Merge branch 'afs-dh' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs

Pull AFS updates from Al Viro:
 "The AFS series posted by dhowells depended upon lookup_one_len()
  rework; now that prereq is in the mainline, that series had been
  rebased on top of it and got some exposure and testing..."

* 'afs-dh' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs:
  afs: Do better accretion of small writes on newly created content
  afs: Add stats for data transfer operations
  afs: Trace protocol errors
  afs: Locally edit directory data for mkdir/create/unlink/...
  afs: Adjust the directory XDR structures
  afs: Split the directory content defs into a header
  afs: Fix directory handling
  afs: Split the dynroot stuff out and give it its own ops tables
  afs: Keep track of invalid-before version for dentry coherency
  afs: Rearrange status mapping
  afs: Make it possible to get the data version in readpage
  afs: Init inode before accessing cache
  afs: Introduce a statistics proc file
  afs: Dump bad status record
  afs: Implement @cell substitution handling
  afs: Implement @sys substitution handling
  afs: Prospectively look up extra files when doing a single lookup
  afs: Don't over-increment the cell usage count when pinning it
  afs: Fix checker warnings
  vfs: Remove the const from dir_context::actor

25 files changed, 2562 insertions, 591 deletions

diff --git a/fs/afs/Makefile b/fs/afs/Makefile
index 45b7fc405fa6..532acae25453 100644
--- a/fs/afs/Makefile
+++ b/fs/afs/Makefile
@@ -12,6 +12,8 @@ kafs-objs := \
 	cell.o \
 	cmservice.o \
 	dir.o \
+	dir_edit.o \
+	dynroot.o \
 	file.o \
 	flock.o \
 	fsclient.o \
diff --git a/fs/afs/addr_list.c b/fs/afs/addr_list.c
index fd9f28b8a933..3bedfed608a2 100644
--- a/fs/afs/addr_list.c
+++ b/fs/afs/addr_list.c
@@ -243,9 +243,9 @@ void afs_merge_fs_addr4(struct afs_addr_list *alist, __be32 xdr, u16 port)
 		    xport == a->sin6_port)
 			return;
 		if (xdr == a->sin6_addr.s6_addr32[3] &&
-		    xport < a->sin6_port)
+		    (u16 __force)xport < (u16 __force)a->sin6_port)
 			break;
-		if (xdr < a->sin6_addr.s6_addr32[3])
+		if ((u32 __force)xdr < (u32 __force)a->sin6_addr.s6_addr32[3])
 			break;
 	}
 
@@ -280,7 +280,7 @@ void afs_merge_fs_addr6(struct afs_addr_list *alist, __be32 *xdr, u16 port)
 		    xport == a->sin6_port)
 			return;
 		if (diff == 0 &&
-		    xport < a->sin6_port)
+		    (u16 __force)xport < (u16 __force)a->sin6_port)
 			break;
 		if (diff < 0)
 			break;
diff --git a/fs/afs/afs.h b/fs/afs/afs.h
index b94d0edc2b78..b4ff1f7ae4ab 100644
--- a/fs/afs/afs.h
+++ b/fs/afs/afs.h
@@ -67,10 +67,14 @@ typedef enum {
 } afs_callback_type_t;
 
 struct afs_callback {
-	struct afs_fid		fid;		/* file identifier */
-	unsigned		version;	/* callback version */
-	unsigned		expiry;		/* time at which expires */
-	afs_callback_type_t	type;		/* type of callback */
+	unsigned		version;	/* Callback version */
+	unsigned		expiry;		/* Time at which expires */
+	afs_callback_type_t	type;		/* Type of callback */
+};
+
+struct afs_callback_break {
+	struct afs_fid		fid;		/* File identifier */
+	struct afs_callback	cb;		/* Callback details */
 };
 
 #define AFSCBMAX 50	/* maximum callbacks transferred per bulk op */
@@ -123,21 +127,20 @@ typedef u32 afs_access_t;
  * AFS file status information
  */
 struct afs_file_status {
-	unsigned		if_version;	/* interface version */
-#define AFS_FSTATUS_VERSION	1
+	u64			size;		/* file size */
+	afs_dataversion_t	data_version;	/* current data version */
+	time_t			mtime_client;	/* last time client changed data */
+	time_t			mtime_server;	/* last time server changed data */
+	unsigned		abort_code;	/* Abort if bulk-fetching this failed */
 
