calldata;
+
+	put_rpccred(data->cred);
+	kfree(calldata);
+}
+
+const static struct rpc_call_ops nfs4_delegreturn_ops = {
+	.rpc_call_prepare = nfs4_delegreturn_prepare,
+	.rpc_call_done = nfs4_delegreturn_done,
+	.rpc_release = nfs4_delegreturn_release,
+};
+
+static int _nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4_stateid *stateid)
+{
+	struct nfs4_delegreturndata *data;
+	struct nfs_server *server = NFS_SERVER(inode);
+	struct rpc_task *task;
+	int status;
+
+	data = kmalloc(sizeof(*data), GFP_KERNEL);
+	if (data == NULL)
+		return -ENOMEM;
+	data->args.fhandle = &data->fh;
+	data->args.stateid = &data->stateid;
+	data->args.bitmask = server->attr_bitmask;
+	nfs_copy_fh(&data->fh, NFS_FH(inode));
+	memcpy(&data->stateid, stateid, sizeof(data->stateid));
+	data->res.fattr = &data->fattr;
+	data->res.server = server;
+	data->cred = get_rpccred(cred);
+	data->timestamp = jiffies;
+	data->rpc_status = 0;
+
+	task = rpc_run_task(NFS_CLIENT(inode), RPC_TASK_ASYNC, &nfs4_delegreturn_ops, data);
+	if (IS_ERR(task)) {
+		nfs4_delegreturn_release(data);
+		return PTR_ERR(task);
+	}
+	status = nfs4_wait_for_completion_rpc_task(task);
+	if (status == 0) {
+		status = data->rpc_status;
+		if (status == 0)
+			nfs_post_op_update_inode(inode, &data->fattr);
+	}
+	rpc_release_task(task);
+	return status;
 }
 
 int nfs4_proc_delegreturn(struct inode *inode, struct rpc_cred *cred, const nfs4_stateid *stateid)
@@ -2734,43 +3035,17 @@ nfs4_set_lock_task_retry(unsigned long timeout)
 	return timeout;
 }
 
-static inline int
-nfs4_lck_type(int cmd, struct file_lock *request)
-{
-	/* set lock type */
-	switch (request->fl_type) {
-		case F_RDLCK:
-			return IS_SETLKW(cmd) ? NFS4_READW_LT : NFS4_READ_LT;
-		case F_WRLCK:
-			return IS_SETLKW(cmd) ? NFS4_WRITEW_LT : NFS4_WRITE_LT;
-		case F_UNLCK:
-			return NFS4_WRITE_LT; 
-	}
-	BUG();
-	return 0;
-}
-
-static inline uint64_t
-nfs4_lck_length(struct file_lock *request)
-{
-	if (request->fl_end == OFFSET_MAX)
-		return ~(uint64_t)0;
-	return request->fl_end - request->fl_start + 1;
-}
-
 static int _nfs4_proc_getlk(struct nfs4_state *state, int cmd, struct file_lock *request)
 {
 	struct inode *inode = state->inode;
 	struct nfs_server *server = NFS_SERVER(inode);
 	struct nfs4_client *clp = server->nfs4_state;
-	struct nfs_lockargs arg = {
+	struct nfs_lockt_args arg = {
 		.fh = NFS_FH(inode),
-		.type = nfs4_lck_type(cmd, request),
-		.offset = request->fl_start,
-		.length = nfs4_lck_length(request),
+		.fl = request,
 	};
-	struct nfs_lockres res = {
-		.server = server,
+	struct nfs_lockt_res res = {
+		.denied = request,
 	};
 	struct rpc_message msg = {
 		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_LOCKT],
@@ -2778,36 +3053,23 @@ static int _nfs4_proc_getlk(struct nfs4_state *state, int cmd, struct file_lock
 		.rpc_resp       = &res,
 		.rpc_cred	= state->owner->so_cred,
 	};
-	struct nfs_lowner nlo;
 	struct nfs4_lock_state *lsp;
 	int status;
 
 	down_read(&clp->cl_sem);
-	nlo.clientid = clp->cl_clientid;
+	arg.lock_owner.clientid = clp->cl_clientid;
 	status = nfs4_set_lock_state(state, request);
 	if (status != 0)
 		goto out;
 	lsp = request->fl_u.nfs4_fl.owner;
-	nlo.id = lsp->ls_id; 
-	arg.u.lockt = &nlo;
+	arg.lock_owner.id = lsp->ls_id; 
 	status = rpc_call_sync(server->client, &msg, 0);
-	if (!status) {
-		request->fl_type = F_UNLCK;
-	} else if (status == -NFS4ERR_DENIED) {
-		int64_t len, start, end;
-		start = res.u.denied.offset;
-		len = res.u.denied.length;
-		end = start + len - 1;
-		if (end < 0 || len == 0)
-			request->fl_end = OFFSET_MAX;
-		else
-			request->fl_end = (loff_t)end;
-		request->fl_start = (loff_t)start;
-		request->fl_type = F_WRLCK;
-		if (res.u.denied.type & 1)
-			request->fl_type = F_RDLCK;
-		request->fl_pid = 0;
-		status = 0;
+	switch (status) {
+		case 0:
+			request->fl_type = F_UNLCK;
+			break;
+		case -NFS4ERR_DENIED:
+			status = 0;
 	}
 out:
 	up_read(&clp->cl_sem);
@@ -2847,196 +3109,314 @@ static int do_vfs_lock(struct file *file, struct file_lock *fl)
 }
 
 struct nfs4_unlockdata {
-	struct nfs_lockargs arg;
-	struct nfs_locku_opargs luargs;
-	struct nfs_lockres res;
+	struct nfs_locku_args arg;
+	struct nfs_locku_res res;
 	struct nfs4_lock_state *lsp;
 	struct nfs_open_context *ctx;
-	atomic_t refcount;
-	struct completion completion;
+	struct file_lock fl;
+	const struct nfs_server *server;
+	unsigned long timestamp;
 };
 
-static void nfs4_locku_release_calldata(struct nfs4_unlockdata *calldata)
-{
-	if (atomic_dec_and_test(&calldata->refcount)) {
-		nfs_free_seqid(calldata->luargs.seqid);
-		nfs4_put_lock_state(calldata->lsp);
-		put_nfs_open_context(calldata->ctx);
-		kfree(calldata);
-	}
+static struct nfs4_unlockdata *nfs4_alloc_unlockdata(struct file_lock *fl,
+		struct nfs_open_context *ctx,
+		struct nfs4_lock_state *lsp,
+		struct nfs_seqid *seqid)
+{
+	struct nfs4_unlockdata *p;
+	struct inode *inode = lsp->ls_state->inode;
+
+	p = kmalloc(sizeof(*p), GFP_KERNEL);
+	if (p == NULL)
+		return NULL;
+	p->arg.fh = NFS_FH(inode);
+	p->arg.fl = &p->fl;
+	p->arg.seqid = seqid;
+	p->arg.stateid = &lsp->ls_stateid;
+	p->lsp = lsp;
+	atomic_inc(&lsp->ls_count);
+	/* Ensure we don't close file until we're done freeing locks! */
+	p->ctx = get_nfs_open_context(ctx);
+	memcpy(&p->fl, fl, sizeof(p->fl));
+	p->server = NFS_SERVER(inode);
+	return p;
 }
 
-static void nfs4_locku_complete(struct nfs4_unlockdata *calldata)
+static void nfs4_locku_release_calldata(void *data)
 {
-	complete(&calldata->completion);
-	nfs4_locku_release_calldata(calldata);
+	struct nfs4_unlockdata *calldata = data;
+	nfs_free_seqid(calldata->arg.seqid);
+	nfs4_put_lock_state(calldata->lsp);
+	put_nfs_open_context(calldata->ctx);
+	kfree(calldata);
 }
 
-static void nfs4_locku_done(struct rpc_task *task)
+static void nfs4_locku_done(struct rpc_task *task, void *data)
 {
-	struct nfs4_unlockdata *calldata = (struct nfs4_unlockdata *)task->tk_calldata;
+	struct nfs4_unlockdata *calldata = data;
 
-	nfs_increment_lock_seqid(task->tk_status, calldata->luargs.seqid);
+	if (RPC_ASSASSINATED(task))
+		return;
+	nfs_increment_lock_seqid(task->tk_status, calldata->arg.seqid);
 	switch (task->tk_status) {
 		case 0:
 			memcpy(calldata->lsp->ls_stateid.data,
-					calldata->res.u.stateid.data,
+					calldata->res.stateid.data,
 					sizeof(calldata->lsp->ls_stateid.data));
+			renew_lease(calldata->server, calldata->timestamp);
 			break;
 		case -NFS4ERR_STALE_STATEID:
 		case -NFS4ERR_EXPIRED:
-			nfs4_schedule_state_recovery(calldata->res.server->nfs4_state);
+			nfs4_schedule_state_recovery(calldata->server->nfs4_state);
 			break;
 		default:
-			if (nfs4_async_handle_error(task, calldata->res.server) == -EAGAIN) {
+			if (nfs4_async_handle_error(task, calldata->server) == -EAGAIN) {
 				rpc_restart_call(task);
-				return;
 			}
 	}
-	nfs4_locku_complete(calldata);
 }
 
-static void nfs4_locku_begin(struct rpc_task *task)
+static void nfs4_locku_prepare(struct rpc_task *task, void *data)
 {
-	struct nfs4_unlockdata *calldata = (struct nfs4_unlockdata *)task->tk_calldata;
+	struct nfs4_unlockdata *calldata = data;
 	struct rpc_message msg = {
 		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_LOCKU],
 		.rpc_argp       = &calldata->arg,
 		.rpc_resp       = &calldata->res,
 		.rpc_cred	= calldata->lsp->ls_state->owner->so_cred,
 	};
-	int status;
 
