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

Pull parallel filesystem directory handling update from Al Viro.

This is the main parallel directory work by Al that makes the vfs layer
able to do lookup and readdir in parallel within a single directory.
That's a big change, since this used to be all protected by the
directory inode mutex.

The inode mutex is replaced by an rwsem, and serialization of lookups of
a single name is done by a "in-progress" dentry marker.

The series begins with xattr cleanups, and then ends with switching
filesystems over to actually doing the readdir in parallel (switching to
the "iterate_shared()" that only takes the read lock).

A more detailed explanation of the process from Al Viro:
 "The xattr work starts with some acl fixes, then switches ->getxattr to
  passing inode and dentry separately.  This is the point where the
  things start to get tricky - that got merged into the very beginning
  of the -rc3-based #work.lookups, to allow untangling the
  security_d_instantiate() mess.  The xattr work itself proceeds to
  switch a lot of filesystems to generic_...xattr(); no complications
  there.

  After that initial xattr work, the series then does the following:

   - untangle security_d_instantiate()

   - convert a bunch of open-coded lookup_one_len_unlocked() to calls of
     that thing; one such place (in overlayfs) actually yields a trivial
     conflict with overlayfs fixes later in the cycle - overlayfs ended
     up switching to a variant of lookup_one_len_unlocked() sans the
     permission checks.  I would've dropped that commit (it gets
     overridden on merge from #ovl-fixes in #for-next; proper resolution
     is to use the variant in mainline fs/overlayfs/super.c), but I
     didn't want to rebase the damn thing - it was fairly late in the
     cycle...

   - some filesystems had managed to depend on lookup/lookup exclusion
     for *fs-internal* data structures in a way that would break if we
     relaxed the VFS exclusion.  Fixing hadn't been hard, fortunately.

   - core of that series - parallel lookup machinery, replacing
     ->i_mutex with rwsem, making lookup_slow() take it only shared.  At
     that point lookups happen in parallel; lookups on the same name
     wait for the in-progress one to be done with that dentry.

     Surprisingly little code, at that - almost all of it is in
     fs/dcache.c, with fs/namei.c changes limited to lookup_slow() -
     making it use the new primitive and actually switching to locking
     shared.

   - parallel readdir stuff - first of all, we provide the exclusion on
     per-struct file basis, same as we do for read() vs lseek() for
     regular files.  That takes care of most of the needed exclusion in
     readdir/readdir; however, these guys are trickier than lookups, so
     I went for switching them one-by-one.  To do that, a new method
     '->iterate_shared()' is added and filesystems are switched to it
     as they are either confirmed to be OK with shared lock on directory
     or fixed to be OK with that.  I hope to kill the original method
     come next cycle (almost all in-tree filesystems are switched
     already), but it's still not quite finished.

   - several filesystems get switched to parallel readdir.  The
     interesting part here is dealing with dcache preseeding by readdir;
     that needs minor adjustment to be safe with directory locked only
     shared.

     Most of the filesystems doing that got switched to in those
     commits.  Important exception: NFS.  Turns out that NFS folks, with
     their, er, insistence on VFS getting the fuck out of the way of the
     Smart Filesystem Code That Knows How And What To Lock(tm) have
     grown the locking of their own.  They had their own homegrown
     rwsem, with lookup/readdir/atomic_open being *writers* (sillyunlink
     is the reader there).  Of course, with VFS getting the fuck out of
     the way, as requested, the actual smarts of the smart filesystem
     code etc. had become exposed...

   - do_last/lookup_open/atomic_open cleanups.  As the result, open()
     without O_CREAT locks the directory only shared.  Including the
     ->atomic_open() case.  Backmerge from #for-linus in the middle of
     that - atomic_open() fix got brought in.

   - then comes NFS switch to saner (VFS-based ;-) locking, killing the
     homegrown "lookup and readdir are writers" kinda-sorta rwsem.  All
     exclusion for sillyunlink/lookup is done by the parallel lookups
     mechanism.  Exclusion between sillyunlink and rmdir is a real rwsem
     now - rmdir being the writer.

     Result: NFS lookups/readdirs/O_CREAT-less opens happen in parallel
     now.

