vfs: make vfs_create break delegations on parent directory

In order to add directory delegation support, we need to break
delegations on the parent whenever there is going to be a change in the
directory.

Add a delegated_inode parameter to vfs_create. Most callers are
converted to pass in NULL, but do_mknodat() is changed to wait for a
delegation break if there is one.

Reviewed-by: Jan Kara <jack@suse.cz>
Reviewed-by: NeilBrown <neil@brown.name>
Signed-off-by: Jeff Layton <jlayton@kernel.org>
Link: https://patch.msgid.link/20251111-dir-deleg-ro-v6-10-52f3feebb2f2@kernel.org
Signed-off-by: Christian Brauner <brauner@kernel.org>

8 files changed, 21 insertions, 9 deletions

diff --git a/fs/ecryptfs/inode.c b/fs/ecryptfs/inode.c
index d109e3763a88..3341f00dd087 100644
--- a/fs/ecryptfs/inode.c
+++ b/fs/ecryptfs/inode.c
@@ -188,7 +188,7 @@ ecryptfs_do_create(struct inode *directory_inode,
 
 	rc = lock_parent(ecryptfs_dentry, &lower_dentry, &lower_dir);
 	if (!rc)
-		rc = vfs_create(&nop_mnt_idmap, lower_dentry, mode);
+		rc = vfs_create(&nop_mnt_idmap, lower_dentry, mode, NULL);
 	if (rc) {
 		printk(KERN_ERR "%s: Failure to create dentry in lower fs; "
 		       "rc = [%d]\n", __func__, rc);
diff --git a/fs/namei.c b/fs/namei.c
index 9586c6aba6ea..b20f053374a5 100644
--- a/fs/namei.c
+++ b/fs/namei.c
@@ -3463,6 +3463,7 @@ static inline umode_t vfs_prepare_mode(struct mnt_idmap *idmap,
  * @idmap:	idmap of the mount the inode was found from
  * @dentry:	dentry of the child file
  * @mode:	mode of the child file
+ * @di:		returns parent inode, if the inode is delegated.
  *
  * Create a new file.
  *
@@ -3472,7 +3473,8 @@ static inline umode_t vfs_prepare_mode(struct mnt_idmap *idmap,
  * On non-idmapped mounts or if permission checking is to be performed on the
  * raw inode simply pass @nop_mnt_idmap.
  */
-int vfs_create(struct mnt_idmap *idmap, struct dentry *dentry, umode_t mode)
+int vfs_create(struct mnt_idmap *idmap, struct dentry *dentry, umode_t mode,
+	       struct delegated_inode *di)
 {
 	struct inode *dir = d_inode(dentry->d_parent);
 	int error;
@@ -3488,6 +3490,9 @@ int vfs_create(struct mnt_idmap *idmap, struct dentry *dentry, umode_t mode)
 	error = security_inode_create(dir, dentry, mode);
 	if (error)
 		return error;
+	error = try_break_deleg(dir, di);
+	if (error)
+		return error;
 	error = dir->i_op->create(idmap, dir, dentry, mode, true);
 	if (!error)
 		fsnotify_create(dir, dentry);
@@ -4358,6 +4363,7 @@ static int may_mknod(umode_t mode)
 static int do_mknodat(int dfd, struct filename *name, umode_t mode,
 		unsigned int dev)
 {
+	struct delegated_inode di = { };
 	struct mnt_idmap *idmap;
 	struct dentry *dentry;
 	struct path path;
@@ -4381,7 +4387,7 @@ retry:
 	idmap = mnt_idmap(path.mnt);
 	switch (mode & S_IFMT) {
 		case 0: case S_IFREG:
-			error = vfs_create(idmap, dentry, mode);
+			error = vfs_create(idmap, dentry, mode, &di);
 			if (!error)
 				security_path_post_mknod(idmap, dentry);
 			break;
@@ -4396,6 +4402,11 @@ retry:
 	}
 out2:
 	end_creating_path(&path, dentry);
+	if (is_delegated(&di)) {
+		error = break_deleg_wait(&di);
+		if (!error)
+			goto retry;
+	}
 	if (retry_estale(error, lookup_flags)) {
 		lookup_flags |= LOOKUP_REVAL;
 		goto retry;
diff --git a/fs/nfsd/nfs3proc.c b/fs/nfsd/nfs3proc.c
index 30ea7ffa2aff..2cb972b5ed99 100644
--- a/fs/nfsd/nfs3proc.c
+++ b/fs/nfsd/nfs3proc.c
@@ -344,7 +344,7 @@ nfsd3_create_file(struct svc_rqst *rqstp, struct svc_fh *fhp,
 	status = fh_fill_pre_attrs(fhp);
 	if (status != nfs_ok)
 		goto out;
-	host_err = vfs_create(&nop_mnt_idmap, child, iap->ia_mode);
+	host_err = vfs_create(&nop_mnt_idmap, child, iap->ia_mode, NULL);
 	if (host_err < 0) {
 		status = nfserrno(host_err);
 		goto out;
diff --git a/fs/nfsd/vfs.c b/fs/nfsd/vfs.c
index 464fd54675f3..de5f46f8c6d3 100644
--- a/fs/nfsd/vfs.c
+++ b/fs/nfsd/vfs.c
@@ -1552,7 +1552,7 @@ nfsd_create_locked(struct svc_rqst *rqstp, struct svc_fh *fhp,
 	err = 0;
 	switch (type) {
 	case S_IFREG:
-		host_err = vfs_create(&nop_mnt_idmap, dchild, iap->ia_mode);
+		host_err = vfs_create(&nop_mnt_idmap, dchild, iap->ia_mode, NULL);
 		if (!host_err)
 			nfsd_check_ignore_resizing(iap);
 		break;
diff --git a/fs/open.c b/fs/open.c
index e440f58e3ce8..92cf2e11781b 100644
--- a/fs/open.c
+++ b/fs/open.c
@@ -1171,7 +1171,7 @@ struct file *dentry_create(const struct path *path, int flags, umode_t mode,
 	if (IS_ERR(f))
 		return f;
 
