Merge branch 'vfs-scale-working' of git://git.kernel.org/pub/scm/linux/kernel/git/npiggin/linux-npiggin
* 'vfs-scale-working' of git://git.kernel.org/pub/scm/linux/kernel/git/npiggin/linux-npiggin:
fs: fix do_last error case when need_reval_dot
nfs: add missing rcu-walk check
fs: hlist UP debug fixup
fs: fix dropping of rcu-walk from force_reval_path
fs: force_reval_path drop rcu-walk before d_invalidate
fs: small rcu-walk documentation fixes
Fixed up trivial conflicts in Documentation/filesystems/porting
diff --git a/Documentation/filesystems/porting b/Documentation/filesystems/porting
index 266d205..dfbcd1b 100644
--- a/Documentation/filesystems/porting
+++ b/Documentation/filesystems/porting
@@ -365,8 +365,8 @@
[recommended]
vfs now tries to do path walking in "rcu-walk mode", which avoids
atomic operations and scalability hazards on dentries and inodes (see
-Documentation/filesystems/path-walk.txt). d_hash and d_compare changes (above)
-are examples of the changes required to support this. For more complex
+Documentation/filesystems/path-lookup.txt). d_hash and d_compare changes
+(above) are examples of the changes required to support this. For more complex
filesystem callbacks, the vfs drops out of rcu-walk mode before the fs call, so
no changes are required to the filesystem. However, this is costly and loses
the benefits of rcu-walk mode. We will begin to add filesystem callbacks that
@@ -383,8 +383,8 @@
permission and check_acl are inode permission checks that are called
on many or all directory inodes on the way down a path walk (to check for
-exec permission). These must now be rcu-walk aware (flags & IPERM_RCU). See
-Documentation/filesystems/vfs.txt for more details.
+exec permission). These must now be rcu-walk aware (flags & IPERM_FLAG_RCU).
+See Documentation/filesystems/vfs.txt for more details.
--
[mandatory]
diff --git a/Documentation/filesystems/vfs.txt b/Documentation/filesystems/vfs.txt
index fbb324e..cae6d27 100644
--- a/Documentation/filesystems/vfs.txt
+++ b/Documentation/filesystems/vfs.txt
@@ -415,8 +415,8 @@
permission: called by the VFS to check for access rights on a POSIX-like
filesystem.
- May be called in rcu-walk mode (flags & IPERM_RCU). If in rcu-walk
- mode, the filesystem must check the permission without blocking or
+ May be called in rcu-walk mode (flags & IPERM_FLAG_RCU). If in rcu-walk
+ mode, the filesystem must check the permission without blocking or
storing to the inode.
If a situation is encountered that rcu-walk cannot handle, return
diff --git a/fs/namei.c b/fs/namei.c
index 0b14f69..8664330 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -479,6 +479,14 @@
struct fs_struct *fs = current->fs;
struct dentry *parent = nd->path.dentry;
+ /*
+ * It can be possible to revalidate the dentry that we started
+ * the path walk with. force_reval_path may also revalidate the
+ * dentry already committed to the nameidata.
+ */
+ if (unlikely(parent == dentry))
+ return nameidata_drop_rcu(nd);
+
BUG_ON(!(nd->flags & LOOKUP_RCU));
if (nd->root.mnt) {
spin_lock(&fs->lock);
@@ -583,6 +591,13 @@
fput(nd->intent.open.file);
}
+/*
+ * Call d_revalidate and handle filesystems that request rcu-walk
+ * to be dropped. This may be called and return in rcu-walk mode,
+ * regardless of success or error. If -ECHILD is returned, the caller
+ * must return -ECHILD back up the path walk stack so path walk may
+ * be restarted in ref-walk mode.
+ */
static int d_revalidate(struct dentry *dentry, struct nameidata *nd)
{
int status;
@@ -673,6 +688,9 @@
return 0;
if (!status) {
+ /* Don't d_invalidate in rcu-walk mode */
+ if (nameidata_drop_rcu(nd))
+ return -ECHILD;
d_invalidate(dentry);
status = -ESTALE;
}
@@ -2105,11 +2123,13 @@
dir = nd->path.dentry;
case LAST_DOT:
if (need_reval_dot(dir)) {
- error = d_revalidate(nd->path.dentry, nd);
- if (!error)
- error = -ESTALE;
- if (error < 0)
+ int status = d_revalidate(nd->path.dentry, nd);
+ if (!status)
+ status = -ESTALE;
+ if (status < 0) {
+ error = status;
goto exit;
+ }
}
/* fallthrough */
case LAST_ROOT:
diff --git a/fs/nfs/dir.c b/fs/nfs/dir.c
index 64ee240..df8c03a 100644
--- a/fs/nfs/dir.c
+++ b/fs/nfs/dir.c
@@ -1406,11 +1406,15 @@
static int nfs_open_revalidate(struct dentry *dentry, struct nameidata *nd)
{
struct dentry *parent = NULL;
- struct inode *inode = dentry->d_inode;
+ struct inode *inode;
struct inode *dir;
struct nfs_open_context *ctx;
int openflags, ret = 0;
+ if (nd->flags & LOOKUP_RCU)
+ return -ECHILD;
+
+ inode = dentry->d_inode;
if (!is_atomic_open(nd) || d_mountpoint(dentry))
goto no_open;
diff --git a/include/linux/list_bl.h b/include/linux/list_bl.h
index 9ee97e7..b2adbb4 100644
--- a/include/linux/list_bl.h
+++ b/include/linux/list_bl.h
@@ -16,7 +16,7 @@
* some fast and compact auxiliary data.
*/
-#if defined(CONFIG_SMP) || defined(CONFIG_DEBUG_SPINLOCK)
+#if defined(CONFIG_SMP)
#define LIST_BL_LOCKMASK 1UL
#else
#define LIST_BL_LOCKMASK 0UL
@@ -62,7 +62,8 @@
struct hlist_bl_node *n)
{
LIST_BL_BUG_ON((unsigned long)n & LIST_BL_LOCKMASK);
- LIST_BL_BUG_ON(!((unsigned long)h->first & LIST_BL_LOCKMASK));
+ LIST_BL_BUG_ON(((unsigned long)h->first & LIST_BL_LOCKMASK) !=
+ LIST_BL_LOCKMASK);
h->first = (struct hlist_bl_node *)((unsigned long)n | LIST_BL_LOCKMASK);
}
diff --git a/include/linux/rculist_bl.h b/include/linux/rculist_bl.h
index b872b49..cf1244f 100644
--- a/include/linux/rculist_bl.h
+++ b/include/linux/rculist_bl.h
@@ -11,7 +11,8 @@
struct hlist_bl_node *n)
{
LIST_BL_BUG_ON((unsigned long)n & LIST_BL_LOCKMASK);
- LIST_BL_BUG_ON(!((unsigned long)h->first & LIST_BL_LOCKMASK));
+ LIST_BL_BUG_ON(((unsigned long)h->first & LIST_BL_LOCKMASK) !=
+ LIST_BL_LOCKMASK);
rcu_assign_pointer(h->first,
(struct hlist_bl_node *)((unsigned long)n | LIST_BL_LOCKMASK));
}