-	status = nfs_wait_on_sequence(calldata->luargs.seqid, task);
-	if (status != 0)
+	if (nfs_wait_on_sequence(calldata->arg.seqid, task) != 0)
 		return;
 	if ((calldata->lsp->ls_flags & NFS_LOCK_INITIALIZED) == 0) {
-		nfs4_locku_complete(calldata);
-		task->tk_exit = NULL;
-		rpc_exit(task, 0);
+		/* Note: exit _without_ running nfs4_locku_done */
+		task->tk_action = NULL;
 		return;
 	}
+	calldata->timestamp = jiffies;
 	rpc_call_setup(task, &msg, 0);
 }
 
+static const struct rpc_call_ops nfs4_locku_ops = {
+	.rpc_call_prepare = nfs4_locku_prepare,
+	.rpc_call_done = nfs4_locku_done,
+	.rpc_release = nfs4_locku_release_calldata,
+};
+
+static struct rpc_task *nfs4_do_unlck(struct file_lock *fl,
+		struct nfs_open_context *ctx,
+		struct nfs4_lock_state *lsp,
+		struct nfs_seqid *seqid)
+{
+	struct nfs4_unlockdata *data;
+	struct rpc_task *task;
+
+	data = nfs4_alloc_unlockdata(fl, ctx, lsp, seqid);
+	if (data == NULL) {
+		nfs_free_seqid(seqid);
+		return ERR_PTR(-ENOMEM);
+	}
+
+	/* Unlock _before_ we do the RPC call */
+	do_vfs_lock(fl->fl_file, fl);
+	task = rpc_run_task(NFS_CLIENT(lsp->ls_state->inode), RPC_TASK_ASYNC, &nfs4_locku_ops, data);
+	if (IS_ERR(task))
+		nfs4_locku_release_calldata(data);
+	return task;
+}
+
 static int nfs4_proc_unlck(struct nfs4_state *state, int cmd, struct file_lock *request)
 {
-	struct nfs4_unlockdata *calldata;
-	struct inode *inode = state->inode;
-	struct nfs_server *server = NFS_SERVER(inode);
+	struct nfs_seqid *seqid;
 	struct nfs4_lock_state *lsp;
-	int status;
+	struct rpc_task *task;
+	int status = 0;
 
 	/* Is this a delegated lock? */
 	if (test_bit(NFS_DELEGATED_STATE, &state->flags))
-		return do_vfs_lock(request->fl_file, request);
+		goto out_unlock;
+	/* Is this open_owner holding any locks on the server? */
+	if (test_bit(LK_STATE_IN_USE, &state->flags) == 0)
+		goto out_unlock;
 
 	status = nfs4_set_lock_state(state, request);
 	if (status != 0)
-		return status;
+		goto out_unlock;
 	lsp = request->fl_u.nfs4_fl.owner;
-	/* We might have lost the locks! */
-	if ((lsp->ls_flags & NFS_LOCK_INITIALIZED) == 0)
-		return 0;
-	calldata = kmalloc(sizeof(*calldata), GFP_KERNEL);
-	if (calldata == NULL)
-		return -ENOMEM;
-	calldata->luargs.seqid = nfs_alloc_seqid(&lsp->ls_seqid);
-	if (calldata->luargs.seqid == NULL) {
-		kfree(calldata);
-		return -ENOMEM;
-	}
-	calldata->luargs.stateid = &lsp->ls_stateid;
-	calldata->arg.fh = NFS_FH(inode);
-	calldata->arg.type = nfs4_lck_type(cmd, request);
-	calldata->arg.offset = request->fl_start;
-	calldata->arg.length = nfs4_lck_length(request);
-	calldata->arg.u.locku = &calldata->luargs;
-	calldata->res.server = server;
-	calldata->lsp = lsp;
-	atomic_inc(&lsp->ls_count);
-
-	/* Ensure we don't close file until we're done freeing locks! */
-	calldata->ctx = get_nfs_open_context((struct nfs_open_context*)request->fl_file->private_data);
-
-	atomic_set(&calldata->refcount, 2);
-	init_completion(&calldata->completion);
-
-	status = nfs4_call_async(NFS_SERVER(inode)->client, nfs4_locku_begin,
-			nfs4_locku_done, calldata);
-	if (status == 0)
-		wait_for_completion_interruptible(&calldata->completion);
+	status = -ENOMEM;
+	seqid = nfs_alloc_seqid(&lsp->ls_seqid);
+	if (seqid == NULL)
+		goto out_unlock;
+	task = nfs4_do_unlck(request, request->fl_file->private_data, lsp, seqid);
+	status = PTR_ERR(task);
+	if (IS_ERR(task))
+		goto out_unlock;
+	status = nfs4_wait_for_completion_rpc_task(task);
+	rpc_release_task(task);
+	return status;
+out_unlock:
 	do_vfs_lock(request->fl_file, request);
-	nfs4_locku_release_calldata(calldata);
 	return status;
 }
 
-static int _nfs4_do_setlk(struct nfs4_state *state, int cmd, struct file_lock *request, int reclaim)
+struct nfs4_lockdata {
+	struct nfs_lock_args arg;
+	struct nfs_lock_res res;
+	struct nfs4_lock_state *lsp;
+	struct nfs_open_context *ctx;
+	struct file_lock fl;
+	unsigned long timestamp;
+	int rpc_status;
+	int cancelled;
+};
+
+static struct nfs4_lockdata *nfs4_alloc_lockdata(struct file_lock *fl,
+		struct nfs_open_context *ctx, struct nfs4_lock_state *lsp)
 {
-	struct inode *inode = state->inode;
+	struct nfs4_lockdata *p;
+	struct inode *inode = lsp->ls_state->inode;
 	struct nfs_server *server = NFS_SERVER(inode);
-	struct nfs4_lock_state *lsp = request->fl_u.nfs4_fl.owner;
-	struct nfs_lock_opargs largs = {
-		.lock_stateid = &lsp->ls_stateid,
-		.open_stateid = &state->stateid,
-		.lock_owner = {
-			.clientid = server->nfs4_state->cl_clientid,
-			.id = lsp->ls_id,
-		},
-		.reclaim = reclaim,
-	};
-	struct nfs_lockargs arg = {
-		.fh = NFS_FH(inode),
-		.type = nfs4_lck_type(cmd, request),
-		.offset = request->fl_start,
-		.length = nfs4_lck_length(request),
-		.u = {
-			.lock = &largs,
-		},
-	};
-	struct nfs_lockres res = {
-		.server = server,
-	};
+
+	p = kzalloc(sizeof(*p), GFP_KERNEL);
+	if (p == NULL)
+		return NULL;
+
+	p->arg.fh = NFS_FH(inode);
+	p->arg.fl = &p->fl;
+	p->arg.lock_seqid = nfs_alloc_seqid(&lsp->ls_seqid);
+	if (p->arg.lock_seqid == NULL)
+		goto out_free;
+	p->arg.lock_stateid = &lsp->ls_stateid;
+	p->arg.lock_owner.clientid = server->nfs4_state->cl_clientid;
+	p->arg.lock_owner.id = lsp->ls_id;
+	p->lsp = lsp;
+	atomic_inc(&lsp->ls_count);
+	p->ctx = get_nfs_open_context(ctx);
+	memcpy(&p->fl, fl, sizeof(p->fl));
+	return p;
+out_free:
+	kfree(p);
+	return NULL;
+}
+
+static void nfs4_lock_prepare(struct rpc_task *task, void *calldata)
+{
+	struct nfs4_lockdata *data = calldata;
+	struct nfs4_state *state = data->lsp->ls_state;
+	struct nfs4_state_owner *sp = state->owner;
 	struct rpc_message msg = {
-		.rpc_proc	= &nfs4_procedures[NFSPROC4_CLNT_LOCK],
-		.rpc_argp       = &arg,
-		.rpc_resp       = &res,
-		.rpc_cred	= state->owner->so_cred,
+		.rpc_proc = &nfs4_procedures[NFSPROC4_CLNT_LOCK],
+		.rpc_argp = &data->arg,
+		.rpc_resp = &data->res,
+		.rpc_cred = sp->so_cred,
 	};
-	int status = -ENOMEM;
-
-	largs.lock_seqid = nfs_alloc_seqid(&lsp->ls_seqid);
-	if (largs.lock_seqid == NULL)
-		return -ENOMEM;
-	if (!(lsp->ls_seqid.flags & NFS_SEQID_CONFIRMED)) {
-		struct nfs4_state_owner *owner = state->owner;
 
