fs: rcu-walk aware d_revalidate method

Require filesystems be aware of .d_revalidate being called in rcu-walk
mode (nd->flags & LOOKUP_RCU). For now do a simple push down, returning
-ECHILD from all implementations.

Signed-off-by: Nick Piggin <npiggin@kernel.dk>

Documentation/filesystems/Locking

@@ -9,7 +9,7 @@ be able to use diff(1).
 
 --------------------------- dentry_operations --------------------------
 prototypes:
-	int (*d_revalidate)(struct dentry *, int);
+	int (*d_revalidate)(struct dentry *, struct nameidata *);
 	int (*d_hash)(const struct dentry *, const struct inode *,
 			struct qstr *);
 	int (*d_compare)(const struct dentry *, const struct inode *,
@@ -21,14 +21,14 @@ prototypes:
 	char *(*d_dname)((struct dentry *dentry, char *buffer, int buflen);
 
 locking rules:
-		rename_lock	->d_lock	may block
-d_revalidate:	no		no		yes
-d_hash		no		no		no
-d_compare:	yes		no		no
-d_delete:	no		yes		no
-d_release:	no		no		yes
-d_iput:		no		no		yes
-d_dname:	no		no		no
+		rename_lock	->d_lock	may block	rcu-walk
+d_revalidate:	no		no		yes (ref-walk)	maybe
+d_hash		no		no		no		maybe
+d_compare:	yes		no		no		maybe
+d_delete:	no		yes		no		no
+d_release:	no		no		yes		no
+d_iput:		no		no		yes		no
+d_dname:	no		no		no		no
 
 --------------------------- inode_operations --------------------------- 
 prototypes:

Documentation/filesystems/path-lookup.txt

@@ -317,11 +317,10 @@ The detailed design for rcu-walk is like this:
 The cases where rcu-walk cannot continue are:
 * NULL dentry (ie. any uncached path element)
 * parent with d_inode->i_op->permission or ACLs
-* dentries with d_revalidate
 * Following links
 
-In future patches, permission checks and d_revalidate become rcu-walk aware. It
-may be possible eventually to make following links rcu-walk aware.
+In future patches, permission checks become rcu-walk aware. It may be possible
+eventually to make following links rcu-walk aware.
 
 Uncached path elements will always require dropping to ref-walk mode, at the
 very least because i_mutex needs to be grabbed, and objects allocated.

Documentation/filesystems/porting

@@ -360,3 +360,23 @@ i_dentry to be reinitialized before it is freed, so an:
   INIT_LIST_HEAD(&inode->i_dentry);
 
 must be done in the RCU callback.
+
+--
+[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
+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
+are rcu-walk aware, shown below. Filesystems should take advantage of this
+where possible.
+
+--
+[mandatory]
+	d_revalidate is a callback that is made on every path element (if
+the filesystem provides it), which requires dropping out of rcu-walk mode. This
+may now be called in rcu-walk mode (nd->flags & LOOKUP_RCU). -ECHILD should be
+returned if the filesystem cannot handle rcu-walk. See
+Documentation/filesystems/vfs.txt for more details.

Documentation/filesystems/vfs.txt

@@ -863,6 +863,15 @@ struct dentry_operations {
 	dcache. Most filesystems leave this as NULL, because all their
 	dentries in the dcache are valid
 
+	d_revalidate may be called in rcu-walk mode (nd->flags & LOOKUP_RCU).
+	If in rcu-walk mode, the filesystem must revalidate the dentry without
+	blocking or storing to the dentry, d_parent and d_inode should not be
+	used without care (because they can go NULL), instead nd->inode should
+	be used.
+
+	If a situation is encountered that rcu-walk cannot handle, return
+	-ECHILD and it will be called again in ref-walk mode.
+
   d_hash: called when the VFS adds a dentry to the hash table. The first
 	dentry passed to d_hash is the parent directory that the name is
 	to be hashed into. The inode is the dentry's inode.