   - the rest of the series consists of switching a lot of filesystems
     to parallel readdir; in a lot of cases ->llseek() gets simplified
     as well.  One backmerge in there (again, #for-linus - rockridge
     fix)"

* 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs: (74 commits)
  ext4: switch to ->iterate_shared()
  hfs: switch to ->iterate_shared()
  hfsplus: switch to ->iterate_shared()
  hostfs: switch to ->iterate_shared()
  hpfs: switch to ->iterate_shared()
  hpfs: handle allocation failures in hpfs_add_pos()
  gfs2: switch to ->iterate_shared()
  f2fs: switch to ->iterate_shared()
  afs: switch to ->iterate_shared()
  befs: switch to ->iterate_shared()
  befs: constify stuff a bit
  isofs: switch to ->iterate_shared()
  get_acorn_filename(): deobfuscate a bit
  btrfs: switch to ->iterate_shared()
  logfs: no need to lock directory in lseek
  switch ecryptfs to ->iterate_shared
  9p: switch to ->iterate_shared()
  fat: switch to ->iterate_shared()
  romfs, squashfs: switch to ->iterate_shared()
  more trivial ->iterate_shared conversions
  ...

7 files changed, 147 insertions, 190 deletions

diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index 33eb81738d03..aaf7bd0cbae2 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -57,7 +57,7 @@ static void nfs_readdir_clear_array(struct page*);
 const struct file_operations nfs_dir_operations = {
 	.llseek		= nfs_llseek_dir,
 	.read		= generic_read_dir,
-	.iterate	= nfs_readdir,
+	.iterate_shared	= nfs_readdir,
 	.open		= nfs_opendir,
 	.release	= nfs_closedir,
 	.fsync		= nfs_fsync_dir,
@@ -145,6 +145,7 @@ struct nfs_cache_array_entry {
 };
 
 struct nfs_cache_array {
+	atomic_t refcount;
 	int size;
 	int eof_index;
 	u64 last_cookie;
@@ -200,11 +201,20 @@ void nfs_readdir_clear_array(struct page *page)
 	int i;
 
 	array = kmap_atomic(page);
-	for (i = 0; i < array->size; i++)
-		kfree(array->array[i].string.name);
+	if (atomic_dec_and_test(&array->refcount))
+		for (i = 0; i < array->size; i++)
+			kfree(array->array[i].string.name);
 	kunmap_atomic(array);
 }
 
+static bool grab_page(struct page *page)
+{
+	struct nfs_cache_array *array = kmap_atomic(page);
+	bool res = atomic_inc_not_zero(&array->refcount);
+	kunmap_atomic(array);
+	return res;
+}
+
 /*
  * the caller is responsible for freeing qstr.name
  * when called by nfs_readdir_add_to_array, the strings will be freed in
@@ -470,6 +480,7 @@ static
 void nfs_prime_dcache(struct dentry *parent, struct nfs_entry *entry)
 {
 	struct qstr filename = QSTR_INIT(entry->name, entry->len);
+	DECLARE_WAIT_QUEUE_HEAD_ONSTACK(wq);
 	struct dentry *dentry;
 	struct dentry *alias;
 	struct inode *dir = d_inode(parent);
@@ -489,7 +500,13 @@ void nfs_prime_dcache(struct dentry *parent, struct nfs_entry *entry)
 	filename.hash = full_name_hash(filename.name, filename.len);
 
 	dentry = d_lookup(parent, &filename);
-	if (dentry != NULL) {
+again:
+	if (!dentry) {
+		dentry = d_alloc_parallel(parent, &filename, &wq);
+		if (IS_ERR(dentry))
+			return;
+	}
+	if (!d_in_lookup(dentry)) {
 		/* Is there a mountpoint here? If so, just exit */
 		if (!nfs_fsid_equal(&NFS_SB(dentry->d_sb)->fsid,
 					&entry->fattr->fsid))
@@ -503,26 +520,21 @@ void nfs_prime_dcache(struct dentry *parent, struct nfs_entry *entry)
 		} else {
 			d_invalidate(dentry);
 			dput(dentry);
+			dentry = NULL;
+			goto again;
 		}
 	}
 
