Re: [V9fs-developer] [GIT PULL] 9p file system bug fixes for 2.6.35-rc2

["Aneesh Kumar K. V" <aneesh.kumar@linux.vnet.ibm.com>]

On Tue, 29 Jun 2010 13:38:57 -0700, Linus Torvalds <torvalds@linux-foundation.org> wrote:
> On Tue, Jun 29, 2010 at 1:18 PM, Eric Van Hensbergen <ericvh@gmail.com> wrote:
> >
> > Does this approach satisfy your concerns?  We've been going over
> > several different options on how to proceed, but this seems to be the
> > best option.
> 
> Using a p9fs rename lock does seem to be a reasonable option.
> 
> That said, the patch itself seems to not be valid. You drop the lock
> too early in v9fs_fid_lookup() as far as I can tell. You then re-take
> it before doing that whole
> 
>    for (d = dentry, i = (n-1); i >= 0; i--, d = d->d_parent)
> 
> loop with it held again, but that's totally bogus - because you
> dropped the lock, your 'n-1' count has absolutely no meaning any more,
> since a cross-directory rename might have changed the depth of the
> thing in the meantime.
> 
> And if the depth changes, you aren't at all guaranteed to stay on the
> same p9fs filesystem, so now you're doing that d_parent access without
> the proper locking (sure: you hold the rename lock, but it's not at
> all guaranteed that the rename lock is the _right_ lock any more as
> you traverse the list down!)
> 
> But I didn't look deeply at the patch. There might be some reason why
> it's safe (I doubt it, though), and there might be other places where
> you do the same. But in general, dropping and re-taking a lock is a
> bad idea. If you dropped the lock, you can't depend on anything you
> found out while having held it.

You are correct. we cannot drop the rename lock in between. I also found
another issue in that we are using dentry->d_name.name directly. That
would imply we need to hold the rename_lock even during the
client_walk.  How about the patch below ?. I updated the patch to hold
rename_lock during multiple path walk. Also the rename path is updated
to hold the lock during p9_client_rename operations.

commit 79f6f20dbb70ad35db37b674957c95de20662a75
Author: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
Date:   Wed Jun 30 15:37:58 2010 +0530

    fs/9p: Prevent parallel rename when doing fid_lookup
    
    During fid lookup we need to make sure that the dentry->d_parent doesn't
    change so that we can safely walk the parent dentries. To ensure that
    we need to prevent cross directory rename during fid_lookup. Add a
    per superblock rename_lock rwlock to prevent parallel fid lookup and rename.
    
    Signed-off-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>

diff --git a/fs/9p/fid.c b/fs/9p/fid.c
index 5d6cfcb..7b387fe 100644
--- a/fs/9p/fid.c
+++ b/fs/9p/fid.c
@@ -97,6 +97,34 @@ static struct p9_fid *v9fs_fid_find(struct dentry *dentry, u32 uid, int any)
 	return ret;
 }
 
+/*
+ * We need to hold v9ses->rename_lock as long as we hold references
+ * to returned path array. Array element contain pointers to
+ * dentry names.
+ */
+static int build_path_from_dentry(struct v9fs_session_info *v9ses,
+				  struct dentry *dentry, char ***names)
+{
+	int n = 0, i;
+	char **wnames;
+	struct dentry *ds;
+
+	for (ds = dentry; !IS_ROOT(ds); ds = ds->d_parent)
+		n++;
+
+	wnames = kmalloc(sizeof(char *) * n, GFP_KERNEL);
+	if (!wnames)
+		goto err_out;
+
+	for (ds = dentry, i = (n-1); i >= 0; i--, ds = ds->d_parent)
+		wnames[i] = (char  *)ds->d_name.name;
+
+	*names = wnames;
+	return n;
+err_out:
+	return -ENOMEM;
+}
+
 /**
  * v9fs_fid_lookup - lookup for a fid, try to walk if not found
  * @dentry: dentry to look for fid in
@@ -112,7 +140,7 @@ struct p9_fid *v9fs_fid_lookup(struct dentry *dentry)
 	int i, n, l, clone, any, access;
 	u32 uid;
 	struct p9_fid *fid, *old_fid = NULL;
-	struct dentry *d, *ds;
+	struct dentry *ds;
 	struct v9fs_session_info *v9ses;
 	char **wnames, *uname;
 