drivers/staging/autofs/root.c

@@ -154,13 +154,16 @@ static int try_to_fill_dentry(struct dentry *dentry, struct super_block *sb, str
  * yet completely filled in, and revalidate has to delay such
  * lookups..
  */
-static int autofs_revalidate(struct dentry * dentry, struct nameidata *nd)
+static int autofs_revalidate(struct dentry *dentry, struct nameidata *nd)
 {
 	struct inode * dir;
 	struct autofs_sb_info *sbi;
 	struct autofs_dir_ent *ent;
 	int res;
 
+	if (nd->flags & LOOKUP_RCU)
+		return -ECHILD;
+
 	lock_kernel();
 	dir = dentry->d_parent->d_inode;
 	sbi = autofs_sbi(dir->i_sb);

drivers/staging/smbfs/dir.c

@@ -14,6 +14,7 @@
 #include <linux/ctype.h>
 #include <linux/net.h>
 #include <linux/sched.h>
+#include <linux/namei.h>
 
 #include "smb_fs.h"
 #include "smb_mount.h"
@@ -301,13 +302,20 @@ static const struct dentry_operations smbfs_dentry_operations_case =
  * This is the callback when the dcache has a lookup hit.
  */
 static int
-smb_lookup_validate(struct dentry * dentry, struct nameidata *nd)
+smb_lookup_validate(struct dentry *dentry, struct nameidata *nd)
 {
-	struct smb_sb_info *server = server_from_dentry(dentry);
-	struct inode * inode = dentry->d_inode;
-	unsigned long age = jiffies - dentry->d_time;
+	struct smb_sb_info *server;
+	struct inode *inode;
+	unsigned long age;
 	int valid;
 
+	if (nd->flags & LOOKUP_RCU)
+		return -ECHILD;
+
+	server = server_from_dentry(dentry);
+	inode = dentry->d_inode;
+	age = jiffies - dentry->d_time;
+
 	/*
 	 * The default validation is based on dentry age:
 	 * we believe in dentries for a few seconds.  (But each

fs/afs/dir.c

@@ -13,6 +13,7 @@
 #include <linux/module.h>
 #include <linux/init.h>
 #include <linux/fs.h>
+#include <linux/namei.h>
 #include <linux/pagemap.h>
 #include <linux/ctype.h>
 #include <linux/sched.h>
@@ -607,6 +608,9 @@ static int afs_d_revalidate(struct dentry *dentry, struct nameidata *nd)
 	void *dir_version;
 	int ret;
 
+	if (nd->flags & LOOKUP_RCU)
+		return -ECHILD;
+
 	vnode = AFS_FS_I(dentry->d_inode);
 
 	if (dentry->d_inode)

fs/autofs4/root.c

@@ -315,12 +315,19 @@ static void *autofs4_follow_link(struct dentry *dentry, struct nameidata *nd)
  */
 static int autofs4_revalidate(struct dentry *dentry, struct nameidata *nd)
 {
-	struct inode *dir = dentry->d_parent->d_inode;
-	struct autofs_sb_info *sbi = autofs4_sbi(dir->i_sb);
-	int oz_mode = autofs4_oz_mode(sbi);
+	struct inode *dir;
+	struct autofs_sb_info *sbi;
+	int oz_mode;
 	int flags = nd ? nd->flags : 0;
 	int status = 1;
 