-	dentry = d_alloc(parent, &filename);
-	if (dentry == NULL)
-		return;
-
 	inode = nfs_fhget(dentry->d_sb, entry->fh, entry->fattr, entry->label);
-	if (IS_ERR(inode))
-		goto out;
-
 	alias = d_splice_alias(inode, dentry);
-	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));
-
+	d_lookup_done(dentry);
+	if (alias) {
+		if (IS_ERR(alias))
+			goto out;
+		dput(dentry);
+		dentry = alias;
+	}
+	nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
 out:
 	dput(dentry);
 }
@@ -643,6 +655,7 @@ int nfs_readdir_xdr_to_array(nfs_readdir_descriptor_t *desc, struct page *page,
 		goto out_label_free;
 	}
 	memset(array, 0, sizeof(struct nfs_cache_array));
+	atomic_set(&array->refcount, 1);
 	array->eof_index = -1;
 
 	status = nfs_readdir_alloc_pages(pages, array_size);
@@ -705,8 +718,7 @@ int nfs_readdir_filler(nfs_readdir_descriptor_t *desc, struct page* page)
 static
 void cache_page_release(nfs_readdir_descriptor_t *desc)
 {
-	if (!desc->page->mapping)
-		nfs_readdir_clear_array(desc->page);
+	nfs_readdir_clear_array(desc->page);
 	put_page(desc->page);
 	desc->page = NULL;
 }
@@ -714,8 +726,16 @@ void cache_page_release(nfs_readdir_descriptor_t *desc)
 static
 struct page *get_cache_page(nfs_readdir_descriptor_t *desc)
 {
-	return read_cache_page(file_inode(desc->file)->i_mapping,
+	struct page *page;
+
+	for (;;) {
+		page = read_cache_page(file_inode(desc->file)->i_mapping,
 			desc->page_index, (filler_t *)nfs_readdir_filler, desc);
+		if (IS_ERR(page) || grab_page(page))
+			break;
+		put_page(page);
+	}
+	return page;
 }
 
 /*
@@ -889,7 +909,6 @@ static int nfs_readdir(struct file *file, struct dir_context *ctx)
 	desc->decode = NFS_PROTO(inode)->decode_dirent;
 	desc->plus = nfs_use_readdirplus(inode, ctx) ? 1 : 0;
 
-	nfs_block_sillyrename(dentry);
 	if (ctx->pos == 0 || nfs_dir_mapping_need_revalidate(inode))
 		res = nfs_revalidate_mapping(inode, file->f_mapping);
 	if (res < 0)
@@ -925,7 +944,6 @@ static int nfs_readdir(struct file *file, struct dir_context *ctx)
 			break;
 	} while (!desc->eof);
 out:
-	nfs_unblock_sillyrename(dentry);
 	if (res > 0)
 		res = 0;
 	dfprintk(FILE, "NFS: readdir(%pD2) returns %d\n", file, res);
@@ -934,13 +952,11 @@ out:
 
 static loff_t nfs_llseek_dir(struct file *filp, loff_t offset, int whence)
 {
-	struct inode *inode = file_inode(filp);
 	struct nfs_open_dir_context *dir_ctx = filp->private_data;
 
 	dfprintk(FILE, "NFS: llseek dir(%pD2, %lld, %d)\n",
 			filp, offset, whence);
 
-	inode_lock(inode);
 	switch (whence) {
 		case 1:
 			offset += filp->f_pos;
@@ -948,16 +964,13 @@ static loff_t nfs_llseek_dir(struct file *filp, loff_t offset, int whence)
 			if (offset >= 0)
 				break;
 		default:
-			offset = -EINVAL;
-			goto out;
+			return -EINVAL;
 	}
 	if (offset != filp->f_pos) {
 		filp->f_pos = offset;
 		dir_ctx->dir_cookie = 0;
 		dir_ctx->duped = 0;
 	}
-out:
-	inode_unlock(inode);
 	return offset;
 }
 
@@ -1383,7 +1396,6 @@ struct dentry *nfs_lookup(struct inode *dir, struct dentry * dentry, unsigned in
 	parent = dentry->d_parent;
 	/* Protect against concurrent sillydeletes */
 	trace_nfs_lookup_enter(dir, dentry, flags);
-	nfs_block_sillyrename(parent);
 	error = NFS_PROTO(dir)->lookup(dir, &dentry->d_name, fhandle, fattr, label);
 	if (error == -ENOENT)
 		goto no_entry;
@@ -1408,7 +1420,6 @@ no_entry:
 	}
 	nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
 out_unblock_sillyrename:
-	nfs_unblock_sillyrename(parent);
 	trace_nfs_lookup_exit(dir, dentry, flags, error);
 	nfs4_label_free(label);
 out:
@@ -1520,9 +1531,7 @@ int nfs_atomic_open(struct inode *dir, struct dentry *dentry,
 		goto out;
 