 	afs_file_type_t		type;		/* file type */
 	unsigned		nlink;		/* link count */
-	u64			size;		/* file size */
-	afs_dataversion_t	data_version;	/* current data version */
 	u32			author;		/* author ID */
-	kuid_t			owner;		/* owner ID */
-	kgid_t			group;		/* group ID */
+	u32			owner;		/* owner ID */
+	u32			group;		/* group ID */
 	afs_access_t		caller_access;	/* access rights for authenticated caller */
 	afs_access_t		anon_access;	/* access rights for unauthenticated caller */
 	umode_t			mode;		/* UNIX mode */
-	time_t			mtime_client;	/* last time client changed data */
-	time_t			mtime_server;	/* last time server changed data */
 	s32			lock_count;	/* file lock count (0=UNLK -1=WRLCK +ve=#RDLCK */
 };
 
diff --git a/fs/afs/afs_fs.h b/fs/afs/afs_fs.h
index d47b6d01e4c0..ddfa88a7a9c0 100644
--- a/fs/afs/afs_fs.h
+++ b/fs/afs/afs_fs.h
@@ -31,10 +31,12 @@ enum AFS_FS_Operations {
 	FSGETVOLUMEINFO		= 148,	/* AFS Get information about a volume */
 	FSGETVOLUMESTATUS	= 149,	/* AFS Get volume status information */
 	FSGETROOTVOLUME		= 151,	/* AFS Get root volume name */
+	FSBULKSTATUS		= 155,	/* AFS Fetch multiple file statuses */
 	FSSETLOCK		= 156,	/* AFS Request a file lock */
 	FSEXTENDLOCK		= 157,	/* AFS Extend a file lock */
 	FSRELEASELOCK		= 158,	/* AFS Release a file lock */
 	FSLOOKUP		= 161,	/* AFS lookup file in directory */
+	FSINLINEBULKSTATUS	= 65536, /* AFS Fetch multiple file statuses with inline errors */
 	FSFETCHDATA64		= 65537, /* AFS Fetch file data */
 	FSSTOREDATA64		= 65538, /* AFS Store file data */
 	FSGIVEUPALLCALLBACKS	= 65539, /* AFS Give up all outstanding callbacks on a server */
diff --git a/fs/afs/callback.c b/fs/afs/callback.c
index f4291b576054..abd9a84f4e88 100644
--- a/fs/afs/callback.c
+++ b/fs/afs/callback.c
@@ -97,26 +97,6 @@ again:
 }
 
 /*
- * Set a vnode's interest on a server.
- */
-void afs_set_cb_interest(struct afs_vnode *vnode, struct afs_cb_interest *cbi)
-{
-	struct afs_cb_interest *old_cbi = NULL;
-
-	if (vnode->cb_interest == cbi)
-		return;
-
-	write_seqlock(&vnode->cb_lock);
-	if (vnode->cb_interest != cbi) {
-		afs_get_cb_interest(cbi);
-		old_cbi = vnode->cb_interest;
-		vnode->cb_interest = cbi;
-	}
-	write_sequnlock(&vnode->cb_lock);
-	afs_put_cb_interest(afs_v2net(vnode), cbi);
-}
-
-/*
  * Remove an interest on a server.
  */
 void afs_put_cb_interest(struct afs_net *net, struct afs_cb_interest *cbi)
@@ -150,6 +130,7 @@ void afs_break_callback(struct afs_vnode *vnode)
 
 	write_seqlock(&vnode->cb_lock);
 
+	clear_bit(AFS_VNODE_NEW_CONTENT, &vnode->flags);
 	if (test_and_clear_bit(AFS_VNODE_CB_PROMISED, &vnode->flags)) {
 		vnode->cb_break++;
 		afs_clear_permits(vnode);
@@ -207,7 +188,7 @@ static void afs_break_one_callback(struct afs_server *server,
  * allow the fileserver to break callback promises
  */
 void afs_break_callbacks(struct afs_server *server, size_t count,
-			 struct afs_callback callbacks[])
+			 struct afs_callback_break *callbacks)
 {
 	_enter("%p,%zu,", server, count);
 
@@ -219,9 +200,9 @@ void afs_break_callbacks(struct afs_server *server, size_t count,
 		       callbacks->fid.vid,
 		       callbacks->fid.vnode,
 		       callbacks->fid.unique,
-		       callbacks->version,
-		       callbacks->expiry,
-		       callbacks->type
+		       callbacks->cb.version,
+		       callbacks->cb.expiry,
+		       callbacks->cb.type
 		       );
 		afs_break_one_callback(server, &callbacks->fid);
 	}
diff --git a/fs/afs/cell.c b/fs/afs/cell.c
index 4235a05afc76..fdf4c36cff79 100644
--- a/fs/afs/cell.c
+++ b/fs/afs/cell.c
@@ -18,7 +18,7 @@
 #include <keys/rxrpc-type.h>
 #include "internal.h"
 
-unsigned __read_mostly afs_cell_gc_delay = 10;
+static unsigned __read_mostly afs_cell_gc_delay = 10;
 
 static void afs_manage_cell(struct work_struct *);
 
