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

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

From: Alexey Khoroshilov <[email protected]>

commit c91261437985d481c472639d4397931d77f5d4e9 upstream.

sccnxp_probe() returns result of regulator_disable() that may lead
to returning zero, while device is not properly initialized.
Also the driver enables clocks, but it does not disable it.

Found by Linux Driver Verification project (linuxtesting.org).

Signed-off-by: Alexey Khoroshilov <[email protected]>
Cc: Ben Hutchings <[email protected]>
Signed-off-by: Greg Kroah-Hartman <[email protected]>

---
 drivers/tty/serial/sccnxp.c |   13 +++++++++++--
 1 file changed, 11 insertions(+), 2 deletions(-)

--- a/drivers/tty/serial/sccnxp.c
+++ b/drivers/tty/serial/sccnxp.c
@@ -889,7 +889,16 @@ static int sccnxp_probe(struct platform_
                        goto err_out;
                uartclk = 0;
        } else {
-               clk_prepare_enable(clk);
+               ret = clk_prepare_enable(clk);
+               if (ret)
+                       goto err_out;
+
+               ret = devm_add_action_or_reset(&pdev->dev,
+                               (void(*)(void *))clk_disable_unprepare,
+                               clk);
+               if (ret)
+                       goto err_out;
+
                uartclk = clk_get_rate(clk);
        }
 
@@ -988,7 +997,7 @@ static int sccnxp_probe(struct platform_
        uart_unregister_driver(&s->uart);
 err_out:
        if (!IS_ERR(s->regulator))
-               return regulator_disable(s->regulator);
+               regulator_disable(s->regulator);
 
        return ret;
 }


Reply via email to