 	trace_nfs_atomic_open_enter(dir, ctx, open_flags);
-	nfs_block_sillyrename(dentry->d_parent);
 	inode = NFS_PROTO(dir)->open_context(dir, ctx, open_flags, &attr, opened);
-	nfs_unblock_sillyrename(dentry->d_parent);
 	if (IS_ERR(inode)) {
 		err = PTR_ERR(inode);
 		trace_nfs_atomic_open_exit(dir, ctx, open_flags, err);
@@ -1766,7 +1775,7 @@ int nfs_rmdir(struct inode *dir, struct dentry *dentry)
 
 	trace_nfs_rmdir_enter(dir, dentry);
 	if (d_really_is_positive(dentry)) {
-		nfs_wait_on_sillyrename(dentry);
+		down_write(&NFS_I(d_inode(dentry))->rmdir_sem);
 		error = NFS_PROTO(dir)->rmdir(dir, &dentry->d_name);
 		/* Ensure the VFS deletes this inode */
 		switch (error) {
@@ -1776,6 +1785,7 @@ int nfs_rmdir(struct inode *dir, struct dentry *dentry)
 		case -ENOENT:
 			nfs_dentry_handle_enoent(dentry);
 		}
+		up_write(&NFS_I(d_inode(dentry))->rmdir_sem);
 	} else
 		error = NFS_PROTO(dir)->rmdir(dir, &dentry->d_name);
 	trace_nfs_rmdir_exit(dir, dentry, error);
diff --git a/fs/nfs/direct.c b/fs/nfs/direct.c
index c93826e4a8c6..438d4e70742f 100644
--- a/fs/nfs/direct.c
+++ b/fs/nfs/direct.c
@@ -396,7 +396,7 @@ static void nfs_direct_complete(struct nfs_direct_req *dreq, bool write)
 static void nfs_direct_readpage_release(struct nfs_page *req)
 {
 	dprintk("NFS: direct read done (%s/%llu %d@%lld)\n",
-		d_inode(req->wb_context->dentry)->i_sb->s_id,
+		req->wb_context->dentry->d_sb->s_id,
 		(unsigned long long)NFS_FILEID(d_inode(req->wb_context->dentry)),
 		req->wb_bytes,
 		(long long)req_offset(req));
diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c
index 738c84a42eb0..52e7d6869e3b 100644
--- a/fs/nfs/inode.c
+++ b/fs/nfs/inode.c
@@ -1958,9 +1958,7 @@ static void init_once(void *foo)
 	nfsi->nrequests = 0;
 	nfsi->commit_info.ncommit = 0;
 	atomic_set(&nfsi->commit_info.rpcs_out, 0);
-	atomic_set(&nfsi->silly_count, 1);
-	INIT_HLIST_HEAD(&nfsi->silly_list);
-	init_waitqueue_head(&nfsi->waitqueue);
+	init_rwsem(&nfsi->rmdir_sem);
 	nfs4_init_once(nfsi);
 }
 
diff --git a/fs/nfs/nfs3acl.c b/fs/nfs/nfs3acl.c
index 17c0fa1eccfa..720d92f5abfb 100644
--- a/fs/nfs/nfs3acl.c
+++ b/fs/nfs/nfs3acl.c
@@ -11,6 +11,38 @@
 
 #define NFSDBG_FACILITY	NFSDBG_PROC
 
+/*
+ * nfs3_prepare_get_acl, nfs3_complete_get_acl, nfs3_abort_get_acl: Helpers for
+ * caching get_acl results in a race-free way.  See fs/posix_acl.c:get_acl()
+ * for explanations.
+ */
+static void nfs3_prepare_get_acl(struct posix_acl **p)
+{
+	struct posix_acl *sentinel = uncached_acl_sentinel(current);
+
+	if (cmpxchg(p, ACL_NOT_CACHED, sentinel) != ACL_NOT_CACHED) {
+		/* Not the first reader or sentinel already in place. */
+	}
+}
+
+static void nfs3_complete_get_acl(struct posix_acl **p, struct posix_acl *acl)
+{
+	struct posix_acl *sentinel = uncached_acl_sentinel(current);
+
+	/* Only cache the ACL if our sentinel is still in place. */
+	posix_acl_dup(acl);
+	if (cmpxchg(p, sentinel, acl) != sentinel)
+		posix_acl_release(acl);
+}
+
+static void nfs3_abort_get_acl(struct posix_acl **p)
+{
+	struct posix_acl *sentinel = uncached_acl_sentinel(current);
+
+	/* Remove our sentinel upon failure. */
+	cmpxchg(p, sentinel, ACL_NOT_CACHED);
+}
+
 struct posix_acl *nfs3_get_acl(struct inode *inode, int type)
 {
 	struct nfs_server *server = NFS_SERVER(inode);
@@ -55,6 +87,11 @@ struct posix_acl *nfs3_get_acl(struct inode *inode, int type)
 	if (res.fattr == NULL)
 		return ERR_PTR(-ENOMEM);
 