@@ -75,7 +75,7 @@ struct afs_cell *afs_lookup_cell_rcu(struct afs_net *net,
 			cell = rcu_dereference_raw(net->ws_cell);
 			if (cell) {
 				afs_get_cell(cell);
-				continue;
+				break;
 			}
 			ret = -EDESTADDRREQ;
 			continue;
@@ -130,6 +130,8 @@ static struct afs_cell *afs_alloc_cell(struct afs_net *net,
 		_leave(" = -ENAMETOOLONG");
 		return ERR_PTR(-ENAMETOOLONG);
 	}
+	if (namelen == 5 && memcmp(name, "@cell", 5) == 0)
+		return ERR_PTR(-EINVAL);
 
 	_enter("%*.*s,%s", namelen, namelen, name, vllist);
 
@@ -334,8 +336,8 @@ int afs_cell_init(struct afs_net *net, const char *rootcell)
 		return PTR_ERR(new_root);
 	}
 
-	set_bit(AFS_CELL_FL_NO_GC, &new_root->flags);
-	afs_get_cell(new_root);
+	if (!test_and_set_bit(AFS_CELL_FL_NO_GC, &new_root->flags))
+		afs_get_cell(new_root);
 
 	/* install the new cell */
 	write_seqlock(&net->cells_lock);
@@ -411,7 +413,7 @@ static void afs_cell_destroy(struct rcu_head *rcu)
 
 	ASSERTCMP(atomic_read(&cell->usage), ==, 0);
 
-	afs_put_addrlist(cell->vl_addrs);
+	afs_put_addrlist(rcu_access_pointer(cell->vl_addrs));
 	key_put(cell->anonymous_key);
 	kfree(cell);
 
diff --git a/fs/afs/cmservice.c b/fs/afs/cmservice.c
index 41e277f57b20..357de908df3a 100644
--- a/fs/afs/cmservice.c
+++ b/fs/afs/cmservice.c
@@ -178,8 +178,8 @@ static void SRXAFSCB_CallBack(struct work_struct *work)
  */
 static int afs_deliver_cb_callback(struct afs_call *call)
 {
+	struct afs_callback_break *cb;
 	struct sockaddr_rxrpc srx;
-	struct afs_callback *cb;
 	struct afs_server *server;
 	__be32 *bp;
 	int ret, loop;
@@ -201,7 +201,7 @@ static int afs_deliver_cb_callback(struct afs_call *call)
 		call->count = ntohl(call->tmp);
 		_debug("FID count: %u", call->count);
 		if (call->count > AFSCBMAX)
-			return -EBADMSG;
+			return afs_protocol_error(call, -EBADMSG);
 
 		call->buffer = kmalloc(call->count * 3 * 4, GFP_KERNEL);
 		if (!call->buffer)
@@ -218,7 +218,7 @@ static int afs_deliver_cb_callback(struct afs_call *call)
 
 		_debug("unmarshall FID array");
 		call->request = kcalloc(call->count,
-					sizeof(struct afs_callback),
+					sizeof(struct afs_callback_break),
 					GFP_KERNEL);
 		if (!call->request)
 			return -ENOMEM;
@@ -229,7 +229,7 @@ static int afs_deliver_cb_callback(struct afs_call *call)
 			cb->fid.vid	= ntohl(*bp++);
 			cb->fid.vnode	= ntohl(*bp++);
 			cb->fid.unique	= ntohl(*bp++);
-			cb->type	= AFSCM_CB_UNTYPED;
+			cb->cb.type	= AFSCM_CB_UNTYPED;
 		}
 
 		call->offset = 0;
@@ -245,7 +245,7 @@ static int afs_deliver_cb_callback(struct afs_call *call)
 		call->count2 = ntohl(call->tmp);
 		_debug("CB count: %u", call->count2);
 		if (call->count2 != call->count && call->count2 != 0)
-			return -EBADMSG;
+			return afs_protocol_error(call, -EBADMSG);
 		call->offset = 0;
 		call->unmarshall++;
 
@@ -260,9 +260,9 @@ static int afs_deliver_cb_callback(struct afs_call *call)
 		cb = call->request;
 		bp = call->buffer;
 		for (loop = call->count2; loop > 0; loop--, cb++) {
-			cb->version	= ntohl(*bp++);
-			cb->expiry	= ntohl(*bp++);
-			cb->type	= ntohl(*bp++);
+			cb->cb.version	= ntohl(*bp++);
+			cb->cb.expiry	= ntohl(*bp++);
+			cb->cb.type	= ntohl(*bp++);
 		}
 
 		call->offset = 0;
@@ -500,9 +500,9 @@ static int afs_deliver_cb_probe_uuid(struct afs_call *call)
 