-		largs.open_seqid = nfs_alloc_seqid(&owner->so_seqid);
-		if (largs.open_seqid == NULL)
+	if (nfs_wait_on_sequence(data->arg.lock_seqid, task) != 0)
+		return;
+	dprintk("%s: begin!\n", __FUNCTION__);
+	/* Do we need to do an open_to_lock_owner? */
+	if (!(data->arg.lock_seqid->sequence->flags & NFS_SEQID_CONFIRMED)) {
+		data->arg.open_seqid = nfs_alloc_seqid(&sp->so_seqid);
+		if (data->arg.open_seqid == NULL) {
+			data->rpc_status = -ENOMEM;
+			task->tk_action = NULL;
 			goto out;
-		largs.new_lock_owner = 1;
-		status = rpc_call_sync(server->client, &msg, RPC_TASK_NOINTR);
-		/* increment open seqid on success, and seqid mutating errors */
-		if (largs.new_lock_owner != 0) {
-			nfs_increment_open_seqid(status, largs.open_seqid);
-			if (status == 0)
-				nfs_confirm_seqid(&lsp->ls_seqid, 0);
 		}
-		nfs_free_seqid(largs.open_seqid);
-	} else
-		status = rpc_call_sync(server->client, &msg, RPC_TASK_NOINTR);
-	/* increment lock seqid on success, and seqid mutating errors*/
-	nfs_increment_lock_seqid(status, largs.lock_seqid);
-	/* save the returned stateid. */
-	if (status == 0) {
-		memcpy(lsp->ls_stateid.data, res.u.stateid.data,
-				sizeof(lsp->ls_stateid.data));
-		lsp->ls_flags |= NFS_LOCK_INITIALIZED;
-	} else if (status == -NFS4ERR_DENIED)
-		status = -EAGAIN;
+		data->arg.open_stateid = &state->stateid;
+		data->arg.new_lock_owner = 1;
+	}
+	data->timestamp = jiffies;
+	rpc_call_setup(task, &msg, 0);
 out:
-	nfs_free_seqid(largs.lock_seqid);
-	return status;
+	dprintk("%s: done!, ret = %d\n", __FUNCTION__, data->rpc_status);
+}
+
+static void nfs4_lock_done(struct rpc_task *task, void *calldata)
+{
+	struct nfs4_lockdata *data = calldata;
+
+	dprintk("%s: begin!\n", __FUNCTION__);
+
+	data->rpc_status = task->tk_status;
+	if (RPC_ASSASSINATED(task))
+		goto out;
+	if (data->arg.new_lock_owner != 0) {
+		nfs_increment_open_seqid(data->rpc_status, data->arg.open_seqid);
+		if (data->rpc_status == 0)
+			nfs_confirm_seqid(&data->lsp->ls_seqid, 0);
+		else
+			goto out;
+	}
+	if (data->rpc_status == 0) {
+		memcpy(data->lsp->ls_stateid.data, data->res.stateid.data,
+					sizeof(data->lsp->ls_stateid.data));
+		data->lsp->ls_flags |= NFS_LOCK_INITIALIZED;
+		renew_lease(NFS_SERVER(data->ctx->dentry->d_inode), data->timestamp);
+	}
+	nfs_increment_lock_seqid(data->rpc_status, data->arg.lock_seqid);
+out:
+	dprintk("%s: done, ret = %d!\n", __FUNCTION__, data->rpc_status);
+}
+
+static void nfs4_lock_release(void *calldata)
+{
+	struct nfs4_lockdata *data = calldata;
+
+	dprintk("%s: begin!\n", __FUNCTION__);
+	if (data->arg.open_seqid != NULL)
+		nfs_free_seqid(data->arg.open_seqid);
+	if (data->cancelled != 0) {
+		struct rpc_task *task;
+		task = nfs4_do_unlck(&data->fl, data->ctx, data->lsp,
+				data->arg.lock_seqid);
+		if (!IS_ERR(task))
+			rpc_release_task(task);
+		dprintk("%s: cancelling lock!\n", __FUNCTION__);
+	} else
+		nfs_free_seqid(data->arg.lock_seqid);
+	nfs4_put_lock_state(data->lsp);
+	put_nfs_open_context(data->ctx);
+	kfree(data);
+	dprintk("%s: done!\n", __FUNCTION__);
+}
+
+static const struct rpc_call_ops nfs4_lock_ops = {
+	.rpc_call_prepare = nfs4_lock_prepare,
+	.rpc_call_done = nfs4_lock_done,
+	.rpc_release = nfs4_lock_release,
+};
+
+static int _nfs4_do_setlk(struct nfs4_state *state, int cmd, struct file_lock *fl, int reclaim)
+{
+	struct nfs4_lockdata *data;
+	struct rpc_task *task;
+	int ret;
+
+	dprintk("%s: begin!\n", __FUNCTION__);
+	data = nfs4_alloc_lockdata(fl, fl->fl_file->private_data,
+			fl->fl_u.nfs4_fl.owner);
+	if (data == NULL)
+		return -ENOMEM;
+	if (IS_SETLKW(cmd))
+		data->arg.block = 1;
+	if (reclaim != 0)
+		data->arg.reclaim = 1;
+	task = rpc_run_task(NFS_CLIENT(state->inode), RPC_TASK_ASYNC,
+			&nfs4_lock_ops, data);
+	if (IS_ERR(task)) {
+		nfs4_lock_release(data);
+		return PTR_ERR(task);
+	}
+	ret = nfs4_wait_for_completion_rpc_task(task);
+	if (ret == 0) {
+		ret = data->rpc_status;
+		if (ret == -NFS4ERR_DENIED)
+			ret = -EAGAIN;
+	} else
+		data->cancelled = 1;
+	rpc_release_task(task);
+	dprintk("%s: done, ret = %d!\n", __FUNCTION__, ret);
+	return ret;
 }
 
 static int nfs4_lock_reclaim(struct nfs4_state *state, struct file_lock *request)
diff --git a/fs/nfs/nfs4renewd.c b/fs/nfs/nfs4renewd.c
index a3001628ad32..5d764d8e6d8a 100644
--- a/fs/nfs/nfs4renewd.c
+++ b/fs/nfs/nfs4renewd.c
@@ -54,6 +54,7 @@
 #include <linux/nfs4.h>
 #include <linux/nfs_fs.h>
 #include "nfs4_fs.h"
+#include "delegation.h"
 
 #define NFSDBG_FACILITY	NFSDBG_PROC
 
@@ -61,6 +62,7 @@ void
 nfs4_renew_state(void *data)
 {
 	struct nfs4_client *clp = (struct nfs4_client *)data;
+	struct rpc_cred *cred;
 	long lease, timeout;
 	unsigned long last, now;
 
@@ -68,7 +70,7 @@ nfs4_renew_state(void *data)
 	dprintk("%s: start\n", __FUNCTION__);
 	/* Are there any active superblocks? */
 	if (list_empty(&clp->cl_superblocks))
-		goto out; 
+		goto out;
 	spin_lock(&clp->cl_lock);
 	lease = clp->cl_lease_time;
 	last = clp->cl_last_renewal;
@@ -76,9 +78,17 @@ nfs4_renew_state(void *data)
 	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);
+		if (cred == NULL) {
+			set_bit(NFS4CLNT_LEASE_EXPIRED, &clp->cl_state);
+			spin_unlock(&clp->cl_lock);
+			nfs_expire_all_delegations(clp);
+			goto out;
+		}
 		spin_unlock(&clp->cl_lock);
 		/* Queue an asynchronous RENEW. */
-		nfs4_proc_async_renew(clp);
+		nfs4_proc_async_renew(clp, cred);
+		put_rpccred(cred);
 		timeout = (2 * lease) / 3;
 		spin_lock(&clp->cl_lock);
 	} else
diff --git a/fs/nfs/nfs4state.c b/fs/nfs/nfs4state.c
index 5ef4c57618fe..afad0255e7db 100644
--- a/fs/nfs/nfs4state.c
+++ b/fs/nfs/nfs4state.c
@@ -43,6 +43,8 @@
 #include <linux/smp_lock.h>
 #include <linux/nfs_fs.h>
 #include <linux/nfs_idmap.h>
+#include <linux/kthread.h>
+#include <linux/module.h>
 #include <linux/workqueue.h>
 #include <linux/bitops.h>
 
@@ -57,8 +59,6 @@ const nfs4_stateid zero_stateid;
 static DEFINE_SPINLOCK(state_spinlock);
 static LIST_HEAD(nfs4_clientid_list);
 
-static void nfs4_recover_state(void *);
-
 void
 init_nfsv4_state(struct nfs_server *server)
 {
@@ -91,11 +91,10 @@ nfs4_alloc_client(struct in_addr *addr)
 
 	if (nfs_callback_up() < 0)
 		return NULL;
-	if ((clp = kmalloc(sizeof(*clp), GFP_KERNEL)) == NULL) {
+	if ((clp = kzalloc(sizeof(*clp), GFP_KERNEL)) == NULL) {
 		nfs_callback_down();
 		return NULL;
 	}
-	memset(clp, 0, sizeof(*clp));
 	memcpy(&clp->cl_addr, addr, sizeof(clp->cl_addr));
 	init_rwsem(&clp->cl_sem);
 	INIT_LIST_HEAD(&clp->cl_delegations);
@@ -103,14 +102,12 @@ nfs4_alloc_client(struct in_addr *addr)
 	INIT_LIST_HEAD(&clp->cl_unused);
 	spin_lock_init(&clp->cl_lock);
 	atomic_set(&clp->cl_count, 1);
-	INIT_WORK(&clp->cl_recoverd, nfs4_recover_state, clp);
 	INIT_WORK(&clp->cl_renewd, nfs4_renew_state, clp);
 	INIT_LIST_HEAD(&clp->cl_superblocks);
-	init_waitqueue_head(&clp->cl_waitq);
 	rpc_init_wait_queue(&clp->cl_rpcwaitq, "NFS4 client");
 	clp->cl_rpcclient = ERR_PTR(-EINVAL);
 	clp->cl_boot_time = CURRENT_TIME;
-	clp->cl_state = 1 << NFS4CLNT_OK;
+	clp->cl_state = 1 << NFS4CLNT_LEASE_EXPIRED;
 	return clp;
 }
 
@@ -127,8 +124,6 @@ nfs4_free_client(struct nfs4_client *clp)
 		kfree(sp);
 	}
 	BUG_ON(!list_empty(&clp->cl_state_owners));
-	if (clp->cl_cred)
-		put_rpccred(clp->cl_cred);
 	nfs_idmap_delete(clp);
 	if (!IS_ERR(clp->cl_rpcclient))
 		rpc_shutdown_client(clp->cl_rpcclient);
@@ -193,27 +188,22 @@ nfs4_put_client(struct nfs4_client *clp)
 	list_del(&clp->cl_servers);
 	spin_unlock(&state_spinlock);
 	BUG_ON(!list_empty(&clp->cl_superblocks));
-	wake_up_all(&clp->cl_waitq);
 	rpc_wake_up(&clp->cl_rpcwaitq);
 	nfs4_kill_renewd(clp);
 	nfs4_free_client(clp);
 }
 