+	if (args.mask & NFS_ACL)
+		nfs3_prepare_get_acl(&inode->i_acl);
+	if (args.mask & NFS_DFACL)
+		nfs3_prepare_get_acl(&inode->i_default_acl);
+
 	status = rpc_call_sync(server->client_acl, &msg, 0);
 	dprintk("NFS reply getacl: %d\n", status);
 
@@ -89,12 +126,12 @@ struct posix_acl *nfs3_get_acl(struct inode *inode, int type)
 	}
 
 	if (res.mask & NFS_ACL)
-		set_cached_acl(inode, ACL_TYPE_ACCESS, res.acl_access);
+		nfs3_complete_get_acl(&inode->i_acl, res.acl_access);
 	else
 		forget_cached_acl(inode, ACL_TYPE_ACCESS);
 
 	if (res.mask & NFS_DFACL)
-		set_cached_acl(inode, ACL_TYPE_DEFAULT, res.acl_default);
+		nfs3_complete_get_acl(&inode->i_default_acl, res.acl_default);
 	else
 		forget_cached_acl(inode, ACL_TYPE_DEFAULT);
 
@@ -108,6 +145,8 @@ struct posix_acl *nfs3_get_acl(struct inode *inode, int type)
 	}
 
 getout:
+	nfs3_abort_get_acl(&inode->i_acl);
+	nfs3_abort_get_acl(&inode->i_default_acl);
 	posix_acl_release(res.acl_access);
 	posix_acl_release(res.acl_default);
 	nfs_free_fattr(res.fattr);
diff --git a/fs/nfs/nfs4proc.c b/fs/nfs/nfs4proc.c
index 327b8c34d360..084e8570da18 100644
--- a/fs/nfs/nfs4proc.c
+++ b/fs/nfs/nfs4proc.c
@@ -3777,7 +3777,7 @@ static void nfs4_proc_unlink_setup(struct rpc_message *msg, struct inode *dir)
 
 static void nfs4_proc_unlink_rpc_prepare(struct rpc_task *task, struct nfs_unlinkdata *data)
 {
-	nfs4_setup_sequence(NFS_SERVER(data->dir),
+	nfs4_setup_sequence(NFS_SB(data->dentry->d_sb),
 			&data->args.seq_args,
 			&data->res.seq_res,
 			task);
@@ -6263,10 +6263,10 @@ static int nfs4_xattr_set_nfs4_acl(const struct xattr_handler *handler,
 }
 
 static int nfs4_xattr_get_nfs4_acl(const struct xattr_handler *handler,
-				   struct dentry *dentry, const char *key,
-				   void *buf, size_t buflen)
+				   struct dentry *unused, struct inode *inode,
+				   const char *key, void *buf, size_t buflen)
 {
-	return nfs4_proc_get_acl(d_inode(dentry), buf, buflen);
+	return nfs4_proc_get_acl(inode, buf, buflen);
 }
 
 static bool nfs4_xattr_list_nfs4_acl(struct dentry *dentry)
@@ -6288,11 +6288,11 @@ static int nfs4_xattr_set_nfs4_label(const struct xattr_handler *handler,
 }
 
 static int nfs4_xattr_get_nfs4_label(const struct xattr_handler *handler,
-				     struct dentry *dentry, const char *key,
-				     void *buf, size_t buflen)
+				     struct dentry *unused, struct inode *inode,
+				     const char *key, void *buf, size_t buflen)
 {
 	if (security_ismaclabel(key))
-		return nfs4_get_security_label(d_inode(dentry), buf, buflen);
+		return nfs4_get_security_label(inode, buf, buflen);
 	return -EOPNOTSUPP;
 }
 
diff --git a/fs/nfs/nfstrace.h b/fs/nfs/nfstrace.h
index 9f80a086b612..0b9e5cc9a747 100644
--- a/fs/nfs/nfstrace.h
+++ b/fs/nfs/nfstrace.h
@@ -702,7 +702,7 @@ TRACE_EVENT(nfs_sillyrename_unlink,
 		),
 