 		b = call->buffer;
 		r = call->request;
-		r->time_low			= ntohl(b[0]);
-		r->time_mid			= ntohl(b[1]);
-		r->time_hi_and_version		= ntohl(b[2]);
+		r->time_low			= b[0];
+		r->time_mid			= htons(ntohl(b[1]));
+		r->time_hi_and_version		= htons(ntohl(b[2]));
 		r->clock_seq_hi_and_reserved 	= ntohl(b[3]);
 		r->clock_seq_low		= ntohl(b[4]);
 
diff --git a/fs/afs/dir.c b/fs/afs/dir.c
index ba2b458b36d1..5889f70d4d27 100644
--- a/fs/afs/dir.c
+++ b/fs/afs/dir.c
@@ -1,6 +1,6 @@
 /* dir.c: AFS filesystem directory handling
  *
- * Copyright (C) 2002 Red Hat, Inc. All Rights Reserved.
+ * Copyright (C) 2002, 2018 Red Hat, Inc. All Rights Reserved.
  * Written by David Howells (dhowells@redhat.com)
  *
  * This program is free software; you can redistribute it and/or
@@ -10,27 +10,26 @@
  */
 
 #include <linux/kernel.h>
-#include <linux/module.h>
-#include <linux/init.h>
 #include <linux/fs.h>
 #include <linux/namei.h>
 #include <linux/pagemap.h>
+#include <linux/swap.h>
 #include <linux/ctype.h>
 #include <linux/sched.h>
-#include <linux/dns_resolver.h>
+#include <linux/task_io_accounting_ops.h>
 #include "internal.h"
+#include "xdr_fs.h"
 
 static struct dentry *afs_lookup(struct inode *dir, struct dentry *dentry,
 				 unsigned int flags);
-static struct dentry *afs_dynroot_lookup(struct inode *dir, struct dentry *dentry,
-					 unsigned int flags);
 static int afs_dir_open(struct inode *inode, struct file *file);
 static int afs_readdir(struct file *file, struct dir_context *ctx);
 static int afs_d_revalidate(struct dentry *dentry, unsigned int flags);
 static int afs_d_delete(const struct dentry *dentry);
-static void afs_d_release(struct dentry *dentry);
-static int afs_lookup_filldir(struct dir_context *ctx, const char *name, int nlen,
+static int afs_lookup_one_filldir(struct dir_context *ctx, const char *name, int nlen,
 				  loff_t fpos, u64 ino, unsigned dtype);
+static int afs_lookup_filldir(struct dir_context *ctx, const char *name, int nlen,
+			      loff_t fpos, u64 ino, unsigned dtype);
 static int afs_create(struct inode *dir, struct dentry *dentry, umode_t mode,
 		      bool excl);
 static int afs_mkdir(struct inode *dir, struct dentry *dentry, umode_t mode);
@@ -43,6 +42,14 @@ static int afs_symlink(struct inode *dir, struct dentry *dentry,
 static int afs_rename(struct inode *old_dir, struct dentry *old_dentry,
 		      struct inode *new_dir, struct dentry *new_dentry,
 		      unsigned int flags);
+static int afs_dir_releasepage(struct page *page, gfp_t gfp_flags);
+static void afs_dir_invalidatepage(struct page *page, unsigned int offset,
+				   unsigned int length);
+
+static int afs_dir_set_page_dirty(struct page *page)
+{
+	BUG(); /* This should never happen. */
+}
 
 const struct file_operations afs_dir_file_operations = {
 	.open		= afs_dir_open,
@@ -67,15 +74,10 @@ const struct inode_operations afs_dir_inode_operations = {
 	.listxattr	= afs_listxattr,
 };
 
-const struct file_operations afs_dynroot_file_operations = {
-	.open		= dcache_dir_open,
-	.release	= dcache_dir_close,
-	.iterate_shared	= dcache_readdir,
-	.llseek		= dcache_dir_lseek,
-};
-
-const struct inode_operations afs_dynroot_inode_operations = {
-	.lookup		= afs_dynroot_lookup,
+const struct address_space_operations afs_dir_aops = {
+	.set_page_dirty	= afs_dir_set_page_dirty,
+	.releasepage	= afs_dir_releasepage,
+	.invalidatepage	= afs_dir_invalidatepage,
 };
 
 const struct dentry_operations afs_fs_dentry_operations = {
@@ -85,91 +87,38 @@ const struct dentry_operations afs_fs_dentry_operations = {
 	.d_automount	= afs_d_automount,
 };
 
