Merge branch 'nfs-for-2.6.37' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6

* 'nfs-for-2.6.37' of git://git.linux-nfs.org/projects/trondmy/nfs-2.6: (67 commits)
  SUNRPC: Cleanup duplicate assignment in rpcauth_refreshcred
  nfs: fix unchecked value
  Ask for time_delta during fsinfo probe
  Revalidate caches on lock
  SUNRPC: After calling xprt_release(), we must restart from call_reserve
  NFSv4: Fix up the 'dircount' hint in encode_readdir
  NFSv4: Clean up nfs4_decode_dirent
  NFSv4: nfs4_decode_dirent must clear entry->fattr->valid
  NFSv4: Fix a regression in decode_getfattr
  NFSv4: Fix up decode_attr_filehandle() to handle the case of empty fh pointer
  NFS: Ensure we check all allocation return values in new readdir code
  NFS: Readdir plus in v4
  NFS: introduce generic decode_getattr function
  NFS: check xdr_decode for errors
  NFS: nfs_readdir_filler catch all errors
  NFS: readdir with vmapped pages
  NFS: remove page size checking code
  NFS: decode_dirent should use an xdr_stream
  SUNRPC: Add a helper function xdr_inline_peek
  NFS: remove readdir plus limit
  ...

24 files changed, 2000 insertions, 1392 deletions

diff --git a/fs/lockd/clntlock.c b/fs/lockd/clntlock.c
index 64fd427c993..d5bb86866e6 100644
--- a/fs/lockd/clntlock.c
+++ b/fs/lockd/clntlock.c
@@ -42,6 +42,7 @@ struct nlm_wait {
 };
 
 static LIST_HEAD(nlm_blocked);
+static DEFINE_SPINLOCK(nlm_blocked_lock);
 
 /**
  * nlmclnt_init - Set up per-NFS mount point lockd data structures
@@ -97,7 +98,10 @@ struct nlm_wait *nlmclnt_prepare_block(struct nlm_host *host, struct file_lock *
 		block->b_lock = fl;
 		init_waitqueue_head(&block->b_wait);
 		block->b_status = nlm_lck_blocked;
+
+		spin_lock(&nlm_blocked_lock);
 		list_add(&block->b_list, &nlm_blocked);
+		spin_unlock(&nlm_blocked_lock);
 	}
 	return block;
 }
@@ -106,7 +110,9 @@ void nlmclnt_finish_block(struct nlm_wait *block)
 {
 	if (block == NULL)
 		return;
+	spin_lock(&nlm_blocked_lock);
 	list_del(&block->b_list);
+	spin_unlock(&nlm_blocked_lock);
 	kfree(block);
 }
 
@@ -154,6 +160,7 @@ __be32 nlmclnt_grant(const struct sockaddr *addr, const struct nlm_lock *lock)
 	 * Look up blocked request based on arguments. 
 	 * Warning: must not use cookie to match it!
 	 */
+	spin_lock(&nlm_blocked_lock);
 	list_for_each_entry(block, &nlm_blocked, b_list) {
 		struct file_lock *fl_blocked = block->b_lock;
 
@@ -178,6 +185,7 @@ __be32 nlmclnt_grant(const struct sockaddr *addr, const struct nlm_lock *lock)
 		wake_up(&block->b_wait);
 		res = nlm_granted;
 	}
+	spin_unlock(&nlm_blocked_lock);
 	return res;
 }
 
@@ -216,10 +224,6 @@ reclaimer(void *ptr)
 	allow_signal(SIGKILL);
 
 	down_write(&host->h_rwsem);
-
-	/* This one ensures that our parent doesn't terminate while the
-	 * reclaim is in progress */
-	lock_kernel();
 	lockd_up();	/* note: this cannot fail as lockd is already running */
 
 	dprintk("lockd: reclaiming locks for host %s\n", host->h_name);
@@ -260,16 +264,17 @@ restart:
 	dprintk("NLM: done reclaiming locks for host %s\n", host->h_name);
 
 	/* Now, wake up all processes that sleep on a blocked lock */
+	spin_lock(&nlm_blocked_lock);
 	list_for_each_entry(block, &nlm_blocked, b_list) {
 		if (block->b_host == host) {
 			block->b_status = nlm_lck_denied_grace_period;
 			wake_up(&block->b_wait);
 		}
 	}
+	spin_unlock(&nlm_blocked_lock);
 