 		TP_fast_assign(
-			struct inode *dir = data->dir;
+			struct inode *dir = d_inode(data->dentry->d_parent);
 			size_t len = data->args.name.len;
 			__entry->dev = dir->i_sb->s_dev;
 			__entry->dir = NFS_FILEID(dir);
diff --git a/fs/nfs/unlink.c b/fs/nfs/unlink.c
index fa538b2ba251..1868246f56e6 100644
--- a/fs/nfs/unlink.c
+++ b/fs/nfs/unlink.c
@@ -30,45 +30,11 @@
 static void
 nfs_free_unlinkdata(struct nfs_unlinkdata *data)
 {
-	iput(data->dir);
 	put_rpccred(data->cred);
 	kfree(data->args.name.name);
 	kfree(data);
 }
 
-#define NAME_ALLOC_LEN(len)	((len+16) & ~15)
-/**
- * nfs_copy_dname - copy dentry name to data structure
- * @dentry: pointer to dentry
- * @data: nfs_unlinkdata
- */
-static int nfs_copy_dname(struct dentry *dentry, struct nfs_unlinkdata *data)
-{
-	char		*str;
-	int		len = dentry->d_name.len;
-
-	str = kmemdup(dentry->d_name.name, NAME_ALLOC_LEN(len), GFP_KERNEL);
-	if (!str)
-		return -ENOMEM;
-	data->args.name.len = len;
-	data->args.name.name = str;
-	return 0;
-}
-
-static void nfs_free_dname(struct nfs_unlinkdata *data)
-{
-	kfree(data->args.name.name);
-	data->args.name.name = NULL;
-	data->args.name.len = 0;
-}
-
-static void nfs_dec_sillycount(struct inode *dir)
-{
-	struct nfs_inode *nfsi = NFS_I(dir);
-	if (atomic_dec_return(&nfsi->silly_count) == 1)
-		wake_up(&nfsi->waitqueue);
-}
-
 /**
  * nfs_async_unlink_done - Sillydelete post-processing
  * @task: rpc_task of the sillydelete
@@ -78,7 +44,7 @@ static void nfs_dec_sillycount(struct inode *dir)
 static void nfs_async_unlink_done(struct rpc_task *task, void *calldata)
 {
 	struct nfs_unlinkdata *data = calldata;
-	struct inode *dir = data->dir;
+	struct inode *dir = d_inode(data->dentry->d_parent);
 
 	trace_nfs_sillyrename_unlink(data, task->tk_status);
 	if (!NFS_PROTO(dir)->unlink_done(task, dir))
@@ -95,17 +61,21 @@ static void nfs_async_unlink_done(struct rpc_task *task, void *calldata)
 static void nfs_async_unlink_release(void *calldata)
 {
 	struct nfs_unlinkdata	*data = calldata;
-	struct super_block *sb = data->dir->i_sb;
+	struct dentry *dentry = data->dentry;
+	struct super_block *sb = dentry->d_sb;
 
-	nfs_dec_sillycount(data->dir);
+	up_read_non_owner(&NFS_I(d_inode(dentry->d_parent))->rmdir_sem);
+	d_lookup_done(dentry);
 	nfs_free_unlinkdata(data);
+	dput(dentry);
 	nfs_sb_deactive(sb);
 }
 
 static void nfs_unlink_prepare(struct rpc_task *task, void *calldata)
 {
 	struct nfs_unlinkdata *data = calldata;
-	NFS_PROTO(data->dir)->unlink_rpc_prepare(task, data);
+	struct inode *dir = d_inode(data->dentry->d_parent);
+	NFS_PROTO(dir)->unlink_rpc_prepare(task, data);
 }
 
 static const struct rpc_call_ops nfs_unlink_ops = {
@@ -114,7 +84,7 @@ static const struct rpc_call_ops nfs_unlink_ops = {
 	.rpc_call_prepare = nfs_unlink_prepare,
 };
 
-static int nfs_do_call_unlink(struct dentry *parent, struct inode *dir, struct nfs_unlinkdata *data)
+static void nfs_do_call_unlink(struct nfs_unlinkdata *data)
 {
 	struct rpc_message msg = {
 		.rpc_argp = &data->args,
@@ -129,10 +99,31 @@ static int nfs_do_call_unlink(struct dentry *parent, struct inode *dir, struct n
 		.flags = RPC_TASK_ASYNC,
 	};
 	struct rpc_task *task;
+	struct inode *dir = d_inode(data->dentry->d_parent);
+	nfs_sb_active(dir->i_sb);
+	data->args.fh = NFS_FH(dir);
+	nfs_fattr_init(data->res.dir_attr);
+
+	NFS_PROTO(dir)->unlink_setup(&msg, dir);
+
+	task_setup_data.rpc_client = NFS_CLIENT(dir);
+	task = rpc_run_task(&task_setup_data);
+	if (!IS_ERR(task))
+		rpc_put_task_async(task);
+}
+
+static int nfs_call_unlink(struct dentry *dentry, struct nfs_unlinkdata *data)
+{
+	struct inode *dir = d_inode(dentry->d_parent);
 	struct dentry *alias;
 
