From: Trond Myklebust Subject: [PATCH 09/15] NFS: Remove BKL from NFS lookup code Date: Thu, 12 Jun 2008 15:28:17 -0400 Message-ID: <20080612192817.24916.44975.stgit@localhost.localdomain> References: <20080612192817.24916.52065.stgit@localhost.localdomain> Mime-Version: 1.0 Content-Type: text/plain; charset="utf-8" Cc: Trond Myklebust To: linux-nfs@vger.kernel.org Return-path: Received: from mx2.netapp.com ([216.240.18.37]:62801 "EHLO mx2.netapp.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753047AbYFLTtZ (ORCPT ); Thu, 12 Jun 2008 15:49:25 -0400 Received: from svlexrs02.hq.netapp.com (svlexrs02.corp.netapp.com [10.57.156.154]) by smtp2.corp.netapp.com (8.13.1/8.13.1/NTAP-1.6) with ESMTP id m5CJnN15006182 for ; Thu, 12 Jun 2008 12:49:24 -0700 (PDT) In-Reply-To: <20080612192817.24916.52065.stgit-bi+AKbBUZKY6gyzm1THtWbp2dZbC/Bob@public.gmane.org> Sender: linux-nfs-owner@vger.kernel.org List-ID: All dentry-related operations are already BKL-safe, since they are protected by the VFS locking. No extra locks should be needed in the NFS code. In the case of nfs_revalidate_inode(), we're only doing an attribute update (protected by the inode->i_lock). In the case of nfs_lookup(), we're instantiating a new dentry, so there should be no contention possible until after we call d_materialise_unique. Signed-off-by: Trond Myklebust --- fs/nfs/dir.c | 9 +-------- 1 files changed, 1 insertions(+), 8 deletions(-) diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c index cdcc0c3..c936102 100644 --- a/fs/nfs/dir.c +++ b/fs/nfs/dir.c @@ -777,7 +777,6 @@ static int nfs_lookup_revalidate(struct dentry * dentry, struct nameidata *nd) struct nfs_fattr fattr; parent = dget_parent(dentry); - lock_kernel(); dir = parent->d_inode; nfs_inc_stats(dir, NFSIOS_DENTRYREVALIDATE); inode = dentry->d_inode; @@ -815,7 +814,6 @@ static int nfs_lookup_revalidate(struct dentry * dentry, struct nameidata *nd) nfs_set_verifier(dentry, nfs_save_change_attribute(dir)); out_valid: - unlock_kernel(); dput(parent); dfprintk(LOOKUPCACHE, "NFS: %s(%s/%s) is valid\n", __func__, dentry->d_parent->d_name.name, @@ -834,7 +832,6 @@ out_zap_parent: shrink_dcache_parent(dentry); } d_drop(dentry); - unlock_kernel(); dput(parent); dfprintk(LOOKUPCACHE, "NFS: %s(%s/%s) is invalid\n", __func__, dentry->d_parent->d_name.name, @@ -921,8 +918,6 @@ static struct dentry *nfs_lookup(struct inode *dir, struct dentry * dentry, stru res = ERR_PTR(-ENOMEM); dentry->d_op = NFS_PROTO(dir)->dentry_ops; - lock_kernel(); - /* * If we're doing an exclusive create, optimize away the lookup * but don't hash the dentry. @@ -930,7 +925,7 @@ static struct dentry *nfs_lookup(struct inode *dir, struct dentry * dentry, stru if (nfs_is_exclusive_create(dir, nd)) { d_instantiate(dentry, NULL); res = NULL; - goto out_unlock; + goto out; } parent = dentry->d_parent; @@ -958,8 +953,6 @@ no_entry: nfs_set_verifier(dentry, nfs_save_change_attribute(dir)); out_unblock_sillyrename: nfs_unblock_sillyrename(parent); -out_unlock: - unlock_kernel(); out: return res; }