+	if (flags & LOOKUP_RCU)
+		return -ECHILD;
+
+	dir = dentry->d_parent->d_inode;
+	sbi = autofs4_sbi(dir->i_sb);
+	oz_mode = autofs4_oz_mode(sbi);
+
 	/* Pending dentry */
 	spin_lock(&sbi->fs_lock);
 	if (autofs4_ispending(dentry)) {

fs/ceph/dir.c

@@ -990,7 +990,12 @@ static int dir_lease_is_valid(struct inode *dir, struct dentry *dentry)
  */
 static int ceph_d_revalidate(struct dentry *dentry, struct nameidata *nd)
 {
-	struct inode *dir = dentry->d_parent->d_inode;
+	struct inode *dir;
+
+	if (nd->flags & LOOKUP_RCU)
+		return -ECHILD;
+
+	dir = dentry->d_parent->d_inode;
 
 	dout("d_revalidate %p '%.*s' inode %p offset %lld\n", dentry,
 	     dentry->d_name.len, dentry->d_name.name, dentry->d_inode,

fs/cifs/dir.c

@@ -656,6 +656,9 @@ cifs_lookup(struct inode *parent_dir_inode, struct dentry *direntry,
 static int
 cifs_d_revalidate(struct dentry *direntry, struct nameidata *nd)
 {
+	if (nd->flags & LOOKUP_RCU)
+		return -ECHILD;
+
 	if (direntry->d_inode) {
 		if (cifs_revalidate_dentry(direntry))
 			return 0;

fs/coda/dir.c

@@ -18,6 +18,7 @@
 #include <linux/errno.h>
 #include <linux/string.h>
 #include <linux/spinlock.h>
+#include <linux/namei.h>
 
 #include <asm/uaccess.h>
 
@@ -541,9 +542,13 @@ static int coda_venus_readdir(struct file *coda_file, void *buf,
 /* called when a cache lookup succeeds */
 static int coda_dentry_revalidate(struct dentry *de, struct nameidata *nd)
 {
-	struct inode *inode = de->d_inode;
+	struct inode *inode;
 	struct coda_inode_info *cii;
 
+	if (nd->flags & LOOKUP_RCU)
+		return -ECHILD;
+
+	inode = de->d_inode;
 	if (!inode || coda_isroot(inode))
 		goto out;
 	if (is_bad_inode(inode))

fs/ecryptfs/dentry.c

@@ -44,12 +44,17 @@
  */
 static int ecryptfs_d_revalidate(struct dentry *dentry, struct nameidata *nd)
 {
-	struct dentry *lower_dentry = ecryptfs_dentry_to_lower(dentry);
-	struct vfsmount *lower_mnt = ecryptfs_dentry_to_lower_mnt(dentry);
+	struct dentry *lower_dentry;
+	struct vfsmount *lower_mnt;
 	struct dentry *dentry_save;
 	struct vfsmount *vfsmount_save;
 	int rc = 1;
 
+	if (nd->flags & LOOKUP_RCU)
+		return -ECHILD;
+
+	lower_dentry = ecryptfs_dentry_to_lower(dentry);
+	lower_mnt = ecryptfs_dentry_to_lower_mnt(dentry);
 	if (!lower_dentry->d_op || !lower_dentry->d_op->d_revalidate)
 		goto out;
 	dentry_save = nd->path.dentry;

fs/fat/namei_vfat.c

@@ -43,6 +43,9 @@ static int vfat_revalidate_shortname(struct dentry *dentry)
 
 static int vfat_revalidate(struct dentry *dentry, struct nameidata *nd)
 {
+	if (nd->flags & LOOKUP_RCU)
+		return -ECHILD;
+
 	/* This is not negative dentry. Always valid. */
 	if (dentry->d_inode)
 		return 1;
@@ -51,6 +54,9 @@ static int vfat_revalidate(struct dentry *dentry, struct nameidata *nd)
 
 static int vfat_revalidate_ci(struct dentry *dentry, struct nameidata *nd)
 {
+	if (nd->flags & LOOKUP_RCU)
+		return -ECHILD;
+
 	/*
 	 * This is not negative dentry. Always valid.
 	 *

fs/fuse/dir.c

@@ -156,8 +156,12 @@ u64 fuse_get_attr_version(struct fuse_conn *fc)
  */
 static int fuse_dentry_revalidate(struct dentry *entry, struct nameidata *nd)
 {
-	struct inode *inode = entry->d_inode;
+	struct inode *inode;
+
+	if (nd->flags & LOOKUP_RCU)
+		return -ECHILD;
 