-	alias = d_lookup(parent, &data->args.name);
-	if (alias != NULL) {
+	down_read_non_owner(&NFS_I(dir)->rmdir_sem);
+	alias = d_alloc_parallel(dentry->d_parent, &data->args.name, &data->wq);
+	if (IS_ERR(alias)) {
+		up_read_non_owner(&NFS_I(dir)->rmdir_sem);
+		return 0;
+	}
+	if (!d_in_lookup(alias)) {
 		int ret;
 		void *devname_garbage = NULL;
 
@@ -140,10 +131,8 @@ static int nfs_do_call_unlink(struct dentry *parent, struct inode *dir, struct n
 		 * Hey, we raced with lookup... See if we need to transfer
 		 * the sillyrename information to the aliased dentry.
 		 */
-		nfs_free_dname(data);
-		ret = nfs_copy_dname(alias, data);
 		spin_lock(&alias->d_lock);
-		if (ret == 0 && d_really_is_positive(alias) &&
+		if (d_really_is_positive(alias) &&
 		    !(alias->d_flags & DCACHE_NFSFS_RENAMED)) {
 			devname_garbage = alias->d_fsdata;
 			alias->d_fsdata = data;
@@ -152,8 +141,8 @@ static int nfs_do_call_unlink(struct dentry *parent, struct inode *dir, struct n
 		} else
 			ret = 0;
 		spin_unlock(&alias->d_lock);
-		nfs_dec_sillycount(dir);
 		dput(alias);
+		up_read_non_owner(&NFS_I(dir)->rmdir_sem);
 		/*
 		 * If we'd displaced old cached devname, free it.  At that
 		 * point dentry is definitely not a root, so we won't need
@@ -162,94 +151,18 @@ static int nfs_do_call_unlink(struct dentry *parent, struct inode *dir, struct n
 		kfree(devname_garbage);
 		return ret;
 	}
-	data->dir = igrab(dir);
-	if (!data->dir) {
-		nfs_dec_sillycount(dir);
-		return 0;
-	}
-	nfs_sb_active(dir->i_sb);
-	data->args.fh = NFS_FH(dir);
-	nfs_fattr_init(data->res.dir_attr);
-
-	NFS_PROTO(dir)->unlink_setup(&msg, dir);
-
-	task_setup_data.rpc_client = NFS_CLIENT(dir);
-	task = rpc_run_task(&task_setup_data);
-	if (!IS_ERR(task))
-		rpc_put_task_async(task);
+	data->dentry = alias;
+	nfs_do_call_unlink(data);
 	return 1;
 }
 