-#define AFS_DIR_HASHTBL_SIZE	128
-#define AFS_DIR_DIRENT_SIZE	32
-#define AFS_DIRENT_PER_BLOCK	64
-
-union afs_dirent {
-	struct {
-		uint8_t		valid;
-		uint8_t		unused[1];
-		__be16		hash_next;
-		__be32		vnode;
-		__be32		unique;
-		uint8_t		name[16];
-		uint8_t		overflow[4];	/* if any char of the name (inc
-						 * NUL) reaches here, consume
-						 * the next dirent too */
-	} u;
-	uint8_t	extended_name[32];
-};
-
-/* AFS directory page header (one at the beginning of every 2048-byte chunk) */
-struct afs_dir_pagehdr {
-	__be16		npages;
-	__be16		magic;
-#define AFS_DIR_MAGIC htons(1234)
-	uint8_t		nentries;
-	uint8_t		bitmap[8];
-	uint8_t		pad[19];
-};
-
-/* directory block layout */
-union afs_dir_block {
-
-	struct afs_dir_pagehdr pagehdr;
-
-	struct {
-		struct afs_dir_pagehdr	pagehdr;
-		uint8_t			alloc_ctrs[128];
-		/* dir hash table */
-		uint16_t		hashtable[AFS_DIR_HASHTBL_SIZE];
-	} hdr;
-
-	union afs_dirent dirents[AFS_DIRENT_PER_BLOCK];
-};
-
-/* layout on a linux VM page */
-struct afs_dir_page {
-	union afs_dir_block blocks[PAGE_SIZE / sizeof(union afs_dir_block)];
+struct afs_lookup_one_cookie {
+	struct dir_context	ctx;
+	struct qstr		name;
+	bool			found;
+	struct afs_fid		fid;
 };
 
 struct afs_lookup_cookie {
-	struct dir_context ctx;
-	struct afs_fid	fid;
-	struct qstr name;
-	int		found;
+	struct dir_context	ctx;
+	struct qstr		name;
+	bool			found;
+	bool			one_only;
+	unsigned short		nr_fids;
+	struct afs_file_status	*statuses;
+	struct afs_callback	*callbacks;
+	struct afs_fid		fids[50];
 };
 
 /*
  * check that a directory page is valid
  */
-bool afs_dir_check_page(struct inode *dir, struct page *page)
+static bool afs_dir_check_page(struct afs_vnode *dvnode, struct page *page,
+			       loff_t i_size)
 {
-	struct afs_dir_page *dbuf;
-	struct afs_vnode *vnode = AFS_FS_I(dir);
-	loff_t latter, i_size, off;
+	struct afs_xdr_dir_page *dbuf;
+	loff_t latter, off;
 	int tmp, qty;
 
-#if 0
-	/* check the page count */
-	qty = desc.size / sizeof(dbuf->blocks[0]);
-	if (qty == 0)
-		goto error;
-
-	if (page->index == 0 && qty != ntohs(dbuf->blocks[0].pagehdr.npages)) {
-		printk("kAFS: %s(%lu): wrong number of dir blocks %d!=%hu\n",
-		       __func__, dir->i_ino, qty,
-		       ntohs(dbuf->blocks[0].pagehdr.npages));
-		goto error;
-	}
-#endif
-
 	/* Determine how many magic numbers there should be in this page, but
 	 * we must take care because the directory may change size under us.
 	 */
 	off = page_offset(page);
-	i_size = i_size_read(dir);
 	if (i_size <= off)
 		goto checked;
 
@@ -178,112 +127,225 @@ bool afs_dir_check_page(struct inode *dir, struct page *page)
 		qty = PAGE_SIZE;
 	else
 		qty = latter;
-	qty /= sizeof(union afs_dir_block);
+	qty /= sizeof(union afs_xdr_dir_block);
 
 	/* check them */
-	dbuf = page_address(page);
+	dbuf = kmap(page);
 	for (tmp = 0; tmp < qty; tmp++) {
-		if (dbuf->blocks[tmp].pagehdr.magic != AFS_DIR_MAGIC) {
+		if (dbuf->blocks[tmp].hdr.magic != AFS_DIR_MAGIC) {
 			printk("kAFS: %s(%lx): bad magic %d/%d is %04hx\n",
-			       __func__, dir->i_ino, tmp, qty,
-			       ntohs(dbuf->blocks[tmp].pagehdr.magic));
-			trace_afs_dir_check_failed(vnode, off, i_size);
+			       __func__, dvnode->vfs_inode.i_ino, tmp, qty,
+			       ntohs(dbuf->blocks[tmp].hdr.magic));
+			trace_afs_dir_check_failed(dvnode, off, i_size);
+			kunmap(page);
 			goto error;
 		}
+
+		/* Make sure each block is NUL terminated so we can reasonably
+		 * use string functions on it.  The filenames in the page
+		 * *should* be NUL-terminated anyway.
+		 */
+		((u8 *)&dbuf->blocks[tmp])[AFS_DIR_BLOCK_SIZE - 1] = 0;
 	}
 