+	inode = entry->d_inode;
 	if (inode && is_bad_inode(inode))
 		return 0;
 	else if (fuse_dentry_time(entry) < get_jiffies_64()) {

fs/gfs2/dentry.c

@@ -11,6 +11,7 @@
 #include <linux/completion.h>
 #include <linux/buffer_head.h>
 #include <linux/gfs2_ondisk.h>
+#include <linux/namei.h>
 #include <linux/crc32.h>
 
 #include "gfs2.h"
@@ -34,15 +35,23 @@
 
 static int gfs2_drevalidate(struct dentry *dentry, struct nameidata *nd)
 {
-	struct dentry *parent = dget_parent(dentry);
-	struct gfs2_sbd *sdp = GFS2_SB(parent->d_inode);
-	struct gfs2_inode *dip = GFS2_I(parent->d_inode);
-	struct inode *inode = dentry->d_inode;
+	struct dentry *parent;
+	struct gfs2_sbd *sdp;
+	struct gfs2_inode *dip;
+	struct inode *inode;
 	struct gfs2_holder d_gh;
 	struct gfs2_inode *ip = NULL;
 	int error;
 	int had_lock = 0;
 
+	if (nd->flags & LOOKUP_RCU)
+		return -ECHILD;
+
+	parent = dget_parent(dentry);
+	sdp = GFS2_SB(parent->d_inode);
+	dip = GFS2_I(parent->d_inode);
+	inode = dentry->d_inode;
+
 	if (inode) {
 		if (is_bad_inode(inode))
 			goto invalid;

fs/hfs/sysdep.c

@@ -8,15 +8,20 @@
  * This file contains the code to do various system dependent things.
  */
 
+#include <linux/namei.h>
 #include "hfs_fs.h"
 
 /* dentry case-handling: just lowercase everything */
 
 static int hfs_revalidate_dentry(struct dentry *dentry, struct nameidata *nd)
 {
-	struct inode *inode = dentry->d_inode;
+	struct inode *inode;
 	int diff;
 
+	if (nd->flags & LOOKUP_RCU)
+		return -ECHILD;
+
+	inode = dentry->d_inode;
 	if(!inode)
 		return 1;
 

fs/jfs/namei.c

@@ -1608,6 +1608,8 @@ static int jfs_ci_compare(const struct dentry *parent,
 
 static int jfs_ci_revalidate(struct dentry *dentry, struct nameidata *nd)
 {
+	if (nd->flags & LOOKUP_RCU)
+		return -ECHILD;
 	/*
 	 * This is not negative dentry. Always valid.
 	 *

fs/namei.c

@@ -563,10 +563,26 @@ void release_open_intent(struct nameidata *nd)
 		fput(nd->intent.open.file);
 }
 
+static int d_revalidate(struct dentry *dentry, struct nameidata *nd)
+{
+	int status;
+
+	status = dentry->d_op->d_revalidate(dentry, nd);
+	if (status == -ECHILD) {
+		if (nameidata_dentry_drop_rcu(nd, dentry))
+			return status;
+		status = dentry->d_op->d_revalidate(dentry, nd);
+	}
+
+	return status;
+}
+
 static inline struct dentry *
 do_revalidate(struct dentry *dentry, struct nameidata *nd)
 {
-	int status = dentry->d_op->d_revalidate(dentry, nd);
+	int status;
+
+	status = d_revalidate(dentry, nd);
 	if (unlikely(status <= 0)) {
 		/*
 		 * The dentry failed validation.
@@ -574,14 +590,20 @@ do_revalidate(struct dentry *dentry, struct nameidata *nd)
 		 * the dentry otherwise d_revalidate is asking us
 		 * to return a fail status.
 		 */
-		if (!status) {
+		if (status < 0) {
+			/* If we're in rcu-walk, we don't have a ref */
+			if (!(nd->flags & LOOKUP_RCU))
+				dput(dentry);
+			dentry = ERR_PTR(status);
+
+		} else {
+			/* Don't d_invalidate in rcu-walk mode */
+			if (nameidata_dentry_drop_rcu_maybe(nd, dentry))
+				return ERR_PTR(-ECHILD);
 			if (!d_invalidate(dentry)) {
 				dput(dentry);
 				dentry = NULL;
 			}
-		} else {
-			dput(dentry);
-			dentry = ERR_PTR(status);
 		}
 	}
 	return dentry;
@@ -626,7 +648,7 @@ force_reval_path(struct path *path, struct nameidata *nd)
 	if (!need_reval_dot(dentry))
 		return 0;
 
