On Wed, Dec 29, 2021 at 12:30:01PM -0800, Stefan Roesch wrote:

+static int io_setxattr_prep(struct io_kiocb *req,
+			const struct io_uring_sqe *sqe)
+{
+	struct io_xattr *ix = &req->xattr;
+	const char __user *path;
+	int ret;
+
+	ret = __io_setxattr_prep(req, sqe);
+	if (ret)
+		return ret;
+
+	path = u64_to_user_ptr(READ_ONCE(sqe->addr3));
+
+	ix->filename = getname_flags(path, LOOKUP_FOLLOW, NULL);
+	if (IS_ERR(ix->filename)) {
+		ret = PTR_ERR(ix->filename);
+		ix->filename = NULL;
+	}
+
+	return ret;
+}

Same question as for getxattr side.  Why bother doing getname in prep
and open-coding the ESTALE retry loop in io_setxattr() proper?

Again, if you have hit the retry_estale() returning true, you are already
on a very slow path; trying to save on getname is completely pointless.
Moreover, had there been a situation where it might have been warranted
(and I really can't imagine one), why would that only be applicable in
io_uring case?