 	/* Release host handle after use */
 	nlm_release_host(host);
 	lockd_down();
-	unlock_kernel();
 	return 0;
 }
diff --git a/fs/lockd/clntproc.c b/fs/lockd/clntproc.c
index 7932c399fab..47ea1e1925b 100644
--- a/fs/lockd/clntproc.c
+++ b/fs/lockd/clntproc.c
@@ -166,7 +166,6 @@ int nlmclnt_proc(struct nlm_host *host, int cmd, struct file_lock *fl)
 	/* Set up the argument struct */
 	nlmclnt_setlockargs(call, fl);
 
-	lock_kernel();
 	if (IS_SETLK(cmd) || IS_SETLKW(cmd)) {
 		if (fl->fl_type != F_UNLCK) {
 			call->a_args.block = IS_SETLKW(cmd) ? 1 : 0;
@@ -177,10 +176,8 @@ int nlmclnt_proc(struct nlm_host *host, int cmd, struct file_lock *fl)
 		status = nlmclnt_test(call, fl);
 	else
 		status = -EINVAL;
-
 	fl->fl_ops->fl_release_private(fl);
 	fl->fl_ops = NULL;
-	unlock_kernel();
 
 	dprintk("lockd: clnt proc returns %d\n", status);
 	return status;
@@ -226,9 +223,7 @@ void nlm_release_call(struct nlm_rqst *call)
 
 static void nlmclnt_rpc_release(void *data)
 {
-	lock_kernel();
 	nlm_release_call(data);
-	unlock_kernel();
 }
 
 static int nlm_wait_on_grace(wait_queue_head_t *queue)
@@ -448,14 +443,18 @@ out:
 
 static void nlmclnt_locks_copy_lock(struct file_lock *new, struct file_lock *fl)
 {
+	spin_lock(&fl->fl_u.nfs_fl.owner->host->h_lock);
 	new->fl_u.nfs_fl.state = fl->fl_u.nfs_fl.state;
 	new->fl_u.nfs_fl.owner = nlm_get_lockowner(fl->fl_u.nfs_fl.owner);
 	list_add_tail(&new->fl_u.nfs_fl.list, &fl->fl_u.nfs_fl.owner->host->h_granted);
+	spin_unlock(&fl->fl_u.nfs_fl.owner->host->h_lock);
 }
 
 static void nlmclnt_locks_release_private(struct file_lock *fl)
 {
+	spin_lock(&fl->fl_u.nfs_fl.owner->host->h_lock);
 	list_del(&fl->fl_u.nfs_fl.list);
+	spin_unlock(&fl->fl_u.nfs_fl.owner->host->h_lock);
 	nlm_put_lockowner(fl->fl_u.nfs_fl.owner);
 }
 
@@ -721,9 +720,7 @@ static void nlmclnt_unlock_callback(struct rpc_task *task, void *data)
 die:
 	return;
  retry_rebind:
-	lock_kernel();
 	nlm_rebind_host(req->a_host);
-	unlock_kernel();
  retry_unlock:
 	rpc_restart_call(task);
 }
@@ -801,9 +798,7 @@ retry_cancel:
 	/* Don't ever retry more than 3 times */
 	if (req->a_retries++ >= NLMCLNT_MAX_RETRIES)
 		goto die;
-	lock_kernel();
 	nlm_rebind_host(req->a_host);
-	unlock_kernel();
 	rpc_restart_call(task);
 	rpc_delay(task, 30 * HZ);
 }
diff --git a/fs/nfs/Kconfig b/fs/nfs/Kconfig
index b950415d7c4..5c55c26af16 100644
--- a/fs/nfs/Kconfig
+++ b/fs/nfs/Kconfig
@@ -118,3 +118,14 @@ config NFS_USE_KERNEL_DNS
 	select DNS_RESOLVER
 	select KEYS
 	default y
+
+config NFS_USE_NEW_IDMAPPER
+	bool "Use the new idmapper upcall routine"
+	depends on NFS_V4 && KEYS
+	help
+	  Say Y here if you want NFS to use the new idmapper upcall functions.
+	  You will need /sbin/request-key (usually provided by the keyutils
+	  package).  For details, read
+	  <file:Documentation/filesystems/nfs/idmapper.txt>.
+
+	  If you are unsure, say N.
diff --git a/fs/nfs/client.c b/fs/nfs/client.c
index e7340729af8..a882785eba4 100644
--- a/fs/nfs/client.c
+++ b/fs/nfs/client.c
@@ -635,7 +635,8 @@ static int nfs_create_rpc_client(struct nfs_client *clp,
  */
 static void nfs_destroy_server(struct nfs_server *server)
 {
-	if (!(server->flags & NFS_MOUNT_NONLM))
+	if (!(server->flags & NFS_MOUNT_LOCAL_FLOCK) ||
+			!(server->flags & NFS_MOUNT_LOCAL_FCNTL))
 		nlmclnt_done(server->nlm_host);
 }
 