-static int __nfs4_init_client(struct nfs4_client *clp)
+static int nfs4_init_client(struct nfs4_client *clp, struct rpc_cred *cred)
 {
-	int status = nfs4_proc_setclientid(clp, NFS4_CALLBACK, nfs_callback_tcpport);
+	int status = nfs4_proc_setclientid(clp, NFS4_CALLBACK,
+			nfs_callback_tcpport, cred);
 	if (status == 0)
-		status = nfs4_proc_setclientid_confirm(clp);
+		status = nfs4_proc_setclientid_confirm(clp, cred);
 	if (status == 0)
 		nfs4_schedule_state_renewal(clp);
 	return status;
 }
 
-int nfs4_init_client(struct nfs4_client *clp)
-{
-	return nfs4_map_errors(__nfs4_init_client(clp));
-}
-
 u32
 nfs4_alloc_lockowner_id(struct nfs4_client *clp)
 {
@@ -235,6 +225,32 @@ nfs4_client_grab_unused(struct nfs4_client *clp, struct rpc_cred *cred)
 	return sp;
 }
 
+struct rpc_cred *nfs4_get_renew_cred(struct nfs4_client *clp)
+{
+	struct nfs4_state_owner *sp;
+	struct rpc_cred *cred = NULL;
+
+	list_for_each_entry(sp, &clp->cl_state_owners, so_list) {
+		if (list_empty(&sp->so_states))
+			continue;
+		cred = get_rpccred(sp->so_cred);
+		break;
+	}
+	return cred;
+}
+
+struct rpc_cred *nfs4_get_setclientid_cred(struct nfs4_client *clp)
+{
+	struct nfs4_state_owner *sp;
+
+	if (!list_empty(&clp->cl_state_owners)) {
+		sp = list_entry(clp->cl_state_owners.next,
+				struct nfs4_state_owner, so_list);
+		return get_rpccred(sp->so_cred);
+	}
+	return NULL;
+}
+
 static struct nfs4_state_owner *
 nfs4_find_state_owner(struct nfs4_client *clp, struct rpc_cred *cred)
 {
@@ -349,14 +365,9 @@ nfs4_alloc_open_state(void)
 {
 	struct nfs4_state *state;
 
-	state = kmalloc(sizeof(*state), GFP_KERNEL);
+	state = kzalloc(sizeof(*state), GFP_KERNEL);
 	if (!state)
 		return NULL;
-	state->state = 0;
-	state->nreaders = 0;
-	state->nwriters = 0;
-	state->flags = 0;
-	memset(state->stateid.data, 0, sizeof(state->stateid.data));
 	atomic_set(&state->count, 1);
 	INIT_LIST_HEAD(&state->lock_states);
 	spin_lock_init(&state->state_lock);
@@ -475,15 +486,23 @@ void nfs4_close_state(struct nfs4_state *state, mode_t mode)
 	/* Protect against nfs4_find_state() */
 	spin_lock(&owner->so_lock);
 	spin_lock(&inode->i_lock);
-	if (mode & FMODE_READ)
-		state->nreaders--;
-	if (mode & FMODE_WRITE)
-		state->nwriters--;
+	switch (mode & (FMODE_READ | FMODE_WRITE)) {
+		case FMODE_READ:
+			state->n_rdonly--;
+			break;
+		case FMODE_WRITE:
+			state->n_wronly--;
+			break;
+		case FMODE_READ|FMODE_WRITE:
+			state->n_rdwr--;
+	}
 	oldstate = newstate = state->state;
-	if (state->nreaders == 0)
-		newstate &= ~FMODE_READ;
-	if (state->nwriters == 0)
-		newstate &= ~FMODE_WRITE;
+	if (state->n_rdwr == 0) {
+		if (state->n_rdonly == 0)
+			newstate &= ~FMODE_READ;
+		if (state->n_wronly == 0)
+			newstate &= ~FMODE_WRITE;
+	}
 	if (test_bit(NFS_DELEGATED_STATE, &state->flags)) {
 		nfs4_state_set_mode_locked(state, newstate);
 		oldstate = newstate;
@@ -733,45 +752,43 @@ out:
 }
 
 static int reclaimer(void *);
-struct reclaimer_args {
-	struct nfs4_client *clp;
-	struct completion complete;
-};
+
+static inline void nfs4_clear_recover_bit(struct nfs4_client *clp)
+{
+	smp_mb__before_clear_bit();
+	clear_bit(NFS4CLNT_STATE_RECOVER, &clp->cl_state);
+	smp_mb__after_clear_bit();
+	wake_up_bit(&clp->cl_state, NFS4CLNT_STATE_RECOVER);
+	rpc_wake_up(&clp->cl_rpcwaitq);
+}
 
 /*
  * State recovery routine
  */
-void
-nfs4_recover_state(void *data)
+static void nfs4_recover_state(struct nfs4_client *clp)
 {
-	struct nfs4_client *clp = (struct nfs4_client *)data;
-	struct reclaimer_args args = {
-		.clp = clp,
-	};
-	might_sleep();
-
-	init_completion(&args.complete);
+	struct task_struct *task;
 
-	if (kernel_thread(reclaimer, &args, CLONE_KERNEL) < 0)
-		goto out_failed_clear;
-	wait_for_completion(&args.complete);
-	return;
-out_failed_clear:
-	set_bit(NFS4CLNT_OK, &clp->cl_state);
-	wake_up_all(&clp->cl_waitq);
-	rpc_wake_up(&clp->cl_rpcwaitq);
+	__module_get(THIS_MODULE);
+	atomic_inc(&clp->cl_count);
+	task = kthread_run(reclaimer, clp, "%u.%u.%u.%u-reclaim",
+			NIPQUAD(clp->cl_addr));
+	if (!IS_ERR(task))
+		return;
+	nfs4_clear_recover_bit(clp);
+	nfs4_put_client(clp);
+	module_put(THIS_MODULE);
 }
 
 /*
  * Schedule a state recovery attempt
  */
-void
-nfs4_schedule_state_recovery(struct nfs4_client *clp)
+void nfs4_schedule_state_recovery(struct nfs4_client *clp)
 {
 	if (!clp)
 		return;
-	if (test_and_clear_bit(NFS4CLNT_OK, &clp->cl_state))
-		schedule_work(&clp->cl_recoverd);
+	if (test_and_set_bit(NFS4CLNT_STATE_RECOVER, &clp->cl_state) == 0)
+		nfs4_recover_state(clp);
 }
 
 static int nfs4_reclaim_locks(struct nfs4_state_recovery_ops *ops, struct nfs4_state *state)
@@ -887,18 +904,14 @@ static void nfs4_state_mark_reclaim(struct nfs4_client *clp)
 
 static int reclaimer(void *ptr)
 {
-	struct reclaimer_args *args = (struct reclaimer_args *)ptr;
-	struct nfs4_client *clp = args->clp;
+	struct nfs4_client *clp = ptr;
 	struct nfs4_state_owner *sp;
 	struct nfs4_state_recovery_ops *ops;
+	struct rpc_cred *cred;
 	int status = 0;
 
-	daemonize("%u.%u.%u.%u-reclaim", NIPQUAD(clp->cl_addr));
 	allow_signal(SIGKILL);
 
-	atomic_inc(&clp->cl_count);
-	complete(&args->complete);
-
 	/* Ensure exclusive access to NFSv4 state */
 	lock_kernel();
 	down_write(&clp->cl_sem);
@@ -906,20 +919,33 @@ static int reclaimer(void *ptr)
 	if (list_empty(&clp->cl_superblocks))
 		goto out;
 restart_loop:
-	status = nfs4_proc_renew(clp);
-	switch (status) {
-		case 0:
-		case -NFS4ERR_CB_PATH_DOWN:
-			goto out;
-		case -NFS4ERR_STALE_CLIENTID:
-		case -NFS4ERR_LEASE_MOVED:
-			ops = &nfs4_reboot_recovery_ops;
-			break;
-		default:
-			ops = &nfs4_network_partition_recovery_ops;
-	};
+	ops = &nfs4_network_partition_recovery_ops;
+	/* Are there any open files on this volume? */
+	cred = nfs4_get_renew_cred(clp);
+	if (cred != NULL) {
+		/* Yes there are: try to renew the old lease */
+		status = nfs4_proc_renew(clp, cred);
+		switch (status) {
+			case 0:
+			case -NFS4ERR_CB_PATH_DOWN:
+				put_rpccred(cred);
+				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;
+		cred = nfs4_get_setclientid_cred(clp);
+	}
+	/* We're going to have to re-establish a clientid */
 	nfs4_state_mark_reclaim(clp);
-	status = __nfs4_init_client(clp);
+	status = -ENOENT;
+	if (cred != NULL) {
+		status = nfs4_init_client(clp, cred);
+		put_rpccred(cred);
+	}
 	if (status)
 		goto out_error;
 	/* Mark all delegations for reclaim */
@@ -940,14 +966,13 @@ restart_loop:
 	}
 	nfs_delegation_reap_unclaimed(clp);
 out:
-	set_bit(NFS4CLNT_OK, &clp->cl_state);
 	up_write(&clp->cl_sem);
 	unlock_kernel();
-	wake_up_all(&clp->cl_waitq);
-	rpc_wake_up(&clp->cl_rpcwaitq);
 	if (status == -NFS4ERR_CB_PATH_DOWN)
 		nfs_handle_cb_pathdown(clp);
+	nfs4_clear_recover_bit(clp);
 	nfs4_put_client(clp);
+	module_put_and_exit(0);
 	return 0;
 out_error:
 	printk(KERN_WARNING "Error: state recovery failed on NFSv4 server %u.%u.%u.%u with error %d\n",
diff --git a/fs/nfs/nfs4xdr.c b/fs/nfs/nfs4xdr.c
index fbbace8a30c4..4bbf5ef57785 100644
--- a/fs/nfs/nfs4xdr.c
+++ b/fs/nfs/nfs4xdr.c
@@ -392,9 +392,11 @@ static int nfs_stat_to_errno(int);
 				decode_getattr_maxsz)
 #define NFS4_enc_delegreturn_sz	(compound_encode_hdr_maxsz + \
 				encode_putfh_maxsz + \
-				encode_delegreturn_maxsz)
+				encode_delegreturn_maxsz + \
+				encode_getattr_maxsz)
 #define NFS4_dec_delegreturn_sz (compound_decode_hdr_maxsz + \
-				decode_delegreturn_maxsz)
+				decode_delegreturn_maxsz + \
+				decode_getattr_maxsz)
 #define NFS4_enc_getacl_sz	(compound_encode_hdr_maxsz + \
 				encode_putfh_maxsz + \
 				encode_getattr_maxsz)
@@ -564,7 +566,7 @@ static int encode_attrs(struct xdr_stream *xdr, const struct iattr *iap, const s
 	}
 	if (iap->ia_valid & ATTR_MODE) {
 		bmval1 |= FATTR4_WORD1_MODE;
-		WRITE32(iap->ia_mode);
+		WRITE32(iap->ia_mode & S_IALLUGO);
 	}
 	if (iap->ia_valid & ATTR_UID) {
 		bmval1 |= FATTR4_WORD1_OWNER;
@@ -742,69 +744,80 @@ static int encode_link(struct xdr_stream *xdr, const struct qstr *name)
 	return 0;
 }
 