+	kunmap(page);
+
 checked:
-	SetPageChecked(page);
+	afs_stat_v(dvnode, n_read_dir);
 	return true;
 
 error:
-	SetPageError(page);
 	return false;
 }
 
 /*
- * discard a page cached in the pagecache
+ * open an AFS directory file
  */
-static inline void afs_dir_put_page(struct page *page)
+static int afs_dir_open(struct inode *inode, struct file *file)
 {
-	kunmap(page);
-	unlock_page(page);
-	put_page(page);
+	_enter("{%lu}", inode->i_ino);
+
+	BUILD_BUG_ON(sizeof(union afs_xdr_dir_block) != 2048);
+	BUILD_BUG_ON(sizeof(union afs_xdr_dirent) != 32);
+
+	if (test_bit(AFS_VNODE_DELETED, &AFS_FS_I(inode)->flags))
+		return -ENOENT;
+
+	return afs_open(inode, file);
 }
 
 /*
- * get a page into the pagecache
+ * Read the directory into the pagecache in one go, scrubbing the previous
+ * contents.  The list of pages is returned, pinning them so that they don't
+ * get reclaimed during the iteration.
  */
-static struct page *afs_dir_get_page(struct inode *dir, unsigned long index,
-				     struct key *key)
+static struct afs_read *afs_read_dir(struct afs_vnode *dvnode, struct key *key)
 {
-	struct page *page;
-	_enter("{%lu},%lu", dir->i_ino, index);
-
-	page = read_cache_page(dir->i_mapping, index, afs_page_filler, key);
-	if (!IS_ERR(page)) {
-		lock_page(page);
-		kmap(page);
-		if (unlikely(!PageChecked(page))) {
-			if (PageError(page))
-				goto fail;
-		}
+	struct afs_read *req;
+	loff_t i_size;
+	int nr_pages, nr_inline, i, n;
+	int ret = -ENOMEM;
+
+retry:
+	i_size = i_size_read(&dvnode->vfs_inode);
+	if (i_size < 2048)
+		return ERR_PTR(-EIO);
+	if (i_size > 2048 * 1024)
+		return ERR_PTR(-EFBIG);
+
+	_enter("%llu", i_size);
+
+	/* Get a request record to hold the page list.  We want to hold it
+	 * inline if we can, but we don't want to make an order 1 allocation.
+	 */
+	nr_pages = (i_size + PAGE_SIZE - 1) / PAGE_SIZE;
+	nr_inline = nr_pages;
+	if (nr_inline > (PAGE_SIZE - sizeof(*req)) / sizeof(struct page *))
+		nr_inline = 0;
+
+	req = kzalloc(sizeof(*req) + sizeof(struct page *) * nr_inline,
+		      GFP_KERNEL);
+	if (!req)
+		return ERR_PTR(-ENOMEM);
+
+	refcount_set(&req->usage, 1);
+	req->nr_pages = nr_pages;
+	req->actual_len = i_size; /* May change */
+	req->len = nr_pages * PAGE_SIZE; /* We can ask for more than there is */
+	req->data_version = dvnode->status.data_version; /* May change */
+	if (nr_inline > 0) {
+		req->pages = req->array;
+	} else {
+		req->pages = kcalloc(nr_pages, sizeof(struct page *),
+				     GFP_KERNEL);
+		if (!req->pages)
+			goto error;
 	}
-	return page;
 
-fail:
-	afs_dir_put_page(page);
-	_leave(" = -EIO");
-	return ERR_PTR(-EIO);
-}
+	/* Get a list of all the pages that hold or will hold the directory
+	 * content.  We need to fill in any gaps that we might find where the
+	 * memory reclaimer has been at work.  If there are any gaps, we will
+	 * need to reread the entire directory contents.
+	 */
+	i = 0;
+	do {
+		n = find_get_pages_contig(dvnode->vfs_inode.i_mapping, i,
+					  req->nr_pages - i,
+					  req->pages + i);
+		_debug("find %u at %u/%u", n, i, req->nr_pages);
+		if (n == 0) {
+			gfp_t gfp = dvnode->vfs_inode.i_mapping->gfp_mask;
+
+			if (test_and_clear_bit(AFS_VNODE_DIR_VALID, &dvnode->flags))
+				afs_stat_v(dvnode, n_inval);
+
+			ret = -ENOMEM;
+			req->pages[i] = __page_cache_alloc(gfp);
+			if (!req->pages[i])
+				goto error;
+			ret = add_to_page_cache_lru(req->pages[i],
+						    dvnode->vfs_inode.i_mapping,
+						    i, gfp);
+			if (ret < 0)
+				goto error;
+
+			set_page_private(req->pages[i], 1);
+			SetPagePrivate(req->pages[i]);
+			unlock_page(req->pages[i]);
+			i++;
+		} else {
+			i += n;
+		}
+	} while (i < req->nr_pages);
 