@@ -657,7 +658,8 @@ static int nfs_start_lockd(struct nfs_server *server)
 
 	if (nlm_init.nfs_version > 3)
 		return 0;
-	if (server->flags & NFS_MOUNT_NONLM)
+	if ((server->flags & NFS_MOUNT_LOCAL_FLOCK) &&
+			(server->flags & NFS_MOUNT_LOCAL_FCNTL))
 		return 0;
 
 	switch (clp->cl_proto) {
@@ -901,8 +903,8 @@ static void nfs_server_set_fsinfo(struct nfs_server *server, struct nfs_fsinfo *
 	server->wtmult = nfs_block_bits(fsinfo->wtmult, NULL);
 
 	server->dtsize = nfs_block_size(fsinfo->dtpref, NULL);
-	if (server->dtsize > PAGE_CACHE_SIZE)
-		server->dtsize = PAGE_CACHE_SIZE;
+	if (server->dtsize > PAGE_CACHE_SIZE * NFS_MAX_READDIR_PAGES)
+		server->dtsize = PAGE_CACHE_SIZE * NFS_MAX_READDIR_PAGES;
 	if (server->dtsize > server->rsize)
 		server->dtsize = server->rsize;
 
@@ -913,6 +915,8 @@ static void nfs_server_set_fsinfo(struct nfs_server *server, struct nfs_fsinfo *
 
 	server->maxfilesize = fsinfo->maxfilesize;
 
+	server->time_delta = fsinfo->time_delta;
+
 	/* We're airborne Set socket buffersize */
 	rpc_setbufsize(server->client, server->wsize + 100, server->rsize + 100);
 }
@@ -1356,8 +1360,9 @@ static int nfs4_init_server(struct nfs_server *server,
 
 	/* Initialise the client representation from the mount data */
 	server->flags = data->flags;
-	server->caps |= NFS_CAP_ATOMIC_OPEN|NFS_CAP_CHANGE_ATTR|
-		NFS_CAP_POSIX_LOCK;
+	server->caps |= NFS_CAP_ATOMIC_OPEN|NFS_CAP_CHANGE_ATTR|NFS_CAP_POSIX_LOCK;
+	if (!(data->flags & NFS_MOUNT_NORDIRPLUS))
+			server->caps |= NFS_CAP_READDIRPLUS;
 	server->options = data->options;
 
 	/* Get a client record */
diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index e257172d438..257e4052492 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -33,11 +33,12 @@
 #include <linux/namei.h>
 #include <linux/mount.h>
 #include <linux/sched.h>
+#include <linux/vmalloc.h>
 
-#include "nfs4_fs.h"
 #include "delegation.h"
 #include "iostat.h"
 #include "internal.h"
+#include "fscache.h"
 
 /* #define NFS_DEBUG_VERBOSE 1 */
 
@@ -55,6 +56,7 @@ static int nfs_rename(struct inode *, struct dentry *,
 		      struct inode *, struct dentry *);
 static int nfs_fsync_dir(struct file *, int);
 static loff_t nfs_llseek_dir(struct file *, loff_t, int);
+static int nfs_readdir_clear_array(struct page*, gfp_t);
 
 const struct file_operations nfs_dir_operations = {
 	.llseek		= nfs_llseek_dir,
@@ -80,6 +82,10 @@ const struct inode_operations nfs_dir_inode_operations = {
 	.setattr	= nfs_setattr,
 };
 
+const struct address_space_operations nfs_dir_addr_space_ops = {
+	.releasepage = nfs_readdir_clear_array,
+};
+
 #ifdef CONFIG_NFS_V3
 const struct inode_operations nfs3_dir_inode_operations = {
 	.create		= nfs_create,
@@ -104,8 +110,9 @@ const struct inode_operations nfs3_dir_inode_operations = {
 #ifdef CONFIG_NFS_V4
 
 static struct dentry *nfs_atomic_lookup(struct inode *, struct dentry *, struct nameidata *);
+static int nfs_open_create(struct inode *dir, struct dentry *dentry, int mode, struct nameidata *nd);
 const struct inode_operations nfs4_dir_inode_operations = {
-	.create		= nfs_create,
+	.create		= nfs_open_create,
 	.lookup		= nfs_atomic_lookup,
 	.link		= nfs_link,
 	.unlink		= nfs_unlink,
@@ -150,51 +157,197 @@ nfs_opendir(struct inode *inode, struct file *filp)
 	return res;
 }
 