@@ -139,50 +167,63 @@ struct p9_fid *v9fs_fid_lookup(struct dentry *dentry)
 	fid = v9fs_fid_find(dentry, uid, any);
 	if (fid)
 		return fid;
-
+	/*
+	 * we don't have a matching fid. To do a TWALK we need
+	 * parent fid. We need to prevent rename when we want to
+	 * look at the parent.
+	 */
+	read_lock(&v9ses->rename_lock);
 	ds = dentry->d_parent;
 	fid = v9fs_fid_find(ds, uid, any);
-	if (!fid) { /* walk from the root */
-		n = 0;
-		for (ds = dentry; !IS_ROOT(ds); ds = ds->d_parent)
-			n++;
-
-		fid = v9fs_fid_find(ds, uid, any);
-		if (!fid) { /* the user is not attached to the fs yet */
-			if (access == V9FS_ACCESS_SINGLE)
-				return ERR_PTR(-EPERM);
-
-			if (v9fs_proto_dotu(v9ses) ||
-				v9fs_proto_dotl(v9ses))
-				uname = NULL;
-			else
-				uname = v9ses->uname;
+	if (fid) {
+		/* Found the parent fid do a lookup with that */
+		fid = p9_client_walk(fid, 1, (char **)&ds->d_name.name, 1);
+		read_unlock(&v9ses->rename_lock);
+		return fid;
+	}
+	read_unlock(&v9ses->rename_lock);
 
-			fid = p9_client_attach(v9ses->clnt, NULL, uname, uid,
-				v9ses->aname);
+	/* start from the root and try to do a lookup */
+	fid = v9fs_fid_find(dentry->d_sb->s_root, uid, any);
+	if (!fid) {
+		/* the user is not attached to the fs yet */
+		if (access == V9FS_ACCESS_SINGLE)
+			return ERR_PTR(-EPERM);
 
-			if (IS_ERR(fid))
-				return fid;
+		if (v9fs_proto_dotu(v9ses) || v9fs_proto_dotl(v9ses))
+				uname = NULL;
+		else
+			uname = v9ses->uname;
 
-			v9fs_fid_add(ds, fid);
-		}
-	} else /* walk from the parent */
-		n = 1;
+		fid = p9_client_attach(v9ses->clnt, NULL, uname, uid,
+				       v9ses->aname);
+		if (IS_ERR(fid))
+			return fid;
 
-	if (ds == dentry)
+		v9fs_fid_add(dentry->d_sb->s_root, fid);
+	}
+	/* If we are root ourself just return that */
+	if (dentry->d_sb->s_root == dentry)
 		return fid;
-
-	wnames = kmalloc(sizeof(char *) * n, GFP_KERNEL);
-	if (!wnames)
-		return ERR_PTR(-ENOMEM);
-
-	for (d = dentry, i = (n-1); i >= 0; i--, d = d->d_parent)
-		wnames[i] = (char *) d->d_name.name;
-
+	/*
+	 * Do a multipath walk with attached root.
+	 * When walking parent we need to make sure we
+	 * don't have a parallel rename happening
+	 */
+	read_lock(&v9ses->rename_lock);
+	n  = build_path_from_dentry(v9ses, dentry, &wnames);
+	if (n < 0) {
+		fid = ERR_CAST(ERR_PTR(n));
+		goto err_out;
+	}
 	clone = 1;
 	i = 0;
 	while (i < n) {
 		l = min(n - i, P9_MAXWELEM);
+		/*
+		 * We need to hold rename lock when doing a multipath
+		 * walk to ensure none of the patch component change
+		 */
 		fid = p9_client_walk(fid, l, &wnames[i], clone);
 		if (IS_ERR(fid)) {
 			if (old_fid) {
@@ -194,15 +235,16 @@ struct p9_fid *v9fs_fid_lookup(struct dentry *dentry)
 				p9_client_clunk(old_fid);
 			}
 			kfree(wnames);
-			return fid;
+			goto err_out;
 		}
 		old_fid = fid;
 		i += l;
 		clone = 0;
 	}
-
 	kfree(wnames);
 	v9fs_fid_add(dentry, fid);
+err_out:
+	read_unlock(&v9ses->rename_lock);
 	return fid;
 }
 