-/*
- * open an AFS directory file
- */
-static int afs_dir_open(struct inode *inode, struct file *file)
-{
-	_enter("{%lu}", inode->i_ino);
+	/* If we're going to reload, we need to lock all the pages to prevent
+	 * races.
+	 */
+	if (!test_bit(AFS_VNODE_DIR_VALID, &dvnode->flags)) {
+		ret = -ERESTARTSYS;
+		for (i = 0; i < req->nr_pages; i++)
+			if (lock_page_killable(req->pages[i]) < 0)
+				goto error_unlock;
 
-	BUILD_BUG_ON(sizeof(union afs_dir_block) != 2048);
-	BUILD_BUG_ON(sizeof(union afs_dirent) != 32);
+		if (test_bit(AFS_VNODE_DIR_VALID, &dvnode->flags))
+			goto success;
 
-	if (test_bit(AFS_VNODE_DELETED, &AFS_FS_I(inode)->flags))
-		return -ENOENT;
+		ret = afs_fetch_data(dvnode, key, req);
+		if (ret < 0)
+			goto error_unlock_all;
 
-	return afs_open(inode, file);
+		task_io_account_read(PAGE_SIZE * req->nr_pages);
+
+		if (req->len < req->file_size)
+			goto content_has_grown;
+
+		/* Validate the data we just read. */
+		ret = -EIO;
+		for (i = 0; i < req->nr_pages; i++)
+			if (!afs_dir_check_page(dvnode, req->pages[i],
+						req->actual_len))
+				goto error_unlock_all;
+
+		// TODO: Trim excess pages
+
+		set_bit(AFS_VNODE_DIR_VALID, &dvnode->flags);
+	}
+
+success:
+	i = req->nr_pages;
+	while (i > 0)
+		unlock_page(req->pages[--i]);
+	return req;
+
+error_unlock_all:
+	i = req->nr_pages;
+error_unlock:
+	while (i > 0)
+		unlock_page(req->pages[--i]);
+error:
+	afs_put_read(req);
+	_leave(" = %d", ret);
+	return ERR_PTR(ret);
+
+content_has_grown:
+	i = req->nr_pages;
+	while (i > 0)
+		unlock_page(req->pages[--i]);
+	afs_put_read(req);
+	goto retry;
 }
 
 /*
  * deal with one block in an AFS directory
  */
 static int afs_dir_iterate_block(struct dir_context *ctx,
-				 union afs_dir_block *block,
+				 union afs_xdr_dir_block *block,
 				 unsigned blkoff)
 {
-	union afs_dirent *dire;
+	union afs_xdr_dirent *dire;
 	unsigned offset, next, curr;
 	size_t nlen;
 	int tmp;
 
 	_enter("%u,%x,%p,,",(unsigned)ctx->pos,blkoff,block);
 
-	curr = (ctx->pos - blkoff) / sizeof(union afs_dirent);
+	curr = (ctx->pos - blkoff) / sizeof(union afs_xdr_dirent);
 
 	/* walk through the block, an entry at a time */
-	for (offset = AFS_DIRENT_PER_BLOCK - block->pagehdr.nentries;
-	     offset < AFS_DIRENT_PER_BLOCK;
+	for (offset = (blkoff == 0 ? AFS_DIR_RESV_BLOCKS0 : AFS_DIR_RESV_BLOCKS);
+	     offset < AFS_DIR_SLOTS_PER_BLOCK;
 	     offset = next
 	     ) {
 		next = offset + 1;
 
 		/* skip entries marked unused in the bitmap */
-		if (!(block->pagehdr.bitmap[offset / 8] &
+		if (!(block->hdr.bitmap[offset / 8] &
 		      (1 << (offset % 8)))) {
 			_debug("ENT[%zu.%u]: unused",
-			       blkoff / sizeof(union afs_dir_block), offset);
+			       blkoff / sizeof(union afs_xdr_dir_block), offset);
 			if (offset >= curr)
 				ctx->pos = blkoff +
-					next * sizeof(union afs_dirent);
+					next * sizeof(union afs_xdr_dirent);
 			continue;
 		}
 
@@ -291,34 +353,34 @@ static int afs_dir_iterate_block(struct dir_context *ctx,
 		dire = &block->dirents[offset];
 		nlen = strnlen(dire->u.name,
 			       sizeof(*block) -
-			       offset * sizeof(union afs_dirent));
+			       offset * sizeof(union afs_xdr_dirent));
 
 		_debug("ENT[%zu.%u]: %s %zu \"%s\"",
-		       blkoff / sizeof(union afs_dir_block), offset,
+		       blkoff / sizeof(union afs_xdr_dir_block), offset,
 		       (offset < curr ? "skip" : "fill"),
 		       nlen, dire->u.name);
 