-static int nfs_call_unlink(struct dentry *dentry, struct nfs_unlinkdata *data)
-{
-	struct dentry *parent;
-	struct inode *dir;
-	int ret = 0;
-
-
-	parent = dget_parent(dentry);
-	if (parent == NULL)
-		goto out_free;
-	dir = d_inode(parent);
-	/* Non-exclusive lock protects against concurrent lookup() calls */
-	spin_lock(&dir->i_lock);
-	if (atomic_inc_not_zero(&NFS_I(dir)->silly_count) == 0) {
-		/* Deferred delete */
-		hlist_add_head(&data->list, &NFS_I(dir)->silly_list);
-		spin_unlock(&dir->i_lock);
-		ret = 1;
-		goto out_dput;
-	}
-	spin_unlock(&dir->i_lock);
-	ret = nfs_do_call_unlink(parent, dir, data);
-out_dput:
-	dput(parent);
-out_free:
-	return ret;
-}
-
-void nfs_wait_on_sillyrename(struct dentry *dentry)
-{
-	struct nfs_inode *nfsi = NFS_I(d_inode(dentry));
-
-	wait_event(nfsi->waitqueue, atomic_read(&nfsi->silly_count) <= 1);
-}
-
-void nfs_block_sillyrename(struct dentry *dentry)
-{
-	struct nfs_inode *nfsi = NFS_I(d_inode(dentry));
-
-	wait_event(nfsi->waitqueue, atomic_cmpxchg(&nfsi->silly_count, 1, 0) == 1);
-}
-
-void nfs_unblock_sillyrename(struct dentry *dentry)
-{
-	struct inode *dir = d_inode(dentry);
-	struct nfs_inode *nfsi = NFS_I(dir);
-	struct nfs_unlinkdata *data;
-
-	atomic_inc(&nfsi->silly_count);
-	spin_lock(&dir->i_lock);
-	while (!hlist_empty(&nfsi->silly_list)) {
-		if (!atomic_inc_not_zero(&nfsi->silly_count))
-			break;
-		data = hlist_entry(nfsi->silly_list.first, struct nfs_unlinkdata, list);
-		hlist_del(&data->list);
-		spin_unlock(&dir->i_lock);
-		if (nfs_do_call_unlink(dentry, dir, data) == 0)
-			nfs_free_unlinkdata(data);
-		spin_lock(&dir->i_lock);
-	}
-	spin_unlock(&dir->i_lock);
-}
-
 /**
  * nfs_async_unlink - asynchronous unlinking of a file
  * @dir: parent directory of dentry
  * @dentry: dentry to unlink
  */
 static int
-nfs_async_unlink(struct inode *dir, struct dentry *dentry)
+nfs_async_unlink(struct dentry *dentry, struct qstr *name)
 {
 	struct nfs_unlinkdata *data;
 	int status = -ENOMEM;
@@ -258,13 +171,18 @@ nfs_async_unlink(struct inode *dir, struct dentry *dentry)
 	data = kzalloc(sizeof(*data), GFP_KERNEL);
 	if (data == NULL)
 		goto out;
+	data->args.name.name = kstrdup(name->name, GFP_KERNEL);
+	if (!data->args.name.name)
+		goto out_free;
+	data->args.name.len = name->len;
 
 	data->cred = rpc_lookup_cred();
 	if (IS_ERR(data->cred)) {
 		status = PTR_ERR(data->cred);
-		goto out_free;
+		goto out_free_name;
 	}
 	data->res.dir_attr = &data->dir_attr;
+	init_waitqueue_head(&data->wq);
 
 	status = -EBUSY;
 	spin_lock(&dentry->d_lock);
@@ -284,6 +202,8 @@ nfs_async_unlink(struct inode *dir, struct dentry *dentry)
 out_unlock:
 	spin_unlock(&dentry->d_lock);
 	put_rpccred(data->cred);
+out_free_name:
+	kfree(data->args.name.name);
 out_free:
 	kfree(data);
 out:
@@ -302,17 +222,15 @@ out:
 void
 nfs_complete_unlink(struct dentry *dentry, struct inode *inode)
 {
-	struct nfs_unlinkdata	*data = NULL;
+	struct nfs_unlinkdata	*data;
 
 	spin_lock(&dentry->d_lock);
-	if (dentry->d_flags & DCACHE_NFSFS_RENAMED) {
-		dentry->d_flags &= ~DCACHE_NFSFS_RENAMED;
-		data = dentry->d_fsdata;
-		dentry->d_fsdata = NULL;
-	}
+	dentry->d_flags &= ~DCACHE_NFSFS_RENAMED;
+	data = dentry->d_fsdata;
+	dentry->d_fsdata = NULL;
 	spin_unlock(&dentry->d_lock);
 
-	if (data != NULL && (NFS_STALE(inode) || !nfs_call_unlink(dentry, data)))
+	if (NFS_STALE(inode) || !nfs_call_unlink(dentry, data))
 		nfs_free_unlinkdata(data);
 }
 
@@ -559,18 +477,10 @@ nfs_sillyrename(struct inode *dir, struct dentry *dentry)
 	/* queue unlink first. Can't do this from rpc_release as it
 	 * has to allocate memory
 	 */
-	error = nfs_async_unlink(dir, dentry);
+	error = nfs_async_unlink(dentry, &sdentry->d_name);
 	if (error)
 		goto out_dput;
 
-	/* populate unlinkdata with the right dname */
-	error = nfs_copy_dname(sdentry,
-				(struct nfs_unlinkdata *)dentry->d_fsdata);
-	if (error) {
-		nfs_cancel_async_unlink(dentry);
-		goto out_dput;
-	}
-
 	/* run the rename task, undo unlink if it fails */
 	task = nfs_async_rename(dir, dir, dentry, sdentry,
 					nfs_complete_sillyrename);