diff --git a/fs/9p/v9fs.c b/fs/9p/v9fs.c
index 3c49201..b41bcef 100644
--- a/fs/9p/v9fs.c
+++ b/fs/9p/v9fs.c
@@ -237,6 +237,7 @@ struct p9_fid *v9fs_session_init(struct v9fs_session_info *v9ses,
 		__putname(v9ses->uname);
 		return ERR_PTR(-ENOMEM);
 	}
+	rwlock_init(&v9ses->rename_lock);
 
 	rc = bdi_setup_and_register(&v9ses->bdi, "9p", BDI_CAP_MAP_COPY);
 	if (rc) {
diff --git a/fs/9p/v9fs.h b/fs/9p/v9fs.h
index bec4d0b..dee4f26 100644
--- a/fs/9p/v9fs.h
+++ b/fs/9p/v9fs.h
@@ -104,6 +104,7 @@ struct v9fs_session_info {
 	struct p9_client *clnt;	/* 9p client */
 	struct list_head slist; /* list of sessions registered with v9fs */
 	struct backing_dev_info bdi;
+	rwlock_t rename_lock;
 };
 
 struct p9_fid *v9fs_session_init(struct v9fs_session_info *, const char *,
diff --git a/fs/9p/vfs_inode.c b/fs/9p/vfs_inode.c
index 503a6a2..eae89ad 100644
--- a/fs/9p/vfs_inode.c
+++ b/fs/9p/vfs_inode.c
@@ -964,6 +964,7 @@ static struct dentry *v9fs_vfs_lookup(struct inode *dir, struct dentry *dentry,
 
 	sb = dir->i_sb;
 	v9ses = v9fs_inode2v9ses(dir);
+	/* We can walk d_parent because we hold the dir->i_mutex */
 	dfid = v9fs_fid_lookup(dentry->d_parent);
 	if (IS_ERR(dfid))
 		return ERR_CAST(dfid);
@@ -1049,7 +1050,7 @@ v9fs_vfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 	struct p9_fid *olddirfid;
 	struct p9_fid *newdirfid;
 	struct p9_wstat wstat;
-	int retval;
+	int retval, cross_dir_rename = 0;
 
 	P9_DPRINTK(P9_DEBUG_VFS, "\n");
 	retval = 0;
@@ -1070,6 +1071,9 @@ v9fs_vfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 		retval = PTR_ERR(newdirfid);
 		goto clunk_olddir;
 	}
+	cross_dir_rename = (old_dentry->d_parent != new_dentry->d_parent);
+	if (cross_dir_rename)
+		write_lock(&v9ses->rename_lock);
 
 	if (v9fs_proto_dotl(v9ses)) {
 		retval = p9_client_rename(oldfid, newdirfid,
@@ -1077,21 +1081,27 @@ v9fs_vfs_rename(struct inode *old_dir, struct dentry *old_dentry,
 		if (retval != -ENOSYS)
 			goto clunk_newdir;
 	}
+	if (cross_dir_rename) {
+		/*
+		 * 9P .u can only handle file rename in the same directory
+		 */
 
-	/* 9P can only handle file rename in the same directory */
-	if (memcmp(&olddirfid->qid, &newdirfid->qid, sizeof(newdirfid->qid))) {
 		P9_DPRINTK(P9_DEBUG_ERROR,
 				"old dir and new dir are different\n");
 		retval = -EXDEV;
 		goto clunk_newdir;
 	}
-
 	v9fs_blank_wstat(&wstat);
 	wstat.muid = v9ses->uname;
 	wstat.name = (char *) new_dentry->d_name.name;
 	retval = p9_client_wstat(oldfid, &wstat);
 
 clunk_newdir:
+	if (!retval)
+		/* successful rename */
+		d_move(old_dentry, new_dentry);
+	if (cross_dir_rename)
+		write_unlock(&v9ses->rename_lock);
 	p9_client_clunk(newdirfid);
 
 clunk_olddir:
diff --git a/fs/9p/vfs_super.c b/fs/9p/vfs_super.c
index 0740675..3abc3ec 100644
--- a/fs/9p/vfs_super.c
+++ b/fs/9p/vfs_super.c
@@ -287,4 +287,5 @@ struct file_system_type v9fs_fs_type = {
 	.get_sb = v9fs_get_sb,
 	.kill_sb = v9fs_kill_super,
 	.owner = THIS_MODULE,
+	.fs_flags = FS_RENAME_DOES_D_MOVE,
 };