+static inline int nfs4_lock_type(struct file_lock *fl, int block)
+{
+	if ((fl->fl_type & (F_RDLCK|F_WRLCK|F_UNLCK)) == F_RDLCK)
+		return block ? NFS4_READW_LT : NFS4_READ_LT;
+	return block ? NFS4_WRITEW_LT : NFS4_WRITE_LT;
+}
+
+static inline uint64_t nfs4_lock_length(struct file_lock *fl)
+{
+	if (fl->fl_end == OFFSET_MAX)
+		return ~(uint64_t)0;
+	return fl->fl_end - fl->fl_start + 1;
+}
+
 /*
  * opcode,type,reclaim,offset,length,new_lock_owner = 32
  * open_seqid,open_stateid,lock_seqid,lock_owner.clientid, lock_owner.id = 40
  */
-static int encode_lock(struct xdr_stream *xdr, const struct nfs_lockargs *arg)
+static int encode_lock(struct xdr_stream *xdr, const struct nfs_lock_args *args)
 {
 	uint32_t *p;
-	struct nfs_lock_opargs *opargs = arg->u.lock;
 
 	RESERVE_SPACE(32);
 	WRITE32(OP_LOCK);
-	WRITE32(arg->type); 
-	WRITE32(opargs->reclaim);
-	WRITE64(arg->offset);
-	WRITE64(arg->length);
-	WRITE32(opargs->new_lock_owner);
-	if (opargs->new_lock_owner){
+	WRITE32(nfs4_lock_type(args->fl, args->block));
+	WRITE32(args->reclaim);
+	WRITE64(args->fl->fl_start);
+	WRITE64(nfs4_lock_length(args->fl));
+	WRITE32(args->new_lock_owner);
+	if (args->new_lock_owner){
 		RESERVE_SPACE(40);
-		WRITE32(opargs->open_seqid->sequence->counter);
-		WRITEMEM(opargs->open_stateid->data, sizeof(opargs->open_stateid->data));
-		WRITE32(opargs->lock_seqid->sequence->counter);
-		WRITE64(opargs->lock_owner.clientid);
+		WRITE32(args->open_seqid->sequence->counter);
+		WRITEMEM(args->open_stateid->data, sizeof(args->open_stateid->data));
+		WRITE32(args->lock_seqid->sequence->counter);
+		WRITE64(args->lock_owner.clientid);
 		WRITE32(4);
-		WRITE32(opargs->lock_owner.id);
+		WRITE32(args->lock_owner.id);
 	}
 	else {
 		RESERVE_SPACE(20);
-		WRITEMEM(opargs->lock_stateid->data, sizeof(opargs->lock_stateid->data));
-		WRITE32(opargs->lock_seqid->sequence->counter);
+		WRITEMEM(args->lock_stateid->data, sizeof(args->lock_stateid->data));
+		WRITE32(args->lock_seqid->sequence->counter);
 	}
 
 	return 0;
 }
 
-static int encode_lockt(struct xdr_stream *xdr, const struct nfs_lockargs *arg)
+static int encode_lockt(struct xdr_stream *xdr, const struct nfs_lockt_args *args)
 {
 	uint32_t *p;
-	struct nfs_lowner *opargs = arg->u.lockt;
 
 	RESERVE_SPACE(40);
 	WRITE32(OP_LOCKT);
-	WRITE32(arg->type);
-	WRITE64(arg->offset);
-	WRITE64(arg->length);
-	WRITE64(opargs->clientid);
+	WRITE32(nfs4_lock_type(args->fl, 0));
+	WRITE64(args->fl->fl_start);
+	WRITE64(nfs4_lock_length(args->fl));
+	WRITE64(args->lock_owner.clientid);
 	WRITE32(4);
-	WRITE32(opargs->id);
+	WRITE32(args->lock_owner.id);
 
 	return 0;
 }
 
-static int encode_locku(struct xdr_stream *xdr, const struct nfs_lockargs *arg)
+static int encode_locku(struct xdr_stream *xdr, const struct nfs_locku_args *args)
 {
 	uint32_t *p;
-	struct nfs_locku_opargs *opargs = arg->u.locku;
 
 	RESERVE_SPACE(44);
 	WRITE32(OP_LOCKU);
-	WRITE32(arg->type);
-	WRITE32(opargs->seqid->sequence->counter);
-	WRITEMEM(opargs->stateid->data, sizeof(opargs->stateid->data));
-	WRITE64(arg->offset);
-	WRITE64(arg->length);
+	WRITE32(nfs4_lock_type(args->fl, 0));
+	WRITE32(args->seqid->sequence->counter);
+	WRITEMEM(args->stateid->data, sizeof(args->stateid->data));
+	WRITE64(args->fl->fl_start);
+	WRITE64(nfs4_lock_length(args->fl));
 
 	return 0;
 }
@@ -964,9 +977,9 @@ static int encode_open_confirm(struct xdr_stream *xdr, const struct nfs_open_con
 {
 	uint32_t *p;
 
-	RESERVE_SPACE(8+sizeof(arg->stateid.data));
+	RESERVE_SPACE(8+sizeof(arg->stateid->data));
 	WRITE32(OP_OPEN_CONFIRM);
-	WRITEMEM(arg->stateid.data, sizeof(arg->stateid.data));
+	WRITEMEM(arg->stateid->data, sizeof(arg->stateid->data));
 	WRITE32(arg->seqid->sequence->counter);
 
 	return 0;
@@ -1499,9 +1512,6 @@ static int nfs4_xdr_enc_open(struct rpc_rqst *req, uint32_t *p, struct nfs_opena
 	};
 	int status;
 
-	status = nfs_wait_on_sequence(args->seqid, req->rq_task);
-	if (status != 0)
-		goto out;
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
 	status = encode_putfh(&xdr, args->fh);
@@ -1538,9 +1548,6 @@ static int nfs4_xdr_enc_open_confirm(struct rpc_rqst *req, uint32_t *p, struct n
 	};
 	int status;
 
-	status = nfs_wait_on_sequence(args->seqid, req->rq_task);
-	if (status != 0)
-		goto out;
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
 	status = encode_putfh(&xdr, args->fh);
@@ -1558,19 +1565,19 @@ static int nfs4_xdr_enc_open_noattr(struct rpc_rqst *req, uint32_t *p, struct nf
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops   = 2,
+		.nops   = 3,
 	};
 	int status;
 
-	status = nfs_wait_on_sequence(args->seqid, req->rq_task);
-	if (status != 0)
-		goto out;
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
 	status = encode_putfh(&xdr, args->fh);
 	if (status)
 		goto out;
 	status = encode_open(&xdr, args);
+	if (status)
+		goto out;
+	status = encode_getfattr(&xdr, args->bitmask);
 out:
 	return status;
 }
@@ -1602,21 +1609,14 @@ out:
 /*
  * Encode a LOCK request
  */
-static int nfs4_xdr_enc_lock(struct rpc_rqst *req, uint32_t *p, struct nfs_lockargs *args)
+static int nfs4_xdr_enc_lock(struct rpc_rqst *req, uint32_t *p, struct nfs_lock_args *args)
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
 		.nops   = 2,
 	};
-	struct nfs_lock_opargs *opargs = args->u.lock;
 	int status;
 
-	status = nfs_wait_on_sequence(opargs->lock_seqid, req->rq_task);
-	if (status != 0)
-		goto out;
-	/* Do we need to do an open_to_lock_owner? */
-	if (opargs->lock_seqid->sequence->flags & NFS_SEQID_CONFIRMED)
-		opargs->new_lock_owner = 0;
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
 	status = encode_putfh(&xdr, args->fh);
@@ -1630,7 +1630,7 @@ out:
 /*
  * Encode a LOCKT request
  */
-static int nfs4_xdr_enc_lockt(struct rpc_rqst *req, uint32_t *p, struct nfs_lockargs *args)
+static int nfs4_xdr_enc_lockt(struct rpc_rqst *req, uint32_t *p, struct nfs_lockt_args *args)
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
@@ -1651,7 +1651,7 @@ out:
 /*
  * Encode a LOCKU request
  */
-static int nfs4_xdr_enc_locku(struct rpc_rqst *req, uint32_t *p, struct nfs_lockargs *args)
+static int nfs4_xdr_enc_locku(struct rpc_rqst *req, uint32_t *p, struct nfs_locku_args *args)
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
@@ -1985,14 +1985,20 @@ static int nfs4_xdr_enc_delegreturn(struct rpc_rqst *req, uint32_t *p, const str
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr = {
-		.nops = 2,
+		.nops = 3,
 	};
 	int status;
 
 	xdr_init_encode(&xdr, &req->rq_snd_buf, p);
 	encode_compound_hdr(&xdr, &hdr);
-	if ((status = encode_putfh(&xdr, args->fhandle)) == 0)
-		status = encode_delegreturn(&xdr, args->stateid);
+	status = encode_putfh(&xdr, args->fhandle);
+	if (status != 0)
+		goto out;
+	status = encode_delegreturn(&xdr, args->stateid);
+	if (status != 0)
+		goto out;
+	status = encode_getfattr(&xdr, args->bitmask);
+out:
 	return status;
 }
 
