diff options
| author | Trond Myklebust <Trond.Myklebust@netapp.com> | 2009-03-19 22:35:50 +0300 |
|---|---|---|
| committer | Trond Myklebust <Trond.Myklebust@netapp.com> | 2009-03-19 22:35:50 +0300 |
| commit | 7fe5c398fc2186ed586db11106a6692d871d0d58 (patch) | |
| tree | 1133bd775a23fb07ca759e4d6c44132c14576b75 /fs/nfs/file.c | |
| parent | b1e4adf4ea41bb8b5a7bfc1a7001f137e65495df (diff) | |
| download | linux-7fe5c398fc2186ed586db11106a6692d871d0d58.tar.xz | |
NFS: Optimise NFS close()
Close-to-open cache consistency rules really only require us to flush out
writes on calls to close(), and require us to revalidate attributes on the
very last close of the file.
Currently we appear to be doing a lot of extra attribute revalidation
and cache flushes.
Signed-off-by: Trond Myklebust <Trond.Myklebust@netapp.com>
Diffstat (limited to 'fs/nfs/file.c')
| -rw-r--r-- | fs/nfs/file.c | 11 |
1 files changed, 2 insertions, 9 deletions
diff --git a/fs/nfs/file.c b/fs/nfs/file.c index 1eab9c9ad242..d451073c4947 100644 --- a/fs/nfs/file.c +++ b/fs/nfs/file.c @@ -137,9 +137,6 @@ nfs_file_release(struct inode *inode, struct file *filp) dentry->d_parent->d_name.name, dentry->d_name.name); - /* Ensure that dirty pages are flushed out with the right creds */ - if (filp->f_mode & FMODE_WRITE) - nfs_wb_all(dentry->d_inode); nfs_inc_stats(inode, NFSIOS_VFSRELEASE); return nfs_release(inode, filp); } @@ -231,7 +228,6 @@ nfs_file_flush(struct file *file, fl_owner_t id) struct nfs_open_context *ctx = nfs_file_open_context(file); struct dentry *dentry = file->f_path.dentry; struct inode *inode = dentry->d_inode; - int status; dprintk("NFS: flush(%s/%s)\n", dentry->d_parent->d_name.name, @@ -241,11 +237,8 @@ nfs_file_flush(struct file *file, fl_owner_t id) return 0; nfs_inc_stats(inode, NFSIOS_VFSFLUSH); - /* Ensure that data+attribute caches are up to date after close() */ - status = nfs_do_fsync(ctx, inode); - if (!status) - nfs_revalidate_inode(NFS_SERVER(inode), inode); - return status; + /* Flush writes to the server and return any errors */ + return nfs_do_fsync(ctx, inode); } static ssize_t |