-typedef __be32 * (*decode_dirent_t)(__be32 *, struct nfs_entry *, int);
+struct nfs_cache_array_entry {
+	u64 cookie;
+	u64 ino;
+	struct qstr string;
+};
+
+struct nfs_cache_array {
+	unsigned int size;
+	int eof_index;
+	u64 last_cookie;
+	struct nfs_cache_array_entry array[0];
+};
+
+#define MAX_READDIR_ARRAY ((PAGE_SIZE - sizeof(struct nfs_cache_array)) / sizeof(struct nfs_cache_array_entry))
+
+typedef __be32 * (*decode_dirent_t)(struct xdr_stream *, struct nfs_entry *, struct nfs_server *, int);
 typedef struct {
 	struct file	*file;
 	struct page	*page;
 	unsigned long	page_index;
-	__be32		*ptr;
 	u64		*dir_cookie;
 	loff_t		current_index;
-	struct nfs_entry *entry;
 	decode_dirent_t	decode;
-	int		plus;
+
 	unsigned long	timestamp;
 	unsigned long	gencount;
-	int		timestamp_valid;
+	unsigned int	cache_entry_index;
+	unsigned int	plus:1;
+	unsigned int	eof:1;
 } nfs_readdir_descriptor_t;
 
-/* Now we cache directories properly, by stuffing the dirent
- * data directly in the page cache.
- *
- * Inode invalidation due to refresh etc. takes care of
- * _everything_, no sloppy entry flushing logic, no extraneous
- * copying, network direct to page cache, the way it was meant
- * to be.
- *
- * NOTE: Dirent information verification is done always by the
- *	 page-in of the RPC reply, nowhere else, this simplies
- *	 things substantially.
+/*
+ * The caller is responsible for calling nfs_readdir_release_array(page)
  */
 static
-int nfs_readdir_filler(nfs_readdir_descriptor_t *desc, struct page *page)
+struct nfs_cache_array *nfs_readdir_get_array(struct page *page)
+{
+	if (page == NULL)
+		return ERR_PTR(-EIO);
+	return (struct nfs_cache_array *)kmap(page);
+}
+
+static
+void nfs_readdir_release_array(struct page *page)
+{
+	kunmap(page);
+}
+
+/*
+ * we are freeing strings created by nfs_add_to_readdir_array()
+ */
+static
+int nfs_readdir_clear_array(struct page *page, gfp_t mask)
+{
+	struct nfs_cache_array *array = nfs_readdir_get_array(page);
+	int i;
+	for (i = 0; i < array->size; i++)
+		kfree(array->array[i].string.name);
+	nfs_readdir_release_array(page);
+	return 0;
+}
+
+/*
+ * the caller is responsible for freeing qstr.name
+ * when called by nfs_readdir_add_to_array, the strings will be freed in
+ * nfs_clear_readdir_array()
+ */
+static
+int nfs_readdir_make_qstr(struct qstr *string, const char *name, unsigned int len)
+{
+	string->len = len;
+	string->name = kmemdup(name, len, GFP_KERNEL);
+	if (string->name == NULL)
+		return -ENOMEM;
+	string->hash = full_name_hash(name, len);
+	return 0;
+}
+
+static
+int nfs_readdir_add_to_array(struct nfs_entry *entry, struct page *page)
+{
+	struct nfs_cache_array *array = nfs_readdir_get_array(page);
+	struct nfs_cache_array_entry *cache_entry;
+	int ret;
+
+	if (IS_ERR(array))
+		return PTR_ERR(array);
+	ret = -EIO;
+	if (array->size >= MAX_READDIR_ARRAY)
+		goto out;
+
+	cache_entry = &array->array[array->size];
+	cache_entry->cookie = entry->prev_cookie;
+	cache_entry->ino = entry->ino;
+	ret = nfs_readdir_make_qstr(&cache_entry->string, entry->name, entry->len);
+	if (ret)
+		goto out;
+	array->last_cookie = entry->cookie;
+	if (entry->eof == 1)
+		array->eof_index = array->size;
+	array->size++;
+out:
+	nfs_readdir_release_array(page);
+	return ret;
+}
+
+static
+int nfs_readdir_search_for_pos(struct nfs_cache_array *array, nfs_readdir_descriptor_t *desc)
+{
+	loff_t diff = desc->file->f_pos - desc->current_index;
+	unsigned int index;
+
+	if (diff < 0)
+		goto out_eof;
+	if (diff >= array->size) {
+		if (array->eof_index > 0)
+			goto out_eof;
+		desc->current_index += array->size;
+		return -EAGAIN;
+	}
+
+	index = (unsigned int)diff;
+	*desc->dir_cookie = array->array[index].cookie;
+	desc->cache_entry_index = index;
+	if (index == array->eof_index)
+		desc->eof = 1;
+	return 0;
+out_eof:
+	desc->eof = 1;
+	return -EBADCOOKIE;
+}
+
+static
+int nfs_readdir_search_for_cookie(struct nfs_cache_array *array, nfs_readdir_descriptor_t *desc)
+{
+	int i;
+	int status = -EAGAIN;
+
+	for (i = 0; i < array->size; i++) {
+		if (i == array->eof_index) {
+			desc->eof = 1;
+			status = -EBADCOOKIE;
+		}
+		if (array->array[i].cookie == *desc->dir_cookie) {
+			desc->cache_entry_index = i;
+			status = 0;
+			break;
+		}
+	}
+
+	return status;
+}
+
+static
+int nfs_readdir_search_array(nfs_readdir_descriptor_t *desc)
+{
+	struct nfs_cache_array *array;
+	int status = -EBADCOOKIE;
+
+	if (desc->dir_cookie == NULL)
+		goto out;
+
+	array = nfs_readdir_get_array(desc->page);
+	if (IS_ERR(array)) {
+		status = PTR_ERR(array);
+		goto out;
+	}
+
+	if (*desc->dir_cookie == 0)
+		status = nfs_readdir_search_for_pos(array, desc);
+	else
+		status = nfs_readdir_search_for_cookie(array, desc);
+
+	nfs_readdir_release_array(desc->page);
+out:
+	return status;
+}
+
+/* Fill a page with xdr information before transferring to the cache page */
+static
+int nfs_readdir_xdr_filler(struct page **pages, nfs_readdir_descriptor_t *desc,
+			struct nfs_entry *entry, struct file *file, struct inode *inode)
 {
-	struct file	*file = desc->file;
-	struct inode	*inode = file->f_path.dentry->d_inode;
 	struct rpc_cred	*cred = nfs_file_cred(file);
 	unsigned long	timestamp, gencount;
 	int		error;
 