@@ -2955,55 +2961,64 @@ static int decode_link(struct xdr_stream *xdr, struct nfs4_change_info *cinfo)
 /*
  * We create the owner, so we know a proper owner.id length is 4.
  */
-static int decode_lock_denied (struct xdr_stream *xdr, struct nfs_lock_denied *denied)
+static int decode_lock_denied (struct xdr_stream *xdr, struct file_lock *fl)
 {
+	uint64_t offset, length, clientid;
 	uint32_t *p;
-	uint32_t namelen;
+	uint32_t namelen, type;
 
 	READ_BUF(32);
-	READ64(denied->offset);
-	READ64(denied->length);
-	READ32(denied->type);
-	READ64(denied->owner.clientid);
+	READ64(offset);
+	READ64(length);
+	READ32(type);
+	if (fl != NULL) {
+		fl->fl_start = (loff_t)offset;
+		fl->fl_end = fl->fl_start + (loff_t)length - 1;
+		if (length == ~(uint64_t)0)
+			fl->fl_end = OFFSET_MAX;
+		fl->fl_type = F_WRLCK;
+		if (type & 1)
+			fl->fl_type = F_RDLCK;
+		fl->fl_pid = 0;
+	}
+	READ64(clientid);
 	READ32(namelen);
 	READ_BUF(namelen);
-	if (namelen == 4)
-		READ32(denied->owner.id);
 	return -NFS4ERR_DENIED;
 }
 
-static int decode_lock(struct xdr_stream *xdr, struct nfs_lockres *res)
+static int decode_lock(struct xdr_stream *xdr, struct nfs_lock_res *res)
 {
 	uint32_t *p;
 	int status;
 
 	status = decode_op_hdr(xdr, OP_LOCK);
 	if (status == 0) {
-		READ_BUF(sizeof(res->u.stateid.data));
-		COPYMEM(res->u.stateid.data, sizeof(res->u.stateid.data));
+		READ_BUF(sizeof(res->stateid.data));
+		COPYMEM(res->stateid.data, sizeof(res->stateid.data));
 	} else if (status == -NFS4ERR_DENIED)
-		return decode_lock_denied(xdr, &res->u.denied);
+		return decode_lock_denied(xdr, NULL);
 	return status;
 }
 
-static int decode_lockt(struct xdr_stream *xdr, struct nfs_lockres *res)
+static int decode_lockt(struct xdr_stream *xdr, struct nfs_lockt_res *res)
 {
 	int status;
 	status = decode_op_hdr(xdr, OP_LOCKT);
 	if (status == -NFS4ERR_DENIED)
-		return decode_lock_denied(xdr, &res->u.denied);
+		return decode_lock_denied(xdr, res->denied);
 	return status;
 }
 
-static int decode_locku(struct xdr_stream *xdr, struct nfs_lockres *res)
+static int decode_locku(struct xdr_stream *xdr, struct nfs_locku_res *res)
 {
 	uint32_t *p;
 	int status;
 
 	status = decode_op_hdr(xdr, OP_LOCKU);
 	if (status == 0) {
-		READ_BUF(sizeof(res->u.stateid.data));
-		COPYMEM(res->u.stateid.data, sizeof(res->u.stateid.data));
+		READ_BUF(sizeof(res->stateid.data));
+		COPYMEM(res->stateid.data, sizeof(res->stateid.data));
 	}
 	return status;
 }
@@ -3831,6 +3846,9 @@ static int nfs4_xdr_dec_open_noattr(struct rpc_rqst *rqstp, uint32_t *p, struct
         if (status)
                 goto out;
         status = decode_open(&xdr, res);
+        if (status)
+                goto out;
+	decode_getfattr(&xdr, res->f_attr, res->server);
 out:
         return status;
 }
@@ -3864,7 +3882,7 @@ out:
 /*
  * Decode LOCK response
  */
-static int nfs4_xdr_dec_lock(struct rpc_rqst *rqstp, uint32_t *p, struct nfs_lockres *res)
+static int nfs4_xdr_dec_lock(struct rpc_rqst *rqstp, uint32_t *p, struct nfs_lock_res *res)
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr;
@@ -3885,7 +3903,7 @@ out:
 /*
  * Decode LOCKT response
  */
-static int nfs4_xdr_dec_lockt(struct rpc_rqst *rqstp, uint32_t *p, struct nfs_lockres *res)
+static int nfs4_xdr_dec_lockt(struct rpc_rqst *rqstp, uint32_t *p, struct nfs_lockt_res *res)
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr;
@@ -3906,7 +3924,7 @@ out:
 /*
  * Decode LOCKU response
  */
-static int nfs4_xdr_dec_locku(struct rpc_rqst *rqstp, uint32_t *p, struct nfs_lockres *res)
+static int nfs4_xdr_dec_locku(struct rpc_rqst *rqstp, uint32_t *p, struct nfs_locku_res *res)
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr;
@@ -4174,7 +4192,7 @@ static int nfs4_xdr_dec_setclientid_confirm(struct rpc_rqst *req, uint32_t *p, s
 /*
  * DELEGRETURN request
  */
-static int nfs4_xdr_dec_delegreturn(struct rpc_rqst *rqstp, uint32_t *p, void *dummy)
+static int nfs4_xdr_dec_delegreturn(struct rpc_rqst *rqstp, uint32_t *p, struct nfs4_delegreturnres *res)
 {
 	struct xdr_stream xdr;
 	struct compound_hdr hdr;
@@ -4182,11 +4200,14 @@ static int nfs4_xdr_dec_delegreturn(struct rpc_rqst *rqstp, uint32_t *p, void *d
 
 	xdr_init_decode(&xdr, &rqstp->rq_rcv_buf, p);
 	status = decode_compound_hdr(&xdr, &hdr);
-	if (status == 0) {
-		status = decode_putfh(&xdr);
-		if (status == 0)
-			status = decode_delegreturn(&xdr);
-	}
+	if (status != 0)
+		goto out;
+	status = decode_putfh(&xdr);
+	if (status != 0)
+		goto out;
+	status = decode_delegreturn(&xdr);
+	decode_getfattr(&xdr, res->fattr, res->server);
+out:
 	return status;
 }
 
diff --git a/fs/nfs/nfsroot.c b/fs/nfs/nfsroot.c
index 1b272a135a31..985cc53b8dd5 100644
--- a/fs/nfs/nfsroot.c
+++ b/fs/nfs/nfsroot.c
@@ -296,8 +296,8 @@ static int __init root_nfs_name(char *name)
 	nfs_port          = -1;
 	nfs_data.version  = NFS_MOUNT_VERSION;
 	nfs_data.flags    = NFS_MOUNT_NONLM;	/* No lockd in nfs root yet */
-	nfs_data.rsize    = NFS_DEF_FILE_IO_BUFFER_SIZE;
-	nfs_data.wsize    = NFS_DEF_FILE_IO_BUFFER_SIZE;
+	nfs_data.rsize    = NFS_DEF_FILE_IO_SIZE;
+	nfs_data.wsize    = NFS_DEF_FILE_IO_SIZE;
 	nfs_data.acregmin = 3;
 	nfs_data.acregmax = 60;
 	nfs_data.acdirmin = 30;
diff --git a/fs/nfs/proc.c b/fs/nfs/proc.c
index e1e3ca5d746b..f5150d71c03d 100644
--- a/fs/nfs/proc.c
+++ b/fs/nfs/proc.c
@@ -111,6 +111,9 @@ nfs_proc_setattr(struct dentry *dentry, struct nfs_fattr *fattr,
 	};
 	int	status;
 
+	/* Mask out the non-modebit related stuff from attr->ia_mode */
+	sattr->ia_mode &= S_IALLUGO;
+
 	dprintk("NFS call  setattr\n");
 	nfs_fattr_init(fattr);
 	status = rpc_call(NFS_CLIENT(inode), NFSPROC_SETATTR, &arg, fattr, 0);
@@ -547,10 +550,9 @@ nfs_proc_pathconf(struct nfs_server *server, struct nfs_fh *fhandle,
 
 extern u32 * nfs_decode_dirent(u32 *, struct nfs_entry *, int);
 
-static void
-nfs_read_done(struct rpc_task *task)
+static void nfs_read_done(struct rpc_task *task, void *calldata)
 {
-	struct nfs_read_data *data = (struct nfs_read_data *) task->tk_calldata;
+	struct nfs_read_data *data = calldata;
 
 	if (task->tk_status >= 0) {
 		nfs_refresh_inode(data->inode, data->res.fattr);
@@ -560,9 +562,14 @@ nfs_read_done(struct rpc_task *task)
 		if (data->args.offset + data->args.count >= data->res.fattr->size)
 			data->res.eof = 1;
 	}
-	nfs_readpage_result(task);
+	nfs_readpage_result(task, calldata);
 }
 
+static const struct rpc_call_ops nfs_read_ops = {
+	.rpc_call_done = nfs_read_done,
+	.rpc_release = nfs_readdata_release,
+};
+
 static void
 nfs_proc_read_setup(struct nfs_read_data *data)
 {
@@ -580,20 +587,24 @@ nfs_proc_read_setup(struct nfs_read_data *data)
 	flags = RPC_TASK_ASYNC | (IS_SWAPFILE(inode)? NFS_RPC_SWAPFLAGS : 0);
 
 	/* Finalize the task. */
-	rpc_init_task(task, NFS_CLIENT(inode), nfs_read_done, flags);
+	rpc_init_task(task, NFS_CLIENT(inode), flags, &nfs_read_ops, data);
 	rpc_call_setup(task, &msg, 0);
 }
 
