Merge tag 'nfs-for-3.11-2' of git://git.linux-nfs.org/projects/trondmy/linux-nfs

Pull second set of NFS client updates from Trond Myklebust:
 "This mainly contains some small readdir optimisations that had
  dependencies on Al Viro's readdir rewrite.  There is also a fix for a
  nasty deadlock which surfaced earlier in this merge window.

  Highlights include:
   - Fix an_rpc pipefs regression that causes a deadlock on mount
   - Readdir optimisations by Scott Mayhew and Jeff Layton
   - clean up the rpc_pipefs dentry operation setup"

* tag 'nfs-for-3.11-2' of git://git.linux-nfs.org/projects/trondmy/linux-nfs:
  SUNRPC: Fix a deadlock in rpc_client_register()
  rpc_pipe: rpc_dir_inode_operations can be static
  NFS: Allow nfs_updatepage to extend a write under additional circumstances
  NFS: Make nfs_readdir revalidate less often
  NFS: Make nfs_attribute_cache_expired() non-static
  rpc_pipe: set dentry operations at d_alloc time
  nfs: set verifier on existing dentries in nfs_prime_dcache

3 files changed, 28 insertions, 11 deletions

diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index 0fac2cb1ea18..e474ca2b2bfe 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -450,6 +450,7 @@ void nfs_prime_dcache(struct dentry *parent, struct nfs_entry *entry)
 	dentry = d_lookup(parent, &filename);
 	if (dentry != NULL) {
 		if (nfs_same_file(dentry, entry)) {
+			nfs_set_verifier(dentry, nfs_save_change_attribute(dir));
 			status = nfs_refresh_inode(dentry->d_inode, entry->fattr);
 			if (!status)
 				nfs_setsecurity(dentry->d_inode, entry->fattr, entry->label);
@@ -817,7 +818,7 @@ static int nfs_readdir(struct file *file, struct dir_context *ctx)
 	nfs_readdir_descriptor_t my_desc,
 			*desc = &my_desc;
 	struct nfs_open_dir_context *dir_ctx = file->private_data;
-	int res;
+	int res = 0;
 
 	dfprintk(FILE, "NFS: readdir(%s/%s) starting at cookie %llu\n",
 			dentry->d_parent->d_name.name, dentry->d_name.name,
@@ -839,7 +840,8 @@ static int nfs_readdir(struct file *file, struct dir_context *ctx)
 	desc->plus = nfs_use_readdirplus(inode, ctx) ? 1 : 0;
 
 	nfs_block_sillyrename(dentry);
-	res = nfs_revalidate_mapping(inode, file->f_mapping);
+	if (ctx->pos == 0 || nfs_attribute_cache_expired(inode))
+		res = nfs_revalidate_mapping(inode, file->f_mapping);
 	if (res < 0)
 		goto out;
 
diff --git a/fs/nfs/inode.c b/fs/nfs/inode.c
index c93639e6cf68..af6e806044d7 100644
--- a/fs/nfs/inode.c
+++ b/fs/nfs/inode.c
@@ -936,7 +936,7 @@ int nfs_attribute_timeout(struct inode *inode)
 	return !time_in_range_open(jiffies, nfsi->read_cache_jiffies, nfsi->read_cache_jiffies + nfsi->attrtimeo);
 }
 
-static int nfs_attribute_cache_expired(struct inode *inode)
+int nfs_attribute_cache_expired(struct inode *inode)
 {
 	if (nfs_have_delegated_attributes(inode))
 		return 0;
diff --git a/fs/nfs/write.c b/fs/nfs/write.c
index a2c7c28049d5..f1bdb7254776 100644
--- a/fs/nfs/write.c
+++ b/fs/nfs/write.c
@@ -888,6 +888,28 @@ out:
 	return PageUptodate(page) != 0;
 }
 
+/* If we know the page is up to date, and we're not using byte range locks (or
+ * if we have the whole file locked for writing), it may be more efficient to
+ * extend the write to cover the entire page in order to avoid fragmentation
+ * inefficiencies.
+ *
+ * If the file is opened for synchronous writes or if we have a write delegation
+ * from the server then we can just skip the rest of the checks.
+ */
+static int nfs_can_extend_write(struct file *file, struct page *page, struct inode *inode)
+{
+	if (file->f_flags & O_DSYNC)
+		return 0;
+	if (NFS_PROTO(inode)->have_delegation(inode, FMODE_WRITE))
+		return 1;
+	if (nfs_write_pageuptodate(page, inode) && (inode->i_flock == NULL ||
+			(inode->i_flock->fl_start == 0 &&
+			inode->i_flock->fl_end == OFFSET_MAX &&
+			inode->i_flock->fl_type != F_RDLCK)))
+		return 1;
+	return 0;
+}
+
 /*
  * Update and possibly write a cached page of an NFS file.
  *
@@ -908,14 +930,7 @@ int nfs_updatepage(struct file *file, struct page *page,
 		file->f_path.dentry->d_name.name, count,
 		(long long)(page_file_offset(page) + offset));
 
-	/* If we're not using byte range locks, and we know the page
-	 * is up to date, it may be more efficient to extend the write
-	 * to cover the entire page in order to avoid fragmentation
-	 * inefficiencies.
-	 */
-	if (nfs_write_pageuptodate(page, inode) &&
-			inode->i_flock == NULL &&
-			!(file->f_flags & O_DSYNC)) {
+	if (nfs_can_extend_write(file, page, inode)) {
 		count = max(count + offset, nfs_page_length(page));
 		offset = 0;
 	}