 		/* work out where the next possible entry is */
-		for (tmp = nlen; tmp > 15; tmp -= sizeof(union afs_dirent)) {
-			if (next >= AFS_DIRENT_PER_BLOCK) {
+		for (tmp = nlen; tmp > 15; tmp -= sizeof(union afs_xdr_dirent)) {
+			if (next >= AFS_DIR_SLOTS_PER_BLOCK) {
 				_debug("ENT[%zu.%u]:"
 				       " %u travelled beyond end dir block"
 				       " (len %u/%zu)",
-				       blkoff / sizeof(union afs_dir_block),
+				       blkoff / sizeof(union afs_xdr_dir_block),
 				       offset, next, tmp, nlen);
 				return -EIO;
 			}
-			if (!(block->pagehdr.bitmap[next / 8] &
+			if (!(block->hdr.bitmap[next / 8] &
 			      (1 << (next % 8)))) {
 				_debug("ENT[%zu.%u]:"
 				       " %u unmarked extension (len %u/%zu)",
-				       blkoff / sizeof(union afs_dir_block),
+				       blkoff / sizeof(union afs_xdr_dir_block),
 				       offset, next, tmp, nlen);
 				return -EIO;
 			}
 
 			_debug("ENT[%zu.%u]: ext %u/%zu",
-			       blkoff / sizeof(union afs_dir_block),
+			       blkoff / sizeof(union afs_xdr_dir_block),
 			       next, tmp, nlen);
 			next++;
 		}
@@ -330,13 +392,14 @@ static int afs_dir_iterate_block(struct dir_context *ctx,
 		/* found the next entry */
 		if (!dir_emit(ctx, dire->u.name, nlen,
 			      ntohl(dire->u.vnode),
-			      ctx->actor == afs_lookup_filldir ?
+			      (ctx->actor == afs_lookup_filldir ||
+			       ctx->actor == afs_lookup_one_filldir)?
 			      ntohl(dire->u.unique) : DT_UNKNOWN)) {
 			_leave(" = 0 [full]");
 			return 0;
 		}
 
-		ctx->pos = blkoff + next * sizeof(union afs_dirent);
+		ctx->pos = blkoff + next * sizeof(union afs_xdr_dirent);
 	}
 
 	_leave(" = 1 [more]");
@@ -349,8 +412,10 @@ static int afs_dir_iterate_block(struct dir_context *ctx,
 static int afs_dir_iterate(struct inode *dir, struct dir_context *ctx,
 			   struct key *key)
 {
-	union afs_dir_block *dblock;
-	struct afs_dir_page *dbuf;
+	struct afs_vnode *dvnode = AFS_FS_I(dir);
+	struct afs_xdr_dir_page *dbuf;
+	union afs_xdr_dir_block *dblock;
+	struct afs_read *req;
 	struct page *page;
 	unsigned blkoff, limit;
 	int ret;
@@ -362,45 +427,53 @@ static int afs_dir_iterate(struct inode *dir, struct dir_context *ctx,
 		return -ESTALE;
 	}
 
+	req = afs_read_dir(dvnode, key);
+	if (IS_ERR(req))
+		return PTR_ERR(req);
+
 	/* round the file position up to the next entry boundary */
-	ctx->pos += sizeof(union afs_dirent) - 1;
-	ctx->pos &= ~(sizeof(union afs_dirent) - 1);
+	ctx->pos += sizeof(union afs_xdr_dirent) - 1;
+	ctx->pos &= ~(sizeof(union afs_xdr_dirent) - 1);
 
 	/* walk through the blocks in sequence */
 	ret = 0;
-	while (ctx->pos < dir->i_size) {
-		blkoff = ctx->pos & ~(sizeof(union afs_dir_block) - 1);
+	while (ctx->pos < req->actual_len) {
+		blkoff = ctx->pos & ~(sizeof(union afs_xdr_dir_block) - 1);
 
-		/* fetch the appropriate page from the directory */
-		page = afs_dir_get_page(dir, blkoff / PAGE_SIZE, key);
-		if (IS_ERR(page)) {
-			ret = PTR_ERR(page);
+		/* Fetch the appropriate page from the directory and re-add it
+		 * to the LRU.
+		 */
+		page = req->pages[blkoff / PAGE_SIZE];
+		if (!page) {
+			ret = -EIO;
 			break;
 		}
+		mark_page_accessed(page);
 
 		limit = blkoff & ~(PAGE_SIZE - 1);
 
-		dbuf = page_address(page);
+		dbuf = kmap(page);
 
 		/* deal with the individual blocks stashed on this page */
 		do {
 			dblock = &dbuf->blocks[(blkoff % PAGE_SIZE) /
-					       sizeof(union afs_dir_block)];
+					       sizeof(union