-static void
-nfs_write_done(struct rpc_task *task)
+static void nfs_write_done(struct rpc_task *task, void *calldata)
 {
-	struct nfs_write_data *data = (struct nfs_write_data *) task->tk_calldata;
+	struct nfs_write_data *data = calldata;
 
 	if (task->tk_status >= 0)
 		nfs_post_op_update_inode(data->inode, data->res.fattr);
-	nfs_writeback_done(task);
+	nfs_writeback_done(task, calldata);
 }
 
+static const struct rpc_call_ops nfs_write_ops = {
+	.rpc_call_done = nfs_write_done,
+	.rpc_release = nfs_writedata_release,
+};
+
 static void
 nfs_proc_write_setup(struct nfs_write_data *data, int how)
 {
@@ -614,7 +625,7 @@ nfs_proc_write_setup(struct nfs_write_data *data, int how)
 	flags = (how & FLUSH_SYNC) ? 0 : RPC_TASK_ASYNC;
 
 	/* Finalize the task. */
-	rpc_init_task(task, NFS_CLIENT(inode), nfs_write_done, flags);
+	rpc_init_task(task, NFS_CLIENT(inode), flags, &nfs_write_ops, data);
 	rpc_call_setup(task, &msg, 0);
 }
 
diff --git a/fs/nfs/read.c b/fs/nfs/read.c
index 5f20eafba8ec..05eb43fadf8e 100644
--- a/fs/nfs/read.c
+++ b/fs/nfs/read.c
@@ -42,9 +42,8 @@ mempool_t *nfs_rdata_mempool;
 
 #define MIN_POOL_READ	(32)
 
-void nfs_readdata_release(struct rpc_task *task)
+void nfs_readdata_release(void *data)
 {
-        struct nfs_read_data   *data = (struct nfs_read_data *)task->tk_calldata;
         nfs_readdata_free(data);
 }
 
@@ -84,7 +83,7 @@ static int nfs_readpage_sync(struct nfs_open_context *ctx, struct inode *inode,
 	int		result;
 	struct nfs_read_data *rdata;
 
-	rdata = nfs_readdata_alloc();
+	rdata = nfs_readdata_alloc(1);
 	if (!rdata)
 		return -ENOMEM;
 
@@ -220,9 +219,6 @@ static void nfs_read_rpcsetup(struct nfs_page *req, struct nfs_read_data *data,
 	NFS_PROTO(inode)->read_setup(data);
 
 	data->task.tk_cookie = (unsigned long)inode;
-	data->task.tk_calldata = data;
-	/* Release requests */
-	data->task.tk_release = nfs_readdata_release;
 
 	dprintk("NFS: %4d initiated read call (req %s/%Ld, %u bytes @ offset %Lu)\n",
 			data->task.tk_pid,
@@ -287,7 +283,7 @@ static int nfs_pagein_multi(struct list_head *head, struct inode *inode)
 
 	nbytes = req->wb_bytes;
 	for(;;) {
-		data = nfs_readdata_alloc();
+		data = nfs_readdata_alloc(1);
 		if (!data)
 			goto out_bad;
 		INIT_LIST_HEAD(&data->pages);
@@ -343,7 +339,7 @@ static int nfs_pagein_one(struct list_head *head, struct inode *inode)
 	if (NFS_SERVER(inode)->rsize < PAGE_CACHE_SIZE)
 		return nfs_pagein_multi(head, inode);
 
-	data = nfs_readdata_alloc();
+	data = nfs_readdata_alloc(NFS_SERVER(inode)->rpages);
 	if (!data)
 		goto out_bad;
 
@@ -452,9 +448,9 @@ static void nfs_readpage_result_full(struct nfs_read_data *data, int status)
  * This is the callback from RPC telling us whether a reply was
  * received or some error occurred (timeout or socket shutdown).
  */
-void nfs_readpage_result(struct rpc_task *task)
+void nfs_readpage_result(struct rpc_task *task, void *calldata)
 {
-	struct nfs_read_data *data = (struct nfs_read_data *)task->tk_calldata;
+	struct nfs_read_data *data = calldata;
 	struct nfs_readargs *argp = &data->args;
 	struct nfs_readres *resp = &data->res;
 	int status = task->tk_status;
diff --git a/fs/nfs/sysctl.c b/fs/nfs/sysctl.c
new file mode 100644
index 000000000000..4c486eb867ca
--- /dev/null
+++ b/fs/nfs/sysctl.c
@@ -0,0 +1,84 @@
+/*
+ * linux/fs/nfs/sysctl.c
+ *
+ * Sysctl interface to NFS parameters
+ */
+#include <linux/config.h>
+#include <linux/types.h>
+#include <linux/linkage.h>
+#include <linux/ctype.h>
+#include <linux/fs.h>
+#include <linux/sysctl.h>
+#include <linux/module.h>
+#include <linux/nfs4.h>
+#include <linux/nfs_idmap.h>
+
+#include "callback.h"
+
+static const int nfs_set_port_min = 0;
+static const int nfs_set_port_max = 65535;
+static struct ctl_table_header *nfs_callback_sysctl_table;
+/*
+ * Something that isn't CTL_ANY, CTL_NONE or a value that may clash.
+ * Use the same values as fs/lockd/svc.c
+ */
+#define CTL_UNNUMBERED -2
+
+static ctl_table nfs_cb_sysctls[] = {
+#ifdef CONFIG_NFS_V4
+	{
+		.ctl_name = CTL_UNNUMBERED,
+		.procname = "nfs_callback_tcpport",
+		.data = &nfs_callback_set_tcpport,
+		.maxlen = sizeof(int),
+		.mode = 0644,
+		.proc_handler = &proc_dointvec_minmax,
+		.extra1 = (int *)&nfs_set_port_min,
+		.extra2 = (int *)&nfs_set_port_max,
+	},
+	{
+		.ctl_name = CTL_UNNUMBERED,
+		.procname = "idmap_cache_timeout",
+		.data = &nfs_idmap_cache_timeout,
+		.maxlen = sizeof(int),
+		.mode = 0644,
+		.proc_handler = &proc_dointvec_jiffies,
+		.strategy = &sysctl_jiffies,
+	},
+#endif
+	{ .ctl_name = 0 }
+};
+
+static ctl_table nfs_cb_sysctl_dir[] = {
+	{
+		.ctl_name = CTL_UNNUMBERED,
+		.procname = "nfs",
+		.mode = 0555,
+		.child = nfs_cb_sysctls,
+	},
+	{ .ctl_name = 0 }
+};
+
+static ctl_table nfs_cb_sysctl_root[] = {
+	{
+		.ctl_name = CTL_FS,
+		.procname = "fs",
+		.mode = 0555,
+		.child = nfs_cb_sysctl_dir,
+	},
+	{ .ctl_name = 0 }
+};
+
+int nfs_register_sysctl(void)
+{
+	nfs_callback_sysctl_table = register_sysctl_table(nfs_cb_sysctl_root, 0);
+	if (nfs_callback_sysctl_table == NULL)
+		return -ENOMEM;
+	return 0;
+}
+
+void nfs_unregister_sysctl(void)
+{
+	unregister_sysctl_table(nfs_callback_sysctl_table);
+	nfs_callback_sysctl_table = NULL;
+}
diff --git a/fs/nfs/unlink.c b/fs/nfs/unlink.c
index d639d172d568..a65c7b53d558 100644
--- a/fs/nfs/unlink.c
+++ b/fs/nfs/unlink.c
@@ -87,10 +87,9 @@ nfs_copy_dname(struct dentry *dentry, struct nfs_unlinkdata *data)
  * We delay initializing RPC info until after the call to dentry_iput()
  * in order to minimize races against rename().
  */
-static void
-nfs_async_unlink_init(struct rpc_task *task)
+static void nfs_async_unlink_init(struct rpc_task *task, void *calldata)
 {
-	struct nfs_unlinkdata	*data = (struct nfs_unlinkdata *)task->tk_calldata;
+	struct nfs_unlinkdata	*data = calldata;
 	struct dentry		*dir = data->dir;
 	struct rpc_message	msg = {
 		.rpc_cred	= data->cred,
@@ -116,10 +115,9 @@ nfs_async_unlink_init(struct rpc_task *task)
  *
  * Do the directory attribute update.
  */
-static void
-nfs_async_unlink_done(struct rpc_task *task)
+static void nfs_async_unlink_done(struct rpc_task *task, void *calldata)
 {
-	struct nfs_unlinkdata	*data = (struct nfs_unlinkdata *)task->tk_calldata;
+	struct nfs_unlinkdata	*data = calldata;
 	struct dentry		*dir = data->dir;
 	struct inode		*dir_i;
 
@@ -141,13 +139,18 @@ nfs_async_unlink_done(struct rpc_task *task)
  * We need to call nfs_put_unlinkdata as a 'tk_release' task since the
  * rpc_task would be freed too.
  */
-static void
-nfs_async_unlink_release(struct rpc_task *task)
+static void nfs_async_unlink_release(void *calldata)
 {
-	struct nfs_unlinkdata	*data = (struct nfs_unlinkdata *)task->tk_calldata;
+	struct nfs_unlinkdata	*data = calldata;
 	nfs_put_unlinkdata(data);
 }
 
+static const struct rpc_call_ops nfs_unlink_ops = {
+	.rpc_call_prepare = nfs_async_unlink_init,
+	.rpc_call_done = nfs_async_unlink_done,
+	.rpc_release = nfs_async_unlink_release,
+};
+
 /**
  * nfs_async_unlink - asynchronous unlinking of a file
  * @dentry: dentry to unlink
@@ -157,7 +160,6 @@ nfs_async_unlink(struct dentry *dentry)
 {
 	struct dentry	*dir = dentry->d_parent;
 	struct nfs_unlinkdata	*data;
-	struct rpc_task	*task;
 	struct rpc_clnt	*clnt = NFS_CLIENT(dir->d_inode);
 	int		status = -ENOMEM;
 