-	error = vfs_create(mnt_idmap(path->mnt), path->dentry, mode);
+	error = vfs_create(mnt_idmap(path->mnt), path->dentry, mode, NULL);
 	if (!error)
 		error = vfs_open(path, f);
 
diff --git a/fs/overlayfs/overlayfs.h b/fs/overlayfs/overlayfs.h
index 2bdc434941eb..e30441cc9c63 100644
--- a/fs/overlayfs/overlayfs.h
+++ b/fs/overlayfs/overlayfs.h
@@ -235,7 +235,7 @@ static inline int ovl_do_create(struct ovl_fs *ofs,
 				struct inode *dir, struct dentry *dentry,
 				umode_t mode)
 {
-	int err = vfs_create(ovl_upper_mnt_idmap(ofs), dentry, mode);
+	int err = vfs_create(ovl_upper_mnt_idmap(ofs), dentry, mode, NULL);
 
 	pr_debug("create(%pd2, 0%o) = %i\n", dentry, mode, err);
 	return err;
diff --git a/fs/smb/server/vfs.c b/fs/smb/server/vfs.c
index 83ece2de4b23..3747851b61c8 100644
--- a/fs/smb/server/vfs.c
+++ b/fs/smb/server/vfs.c
@@ -188,7 +188,7 @@ int ksmbd_vfs_create(struct ksmbd_work *work, const char *name, umode_t mode)
 	}
 
 	mode |= S_IFREG;
-	err = vfs_create(mnt_idmap(path.mnt), dentry, mode);
+	err = vfs_create(mnt_idmap(path.mnt), dentry, mode, NULL);
 	if (!err) {
 		ksmbd_vfs_inherit_owner(work, d_inode(path.dentry),
 					d_inode(dentry));
diff --git a/include/linux/fs.h b/include/linux/fs.h
index 21876ef1fec9..83b05aec4e10 100644
--- a/include/linux/fs.h
+++ b/include/linux/fs.h
@@ -2111,7 +2111,8 @@ bool inode_owner_or_capable(struct mnt_idmap *idmap,
 /*
  * VFS helper functions..
  */
-int vfs_create(struct mnt_idmap *, struct dentry *, umode_t);
+int vfs_create(struct mnt_idmap *, struct dentry *, umode_t,
+	       struct delegated_inode *);
 struct dentry *vfs_mkdir(struct mnt_idmap *, struct inode *,
 			 struct dentry *, umode_t, struct delegated_inode *);
 int vfs_mknod(struct mnt_idmap *, struct inode *, struct dentry *,