-	status = dentry->d_op->d_revalidate(dentry, nd);
+	status = d_revalidate(dentry, nd);
 	if (status > 0)
 		return 0;
 
@@ -1039,12 +1061,8 @@ static int do_lookup(struct nameidata *nd, struct qstr *name,
 			return -ECHILD;
 
 		nd->seq = seq;
-		if (dentry->d_flags & DCACHE_OP_REVALIDATE) {
-			/* We commonly drop rcu-walk here */
-			if (nameidata_dentry_drop_rcu(nd, dentry))
-				return -ECHILD;
+		if (dentry->d_flags & DCACHE_OP_REVALIDATE)
 			goto need_revalidate;
-		}
 		path->mnt = mnt;
 		path->dentry = dentry;
 		__follow_mount_rcu(nd, path, inode);
@@ -1292,12 +1310,11 @@ static int link_path_walk(const char *name, struct nameidata *nd)
 		 * We may need to check the cached dentry for staleness.
 		 */
 		if (need_reval_dot(nd->path.dentry)) {
-			if (nameidata_drop_rcu_maybe(nd))
-				return -ECHILD;
-			err = -ESTALE;
 			/* Note: we do not d_invalidate() */
-			if (!nd->path.dentry->d_op->d_revalidate(
-					nd->path.dentry, nd))
+			err = d_revalidate(nd->path.dentry, nd);
+			if (!err)
+				err = -ESTALE;
+			if (err < 0)
 				break;
 		}
 return_base:
@@ -2080,10 +2097,11 @@ static struct file *do_last(struct nameidata *nd, struct path *path,
 		dir = nd->path.dentry;
 	case LAST_DOT:
 		if (need_reval_dot(dir)) {
-			if (!dir->d_op->d_revalidate(dir, nd)) {
+			error = d_revalidate(nd->path.dentry, nd);
+			if (!error)
 				error = -ESTALE;
+			if (error < 0)
 				goto exit;
-			}
 		}
 		/* fallthrough */
 	case LAST_ROOT:

fs/ncpfs/dir.c

@@ -17,6 +17,7 @@
 #include <linux/kernel.h>
 #include <linux/vmalloc.h>
 #include <linux/mm.h>
+#include <linux/namei.h>
 #include <asm/uaccess.h>
 #include <asm/byteorder.h>
 
@@ -308,6 +309,9 @@ ncp_lookup_validate(struct dentry *dentry, struct nameidata *nd)
 	int res, val = 0, len;
 	__u8 __name[NCP_MAXPATHLEN + 1];
 
+	if (nd->flags & LOOKUP_RCU)
+		return -ECHILD;
+
 	parent = dget_parent(dentry);
 	dir = parent->d_inode;
 

fs/ncpfs/inode.c

@@ -29,6 +29,7 @@
 #include <linux/vfs.h>
 #include <linux/mount.h>
 #include <linux/seq_file.h>
+#include <linux/namei.h>
 
 #include <linux/ncp_fs.h>