-	dfprintk(DIRCACHE, "NFS: %s: reading cookie %Lu into page %lu\n",
-			__func__, (long long)desc->entry->cookie,
-			page->index);
-
  again:
 	timestamp = jiffies;
 	gencount = nfs_inc_attr_generation_counter();
-	error = NFS_PROTO(inode)->readdir(file->f_path.dentry, cred, desc->entry->cookie, page,
+	error = NFS_PROTO(inode)->readdir(file->f_path.dentry, cred, entry->cookie, pages,
 					  NFS_SERVER(inode)->dtsize, desc->plus);
 	if (error < 0) {
 		/* We requested READDIRPLUS, but the server doesn't grok it */
@@ -208,190 +361,292 @@ int nfs_readdir_filler(nfs_readdir_descriptor_t *desc, struct page *page)
 	}
 	desc->timestamp = timestamp;
 	desc->gencount = gencount;
-	desc->timestamp_valid = 1;
-	SetPageUptodate(page);
-	/* Ensure consistent page alignment of the data.
-	 * Note: assumes we have exclusive access to this mapping either
-	 *	 through inode->i_mutex or some other mechanism.
-	 */
-	if (invalidate_inode_pages2_range(inode->i_mapping, page->index + 1, -1) < 0) {
-		/* Should never happen */
-		nfs_zap_mapping(inode, inode->i_mapping);
-	}
-	unlock_page(page);
-	return 0;
- error:
-	unlock_page(page);
-	return -EIO;
+error:
+	return error;
 }
 
-static inline
-int dir_decode(nfs_readdir_descriptor_t *desc)
+/* Fill in an entry based on the xdr code stored in desc->page */
+static
+int xdr_decode(nfs_readdir_descriptor_t *desc, struct nfs_entry *entry, struct xdr_stream *stream)
 {
-	__be32	*p = desc->ptr;
-	p = desc->decode(p, desc->entry, desc->plus);
+	__be32 *p = desc->decode(stream, entry, NFS_SERVER(desc->file->f_path.dentry->d_inode), desc->plus);
 	if (IS_ERR(p))
 		return PTR_ERR(p);
-	desc->ptr = p;
-	if (desc->timestamp_valid) {
-		desc->entry->fattr->time_start = desc->timestamp;
-		desc->entry->fattr->gencount = desc->gencount;
-	} else
-		desc->entry->fattr->valid &= ~NFS_ATTR_FATTR;
+
+	entry->fattr->time_start = desc->timestamp;
+	entry->fattr->gencount = desc->gencount;
 	return 0;
 }
 
