[PATCH 4.15 45/55] serial: 8250_of: fix return code when probe function fails to get reset

[Greg Kroah-Hartman]

4.15-stable review patch.  If anyone has any objections, please let me know.

------------------

From: Masahiro Yamada <yamada.masahiro@xxxxxxxxxxxxx>

commit b9820a31691b771db37afe2054dd3d3a680c1eed upstream.

The error pointer from devm_reset_control_get_optional_shared() is
not propagated.

One of the most common problem scenarios is it returns -EPROBE_DEFER
when the reset controller has not probed yet.  In this case, the
probe of the reset consumer should be deferred.

Fixes: e2860e1f62f2 ("serial: 8250_of: Add reset support")
Signed-off-by: Masahiro Yamada <yamada.masahiro@xxxxxxxxxxxxx>
Reviewed-by: Philipp Zabel <p.zabel@xxxxxxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>

---
 drivers/tty/serial/8250/8250_of.c |    5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

--- a/drivers/tty/serial/8250/8250_of.c
+++ b/drivers/tty/serial/8250/8250_of.c
@@ -136,8 +136,11 @@ static int of_platform_serial_setup(stru
 	}
 
 	info->rst = devm_reset_control_get_optional_shared(&ofdev->dev, NULL);
-	if (IS_ERR(info->rst))
+	if (IS_ERR(info->rst)) {
+		ret = PTR_ERR(info->rst);
 		goto err_dispose;
+	}
+
 	ret = reset_control_deassert(info->rst);
 	if (ret)
 		goto err_dispose;