@@ -178,17 +180,13 @@ nfs_async_unlink(struct dentry *dentry)
 	nfs_deletes = data;
 	data->count = 1;
 
-	task = &data->task;
-	rpc_init_task(task, clnt, nfs_async_unlink_done , RPC_TASK_ASYNC);
-	task->tk_calldata = data;
-	task->tk_action = nfs_async_unlink_init;
-	task->tk_release = nfs_async_unlink_release;
+	rpc_init_task(&data->task, clnt, RPC_TASK_ASYNC, &nfs_unlink_ops, data);
 
 	spin_lock(&dentry->d_lock);
 	dentry->d_flags |= DCACHE_NFSFS_RENAMED;
 	spin_unlock(&dentry->d_lock);
 
-	rpc_sleep_on(&nfs_delete_queue, task, NULL, NULL);
+	rpc_sleep_on(&nfs_delete_queue, &data->task, NULL, NULL);
 	status = 0;
  out:
 	return status;
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index 3107908e5f3f..9449b6835509 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -89,24 +89,38 @@ static mempool_t *nfs_commit_mempool;
 
 static DECLARE_WAIT_QUEUE_HEAD(nfs_write_congestion);
 
-static inline struct nfs_write_data *nfs_commit_alloc(void)
+static inline struct nfs_write_data *nfs_commit_alloc(unsigned int pagecount)
 {
 	struct nfs_write_data *p = mempool_alloc(nfs_commit_mempool, SLAB_NOFS);
+
 	if (p) {
 		memset(p, 0, sizeof(*p));
 		INIT_LIST_HEAD(&p->pages);
+		if (pagecount < NFS_PAGEVEC_SIZE)
+			p->pagevec = &p->page_array[0];
+		else {
+			size_t size = ++pagecount * sizeof(struct page *);
+			p->pagevec = kmalloc(size, GFP_NOFS);
+			if (p->pagevec) {
+				memset(p->pagevec, 0, size);
+			} else {
+				mempool_free(p, nfs_commit_mempool);
+				p = NULL;
+			}
+		}
 	}
 	return p;
 }
 
 static inline void nfs_commit_free(struct nfs_write_data *p)
 {
+	if (p && (p->pagevec != &p->page_array[0]))
+		kfree(p->pagevec);
 	mempool_free(p, nfs_commit_mempool);
 }
 
-static void nfs_writedata_release(struct rpc_task *task)
+void nfs_writedata_release(void *wdata)
 {
-	struct nfs_write_data	*wdata = (struct nfs_write_data *)task->tk_calldata;
 	nfs_writedata_free(wdata);
 }
 
@@ -168,7 +182,7 @@ static int nfs_writepage_sync(struct nfs_open_context *ctx, struct inode *inode,
 	int		result, written = 0;
 	struct nfs_write_data *wdata;
 
-	wdata = nfs_writedata_alloc();
+	wdata = nfs_writedata_alloc(1);
 	if (!wdata)
 		return -ENOMEM;
 
@@ -232,19 +246,16 @@ static int nfs_writepage_async(struct nfs_open_context *ctx,
 		unsigned int offset, unsigned int count)
 {
 	struct nfs_page	*req;
-	int		status;
 
 	req = nfs_update_request(ctx, inode, page, offset, count);
-	status = (IS_ERR(req)) ? PTR_ERR(req) : 0;
-	if (status < 0)
-		goto out;
+	if (IS_ERR(req))
+		return PTR_ERR(req);
 	/* Update file length */
 	nfs_grow_file(page, offset, count);
 	/* Set the PG_uptodate flag? */
 	nfs_mark_uptodate(page, offset, count);
 	nfs_unlock_request(req);
- out:
-	return status;
+	return 0;
 }
 
 static int wb_priority(struct writeback_control *wbc)
@@ -304,11 +315,8 @@ do_it:
 	lock_kernel();
 	if (!IS_SYNC(inode) && inode_referenced) {
 		err = nfs_writepage_async(ctx, inode, page, 0, offset);
-		if (err >= 0) {
-			err = 0;
-			if (wbc->for_reclaim)
-				nfs_flush_inode(inode, 0, 0, FLUSH_STABLE);
-		}
+		if (!wbc->for_writepages)
+			nfs_flush_inode(inode, 0, 0, wb_priority(wbc));
 	} else {
 		err = nfs_writepage_sync(ctx, inode, page, 0,
 						offset, priority);
@@ -877,9 +885,6 @@ static void nfs_write_rpcsetup(struct nfs_page *req,
 
 	data->task.tk_priority = flush_task_priority(how);
 	data->task.tk_cookie = (unsigned long)inode;
-	data->task.tk_calldata = data;
-	/* Release requests */
-	data->task.tk_release = nfs_writedata_release;
 
 	dprintk("NFS: %4d initiated write call (req %s/%Ld, %u bytes @ offset %Lu)\n",
 		data->task.tk_pid,
@@ -919,7 +924,7 @@ static int nfs_flush_multi(struct list_head *head, struct inode *inode, int how)
 
 	nbytes = req->wb_bytes;
 	for (;;) {
-		data = nfs_writedata_alloc();
+		data = nfs_writedata_alloc(1);
 		if (!data)
 			goto out_bad;
 		list_add(&data->pages, &list);
@@ -983,7 +988,7 @@ static int nfs_flush_one(struct list_head *head, struct inode *inode, int how)
 	if (NFS_SERVER(inode)->wsize < PAGE_CACHE_SIZE)
 		return nfs_flush_multi(head, inode, how);
 
-	data = nfs_writedata_alloc();
+	data = nfs_writedata_alloc(NFS_SERVER(inode)->wpages);
 	if (!data)
 		goto out_bad;
 
@@ -1137,9 +1142,9 @@ static void nfs_writeback_done_full(struct nfs_write_data *data, int status)
 /*
  * This function is called when the WRITE call is complete.
  */
-void nfs_writeback_done(struct rpc_task *task)
+void nfs_writeback_done(struct rpc_task *task, void *calldata)
 {
-	struct nfs_write_data	*data = (struct nfs_write_data *) task->tk_calldata;
+	struct nfs_write_data	*data = calldata;
 	struct nfs_writeargs	*argp = &data->args;
 	struct nfs_writeres	*resp = &data->res;
 
@@ -1206,9 +1211,8 @@ void nfs_writeback_done(struct rpc_task *task)
 
 
 #if defined(CONFIG_NFS_V3) || defined(CONFIG_NFS_V4)
-static void nfs_commit_release(struct rpc_task *task)
+void nfs_commit_release(void *wdata)
 {
-	struct nfs_write_data	*wdata = (struct nfs_write_data *)task->tk_calldata;
 	nfs_commit_free(wdata);
 }
 
@@ -1244,9 +1248,6 @@ static void nfs_commit_rpcsetup(struct list_head *head,
 
 	data->task.tk_priority = flush_task_priority(how);
 	data->task.tk_cookie = (unsigned long)inode;
-	data->task.tk_calldata = data;
-	/* Release requests */
-	data->task.tk_release = nfs_commit_release;
 	
 	dprintk("NFS: %4d initiated commit call\n", data->task.tk_pid);
 }
@@ -1255,12 +1256,12 @@ static void nfs_commit_rpcsetup(struct list_head *head,
  * Commit dirty pages
  */
 static int
-nfs_commit_list(struct list_head *head, int how)
+nfs_commit_list(struct inode *inode, struct list_head *head, int how)
 {
 	struct nfs_write_data	*data;
 	struct nfs_page         *req;
 
-	data = nfs_commit_alloc();
+	data = nfs_commit_alloc(NFS_SERVER(inode)->wpages);
 
 	if (!data)
 		goto out_bad;
@@ -1283,10 +1284,9 @@ nfs_commit_list(struct list_head *head, int how)
 /*
  * COMMIT call returned
  */
-void
-nfs_commit_done(struct rpc_task *task)
+void nfs_commit_done(struct rpc_task *task, void *calldata)
 {
-	struct nfs_write_data	*data = (struct nfs_write_data *)task->tk_calldata;
+	struct nfs_write_data	*data = calldata;
 	struct nfs_page		*req;
 	int res = 0;
 
@@ -1366,7 +1366,7 @@ int nfs_commit_inode(struct inode *inode, int how)
 	res = nfs_scan_commit(inode, &head, 0, 0);
 	spin_unlock(&nfsi->req_lock);
 	if (res) {
-		error = nfs_commit_list(&head, how);
+		error = nfs_commit_list(inode, &head, how);
 		if (error < 0)
 			return error;
 	}
@@ -1377,22 +1377,23 @@ int nfs_commit_inode(struct inode *inode, int how)
 int nfs_sync_inode(struct inode *inode, unsigned long idx_start,
 		  unsigned int npages, int how)
 {
-	int	error,
-		wait;
+	int nocommit = how & FLUSH_NOCOMMIT;
+	int wait = how & FLUSH_WAIT;
+	int error;
 
-	wait = how & FLUSH_WAIT;
-	how &= ~FLUSH_WAIT;
+	how &= ~(FLUSH_WAIT|FLUSH_NOCOMMIT);
 
 	do {
-		error = 0;
-		if (wait)
+		if (wait) {
 			error = nfs_wait_on_requests(inode, idx_start, npages);
-		if (error == 0)
-			error = nfs_flush_inode(inode, idx_start, npages, how);
-#if defined(CONFIG_NFS_V3) || defined(CONFIG_NFS_V4)
-		if (error == 0)
+			if (error != 0)
+				continue;
+		}
+		error = nfs_flush_inode(inode, idx_start, npages, how);
+		if (error != 0)
+			continue;
+		if (!nocommit)
 			error = nfs_commit_inode(inode, how);
-#endif
 	} while (error > 0);
 	return error;
 }