-static inline
-void dir_page_release(nfs_readdir_descriptor_t *desc)
+static
+int nfs_same_file(struct dentry *dentry, struct nfs_entry *entry)
 {
-	kunmap(desc->page);
-	page_cache_release(desc->page);
-	desc->page = NULL;
-	desc->ptr = NULL;
+	struct nfs_inode *node;
+	if (dentry->d_inode == NULL)
+		goto different;
+	node = NFS_I(dentry->d_inode);
+	if (node->fh.size != entry->fh->size)
+		goto different;
+	if (strncmp(node->fh.data, entry->fh->data, node->fh.size) != 0)
+		goto different;
+	return 1;
+different:
+	return 0;
 }
 
-/*
- * Given a pointer to a buffer that has already been filled by a call
- * to readdir, find the next entry with cookie '*desc->dir_cookie'.
- *
- * If the end of the buffer has been reached, return -EAGAIN, if not,
- * return the offset within the buffer of the next entry to be
- * read.
- */
-static inline
-int find_dirent(nfs_readdir_descriptor_t *desc)
+static
+void nfs_prime_dcache(struct dentry *parent, struct nfs_entry *entry)
 {
-	struct nfs_entry *entry = desc->entry;
-	int		loop_count = 0,
-			status;
+	struct qstr filename = {
+		.len = entry->len,
+		.name = entry->name,
+	};
+	struct dentry *dentry;
+	struct dentry *alias;
+	struct inode *dir = parent->d_inode;
+	struct inode *inode;
 
-	while((status = dir_decode(desc)) == 0) {
-		dfprintk(DIRCACHE, "NFS: %s: examining cookie %Lu\n",
-				__func__, (unsigned long long)entry->cookie);
-		if (entry->prev_cookie == *desc->dir_cookie)
-			break;
-		if (loop_count++ > 200) {
-			loop_count = 0;
-			schedule();
+	if (filename.name[0] == '.') {
+		if (filename.len == 1)
+			return;
+		if (filename.len == 2 && filename.name[1] == '.')
+			return;
+	}
+	filename.hash = full_name_hash(filename.name, filename.len);
+
+	dentry = d_lookup(parent, &filename);
+	if (dentry != NULL) {
+		if (nfs_same_file(dentry, entry)) {
+			nfs_refresh_inode(dentry->d_inode, entry->fattr);
+			goto out;
+		} else {
+			d_drop(dentry);
+			dput(dentry);
 		}
 	}
-	return status;
+
+	dentry = d_alloc(parent, &filename);
+	if (dentry == NULL)
+		return;
+
+	dentry->d_op = NFS_PROTO(dir)->dentry_ops;
+	inode = nfs_fhget(dentry->d_sb, entry->fh, entry->fattr);
+	if (IS_ERR(inode))
+		goto out;
+
+	alias = d_materialise_unique(dentry, inode);
+	if (IS_ERR(alias))
+		goto out;
+	else if (alias) {
+		nfs_set_verifier(alias, nfs_save_change_attribute(dir));
+		dput(alias);
+	} else
+		nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
+
+out:
+	dput(dentry);
+}
+
+/* Perform conversion from xdr to cache array */
+static
+void nfs_readdir_page_filler(nfs_readdir_descriptor_t *desc, struct nfs_entry *entry,
+				void *xdr_page, struct page *page, unsigned int buflen)
+{
+	struct xdr_stream stream;
+	struct xdr_buf buf;
+	__be32 *ptr = xdr_page;
+	int status;
+	struct nfs_cache_array *array;
+
+	buf.head->iov_base = xdr_page;
+	buf.head->iov_len = buflen;
+	buf.tail->iov_len = 0;
+	buf.page_base = 0;
+	buf.page_len = 0;
+	buf.buflen = buf.head->iov_len;
+	buf.len = buf.head->iov_len;
+
+	xdr_init_decode(&stream, &buf, ptr);
+
+
+	do {
+		status = xdr_decode(desc, entry, &stream);
+		if (status != 0)
+			break;
+
+		if (nfs_readdir_add_to_array(entry, page) == -1)
+			break;
+		if (desc->plus == 1)
+			nfs_prime_dcache(desc->file->f_path.dentry, entry);
+	} while (!entry->eof);
+
+	if (status == -EBADCOOKIE && entry->eof) {
+		array = nfs_readdir_get_array(page);
+		array->eof_index = array->size - 1;
+		status = 0;
+		nfs_readdir_release_array(page);
+	}
+}
+
+static
+void nfs_readdir_free_pagearray(struct page **pages, unsigned int npages)
+{
+	unsigned int i;
+	for (i = 0; i < npages; i++)
+		put_page(pages[i]);
+}
+
+static
+void nfs_readdir_free_large_page(void *ptr, struct page **pages,
+		unsigned int npages)
+{
+	vm_unmap_ram(ptr, npages);
+	nfs_readdir_free_pagearray(pages, npages);
 }
 
 /*
- * Given a pointer to a buffer that has already been filled by a call
- * to readdir, find the entry at offset 'desc->file->f_pos'.
- *
- * If the end of the buffer has been reached, return -EAGAIN, if not,
- * return the offset within the buffer of the next entry to be
- * read.
+ * nfs_readdir_large_page will allocate pages that must be freed with a call
+ * to nfs_readdir_free_large_page
  */
-static inline
-int find_dirent_index(nfs_readdir_descriptor_t *desc)
+static
+void *nfs_readdir_large_page(struct page **pages, unsigned int npages)
 {
-	struct nfs_entry *entry = desc->entry;
-	int		loop_count = 0,
-			status;
+	void *ptr;
+	unsigned int i;
+
+	for (i = 0; i < npages; i++) {
+		struct page *page = alloc_page(GFP_KERNEL);
+		if (page == NULL)
+			goto out_freepages;
+		pages[i] = page;
+	}
 
-	for(;;) {
-		status = dir_decode(desc);
-		if (status)
-			break;
+	ptr = vm_map_ram(pages, npages, 0, PAGE_KERNEL);
+	if (!IS_ERR_OR_NULL(ptr))
+		return ptr;
+out_freepages:
+	nfs_readdir_free_pagearray(pages, i);
+	return NULL;
+}
+
+static
+int nfs_readdir_xdr_to_array(nfs_readdir_descriptor_t *desc, struct page *page, struct inode *inode)
+{
+	struct page *pages[NFS_MAX_READDIR_PAGES];
+	void *pages_ptr = NULL;
+	struct nfs_entry entry;
+	struct file	*file = desc->file;
+	struct nfs_cache_array *array;
+	int status = 0;
+	unsigned int array_size = ARRAY_SIZE(pages);
+
+	entry.prev_cookie = 0;
+	entry.cookie = *desc->dir_cookie;
+	entry.eof = 0;
+	entry.fh = nfs_alloc_fhandle();
+	entry.fattr = nfs_alloc_fattr();
+	if (entry.fh == NULL || entry.fattr == NULL)
+		goto out;
 
-		dfprintk(DIRCACHE, "NFS: found cookie %Lu at index %Ld\n",
-				(unsigned long long)entry->cookie, desc->current_index);
+	array = nfs_readdir_get_array(page);
+	memset(array, 0, sizeof(struct nfs_cache_array));
+	array->eof_index = -1;
 
-		if (desc->file->f_pos == desc->current_index) {
-			*desc->dir_cookie = entry->cookie;
+	pages_ptr = nfs_readdir_large_page(pages, array_size);
+	if (!pages_ptr)
+		goto out_release_array;
+	do {
+		status = nfs_readdir_xdr_filler(pages, desc, &entry, file, inode);
+
+		if (status < 0)
 			break;
-		}
-		desc->current_index++;
-		if (loop_count++ > 200) {
-			loop_count = 0;
-			schedule();
-		}
-	}
+		nfs_readdir_page_filler(desc, &entry, pages_ptr, page, array_size * PAGE_SIZE);
+	} while (array->eof_index < 0 && array->size < MAX_READDIR_ARRAY);
+
+	nfs_readdir_free_large_page(pages_ptr, pages, array_size);
+out_release_array:
+	nfs_readdir_release_array(page);
+out:
+	nfs_free_fattr(entry.fattr);
+	nfs_free_fhandle(entry.fh);
 	return status;
 }
 
 /*
- * Find the given page, and call find_dirent() or find_dirent_index in
- * order to try to return the next entry.
+ * Now we cache directories properly, by converting xdr information
+ * to an array that can be used for lookups later.  This results in
+ * fewer cache pages, since we can store more information on each page.
+ * We only need to convert from xdr once so future lookups are much simpler
  */
-static inline
-int find_dirent_page(nfs_readdir_descriptor_t *desc)
+static
+int nfs_readdir_filler(nfs_readdir_descriptor_t *desc, struct page* page)
 {
 	struct inode	*inode = desc->file->f_path.dentry->d_inode;
-	struct page	*page;
-	int		status;
 
-	dfprintk(DIRCACHE, "NFS: %s: searching page %ld for target %Lu\n",
-			__func__, desc->page_index,
-			(long long) *desc->dir_cookie);
+	if (nfs_readdir_xdr_to_array(desc, page, inode) < 0)
+		goto error;
+	SetPageUptodate(page);
 
-	/* If we find the page in the page_cache, we cannot be sure
-	 * how fresh the data is, so we will ignore readdir_plus attributes.
-	 */
-	desc->timestamp_valid = 0;
-	page = read_cache_page(inode->i_mapping, desc->page_index,
-			       (filler_t *)nfs_readdir_filler, desc);
-	if (IS_ERR(page)) {
-		status = PTR_ERR(page);
-		goto out;
+	if (invalidate_inode_pages2_range(inode->i_mapping, page->index + 1, -1) < 0) {
+		/* Should never happen */
+		nfs_zap_mapping(inode, inode->i_mapping);
 	}
+	unlock_page(page);
+	return 0;
+ error:
+	unlock_page(page);
+	return -EIO;
+}
 
-	/* NOTE: Someone else may have changed the READDIRPLUS flag */
-	desc->page = page;
-	desc->ptr = kmap(page);		/* matching kunmap in nfs_do_filldir */
-	if (*desc->dir_cookie != 0)
-		status = find_dirent(desc);
-	else
-		status = find_dirent_index(desc);
-	if (status < 0)
-		dir_page_release(desc);
- out:
-	dfprintk(DIRCACHE, "NFS: %s: returns %d\n", __func__, status);
-	return status;
+static
+void cache_page_release(nfs_readdir_descriptor_t *desc)
+{
+	page_cache_release(desc->page);
+	desc->page = NULL;
+}
+
+static
+struct page *get_cache_page(nfs_readdir_descriptor_t *desc)
+{
+	struct page *page;
+	page = read_cache_page(desc->file->f_path.dentry->d_inode->i_mapping,
+			desc->page_index, (filler_t *)nfs_readdir_filler, desc);
+	if (IS_ERR(page))
+		desc->eof = 1;
+	return page;
 }
 
 /*
- * Recurse through the page cache pages, and return a
- * filled nfs_entry structure of the next directory entry if possible.
- *
- * The target for the search is '*desc->dir_cookie' if non-0,
- * 'desc->file->f_pos' otherwise
+ * Returns 0 if desc->dir_cookie was found on page desc->page_index
  */
+static
+int find_cache_page(nfs_readdir_descriptor_t *desc)
+{
+	int res;
+
+	desc->page = get_cache_page(desc);
+	if (IS_ERR(desc->page))
+		return PTR_ERR(desc->page);
+
+	res = nfs_readdir_search_array(desc);
+	if (res == 0)
+		return 0;
+	cache_page_release(desc);
+	return res;
+}
+
+/* Search for desc->dir_cookie from the beginning of the page cache */
 static inline
 int readdir_search_pagecache(nfs_readdir_descriptor_t *desc)
 {
-	int		loop_count = 0;
-	int		res;
-
-	/* Always search-by-index from the beginning of the cache */
-	if (*desc->dir_cookie == 0) {
-		dfprintk(DIRCACHE, "NFS: readdir_search_pagecache() searching for offset %Ld\n",
-				(long long)desc->file->f_pos);
-		desc->page_index = 0;
-		desc->entry->cookie = desc->entry->prev_cookie = 0;
-		desc->entry->eof = 0;
-		desc->current_index = 0;
-	} else
-		dfprintk(DIRCACHE, "NFS: readdir_search_pagecache() searching for cookie %Lu\n",
-				(unsigned long long)*desc->dir_cookie);
+	int res = -EAGAIN;
 
-	for (;;) {
-		res = find_dirent_page(desc);
+	while (1) {
+		res = find_cache_page(desc);
 		if (res != -EAGAIN)
 			break;
-		/* Align to beginning of next page */
-		desc->page_index ++;
-		if (loop_count++ > 200) {
-			loop_count = 0;
-			schedule();
-		}
+		desc->page_index++;
 	}
-
-	dfprintk(DIRCACHE, "NFS: %s: returns %d\n", __func__, res);
 	return res;